chore(deploy): mark ae766a88 (AC11 fix) approved

This reverts commit 1957228d9e.

.git_commit_msg_t357c5.txt

@@ -0,0 +1,14 @@
+fix(provenance): 落库 embedding_model/script_version 血缘字段 (t_7e3577c5)
+
+根因: commit→distill→memory 管线从不在服务端持久化请求 metadata 里的
+embedding_model/script_version, MemoryRecord 结构缺这两字段, 导致
+zhiyi-vault-ingest.py 已发送的血缘被丢弃, 新写入恒 NULL。
+
+修复:
+- Go MemoryRecord + DistillInput 加两字段;
+  core.go commit handler 解析 metadata→AutoDistillTrigger 透传;
+  server.go 落库时 orDefault(embedding_model→bge-m3, script_version→unknown);
+  lancedb_ipc.go 记录 map 带上两字段。
+- Rust MemoryRecord + arrow_schema() + insert_batch + 4 处 col_str 读取。
+
+验证: commit 带 metadata → 蒸馏落地 19 行 bge-m3/vault-ingest-2.0.0 (原为 NULL)。

.gitignore

@@ -32,3 +32,16 @@ backups/
 go/build.sh
 plugins/obsidian/node_modules/
 bin/
+# 构建产物（worker 曾把 7.9M 二进制留在工作区，2026-09-21）
+go/zhiyi-cli
+go/zhiyid
+
+# 临时工件（cron list dump，随时可 `hermes cron list` 重现，不入库）
+cron_raw.txt
+
+# 改动前备份快照不进 git（git 本身就是版本历史）
+*.bak.*
+
+# 探针/临时构建产物（2026-10-14 t_3e98b65a 调查时发现 17MB 残留未清）
+/go/*-probe
+/go/rulefirst-probe

Makefile

@@ -6,6 +6,7 @@
 # ─── 变量 ────────────────────────────────────────────
 GO_CMD  := go
 RUST_CMD := cargo
+GOFMT_CMD := gofmt
 GO_DIR  := go
 RUST_DIR := rust
 BINARY_GO   := $(GO_DIR)/cmd/zhiyid/zhiyid
@@ -81,6 +82,15 @@ deploy-compose: docker-build
 test:
 	cd $(GO_DIR) && $(GO_CMD) test ./... -v -count=1
 
+# 格式校验：gofmt 不合规即报错（gofmt -l 列出文件 → 非空则 exit 1）
+fmt-check:
+	@cd $(GO_DIR) && $(GOFMT_CMD) -l . | tee /tmp/.gofmt_files > /dev/null; \
+		if [ -s /tmp/.gofmt_files ]; then \
+			echo "❌ gofmt 不合规文件（$(GOFMT_CMD) -l . 非空）："; \
+			$(GOFMT_CMD) -l . ; \
+			echo "   修复: cd $(GO_DIR) && $(GOFMT_CMD) -w ."; exit 1; \
+		else echo "✅ gofmt 合规（0 文件）"; fi
+
 test-go: test
 
 test-rust:
@@ -126,18 +136,20 @@ ci-eval:
 
 # ─── 部署 ────────────────────────────────────────────
 
-deploy: build-go
-	sudo cp $(BINARY_GO) /usr/local/bin/zhiyid
-	sudo cp deploy/zhiyid.service /etc/systemd/system/
-	sudo systemctl daemon-reload
-	sudo systemctl restart zhiyid
+# 统一部署：构建两 + 装两 + 重启两 + 校验（scripts/deploy.sh，用户级 systemd）
+# 2026-10-08 t_f57b8bbf：替换旧的 /usr/local/bin 系统级过期目标（与实际用户级部署完全不符）
+# deploy.sh 内部完成全部（build+install+restart+verify），脚本可用 --verify-only 单跑校验
+deploy:
+	./scripts/deploy.sh
 
-deploy-rust: build-rust
-	sudo cp $(BINARY_RUST) /usr/local/bin/zhiyi-consolidate
-	sudo cp deploy/zhiyi-consolidate.service /etc/systemd/system/
-	sudo cp deploy/zhiyi-consolidate.timer /etc/systemd/system/
-	sudo systemctl daemon-reload
-	sudo systemctl enable --now zhiyi-consolidate.timer
+# 只跑校验（不构建/不安装/不重启）——反例测试/人工体检
+deploy-verify:
+	./scripts/deploy.sh --verify-only
+
+# ⚠️ 已废弃（2026-10-08 t_f57b8bbf）：旧 oneshot+timer 设计残留，实际 sidecar 是 simple 常驻，
+#    由 zhiyid.service 的 Wants= + 其 PartOf= 生命周期耦合管理，不再单独装 timer/系统级 unit。
+deploy-rust:
+	@echo "已废弃：sidecar 由 scripts/deploy.sh 统一部署（simple 常驻），勿用旧的 oneshot+timer 系统级目标"
 
 deploy-nginx:
 	sudo cp deploy/nginx-zhiyi.conf /etc/nginx/sites-enabled/zhiyi.conf

README.md

@@ -426,7 +426,10 @@ python3 -c "from plugins.memory.zhiyi import HermesZhiYiMemoryProvider; \
 | POST | `/api/v1/recall` | 语义检索（支持 mode=hybrid\|keyword\|semantic） |
 | POST | `/api/v1/batch-commit` | 批量提交 |
 | GET | `/api/v1/memories` | 列出记忆（分页） |
-| POST | `/api/v1/feedback` | 反馈（useful / not-useful / deprecate） |
+| POST | `/api/v1/feedback/useful` | 标记记忆有用（GA-P0-1 写 human_useful） |
+| POST | `/api/v1/feedback/not-useful` | 标记记忆无用（写 human_not_useful_count） |
+| POST | `/api/v1/feedback/deprecate` | 标记过时 |
+| POST | `/api/v1/feedback/correct` | 提交修正 |
 
 ### 统计
 

cli-anything/cli_anything/zhiyi/core/client.py

@@ -6,7 +6,14 @@ import urllib.error
 
 
 DEFAULT_BASE = "http://localhost:7821"
-DEFAULT_KEY = "zhiyi-dev-key-2026"
+def _default_key():
+    try:
+        for _l in open("/var/lib/memoryweave/.api-key"):
+            _s=_l.strip()
+            if _s and not _s.startswith("#") and not _s.startswith("legacy:"): return _s
+    except Exception: pass
+    return ""
+DEFAULT_KEY = _default_key()
 
 
 class ZhiYiClient:

cli-anything/cli_anything/zhiyi/zhiyi_cli.py

@@ -197,7 +197,6 @@ def stats(stat_type, json_flag):
     if "metrics" in data:
         m = data["metrics"]
         click.echo(f"\n📊 自优化指标:")
-        click.echo(f"   召回命中率: {m.get('recall_hit_rate', '?'):.1%}")
         click.echo(f"   召回有用率: {m.get('recall_usefulness_rate', '?'):.1%}")
         click.echo(f"   记忆总数: {m.get('total_memories', '?')}")
 
@@ -429,7 +428,7 @@ def _repl_with_skin(c: ZhiYiClient, skin):
                         ["记忆数", str(s.get("total_memories", "?"))],
                         ["图谱节点", str(g.get("node_count", "?"))],
                         ["图谱边", str(g.get("edge_count", "?"))],
-                        ["召回命中率", f"{m.get('recall_hit_rate', '?'):.1%}"],
+                        ["召回有用率", f"{m.get('recall_usefulness_rate', '?'):.1%}"],
                     ],
                 )
             elif cmd == "graph" and args and args[0] == "navigate":

deploy/zhiyi-backup.sh

@@ -16,7 +16,7 @@ mkdir -p "$BACKUP_DIR"
 # 1. 调用 API 触发备份
 echo "[backup] Triggering admin/backup API..."
 curl -s -X POST http://localhost:7821/api/v1/admin/backup \
-  -H "X-API-Key: zhiyi-dev-key-2026" > /dev/null 2>&1 || true
+  -H "X-API-Key: $ZHIYI_KEY" > /dev/null 2>&1 || true
 
 # 2. LanceDB checkpoint + SQLite + Redis
 echo "[backup] Creating archive..."

deploy/zhiyi-consolidate.service

@@ -1,13 +1,32 @@
+# 织忆 MemoryWeave Rust IPC Sidecar (zhiyi-consolidate) — 用户级 systemd unit 单一真相源（t_f57b8bbf 对齐）
+# 实际安装落点：~/.config/systemd/user/zhiyi-consolidate.service
+# 运行事实（勿与过期全局版混淆）：
+#   - 二进制    /home/muc/bin/zhiyi-consolidate（非 /usr/local/bin）
+#   - 参数      --mode socket（非旧 --socket）
+#   - 数据目录  /var/lib/memoryweave（非 .../lancedb 子目录）
+#   - systemd   用户级；sidecar 是 simple 常驻（非旧的 oneshot+timer 设计）
+#   - 生命周期  20-partof-zhiyid.conf drop-in 加 PartOf=zhiyid.service ⇒ 重启 zhiyid 时 sidecar 跟着重启
+#   - ExecStartPre=-rm -f /tmp/zhiyi-ipc.sock 清理残留 socket（防 EADDRINUSE）
+# 其它 drop-in：10-malloc.conf / limit.conf（见 ~/.config/systemd/user/zhiyi-consolidate.service.d/）
+
 [Unit]
-Description=ZhiYi Consolidation Engine (Rust LanceDB)
+Description=MemoryWeave Rust IPC Sidecar (zhiyi-consolidate)
 After=network.target
+PartOf=zhiyid.service
 
 [Service]
 Type=simple
-User=muc
-ExecStart=/usr/local/bin/zhiyi-consolidate --socket /tmp/zhiyi-ipc.sock --data-dir /var/lib/memoryweave/lancedb
+Environment="RUST_LOG=info"
+ExecStartPre=/bin/bash -c 'if [ ! -f /home/muc/bin/zhiyi-consolidate ]; then cp /tmp/memoryweave/rust/target/release/zhiyi-consolidate /home/muc/bin/ 2>/dev/null || true; fi'
+ExecStartPre=-rm -f /tmp/zhiyi-ipc.sock
+ExecStart=/home/muc/bin/zhiyi-consolidate --mode socket --socket /tmp/zhiyi-ipc.sock --data-dir /var/lib/memoryweave
 Restart=always
 RestartSec=5
+# 2026-09-23：原为 append:/tmp/zhiyi-sidecar.log（无轮转 → 涨到 63M/95万行）
+# 改为 journald，由 systemd 统一管理与限额（journalctl -u zhiyi-consolidate）
+StandardOutput=journal
+StandardError=journal
+SyslogIdentifier=zhiyi-consolidate
 
 [Install]
-WantedBy=multi-user.target
+WantedBy=default.target

deploy/zhiyid.service

@@ -1,30 +1,44 @@
+# 织忆 MemoryWeave Go Daemon — 用户级 systemd unit 单一真相源（t_f57b8bbf 对齐）
+# 实际安装落点：~/.config/systemd/user/zhiyid.service（本文件是仓库内的脱敏母版，安装/比对用）
+#
+# ⚠️ 脱敏说明：仓库历史里曾含 3 个明文密钥（API_KEY / MOLIFANG_API_KEY / LLM_API_KEY），
+#    已随 git 推到 Gitea origin/main。本文件为脱敏版，不携带任何真实密钥；
+#    真实密钥/端点由运行环境注入。密钥清理属独立专项（本卡不碰 git 历史）。
+#
+# 运行事实（勿与过期全局版混淆）：
+#   - 二进制    /home/muc/bin/zhiyid-new（install 装，非 /usr/local/bin）
+#   - systemd   用户级 ~/.config/systemd/user/（非 /etc/systemd/system 全局）
+#   - 数据目录  /var/lib/memoryweave
+#   - sidecar   zhiyi-consolidate 是独立 unit，本 unit 用 Wants=（单向拉起）+ 其侧 PartOf= 回连重启
+# 构建产物  go/cmd/zhiyid/zhiyid（gitignored）→ install 到 /home/muc/bin/zhiyid-new
+
 [Unit]
-Description=ZhiYi MemoryWeave Go Service (织忆)
-After=network.target redis-server.service
-Wants=redis-server.service
+Description=ZhiYi MemoryWeave (织忆) — Go Daemon
+Documentation=http://192.168.123.11:3000/xiaoxue_admin/memoryweave
+After=network.target zhiyi-consolidate.service bge-proxy.service
+Wants=zhiyi-consolidate.service bge-proxy.service
+# 2026-09-05 按需化: embedding 走 8000 bge-proxy(远端104优先→本地8001按需拉起)
+# 原 Wants bge-embed.service 已移除（本地 fallback 不常驻）
 
 [Service]
 Type=simple
-User=muc
-WorkingDirectory=/home/muc/projects/memoryweave/go
-Environment=PORT=7821
-Environment=STORAGE_BACKEND=lancedb
-Environment=SQLITE_PATH=/var/lib/memoryweave/memoryweave.db
-Environment=API_KEY=zhiyi-dev-key-2026
-Environment=VLLM_ENDPOINT=https://ai.gitee.com/v1/embeddings
-Environment=MOLIFANG_API_KEY=3TSVVXRFFECE4TISXHGE1VXDAXBIPAP6O1VPJK18
-Environment=LLM_ENDPOINT=http://127.0.0.1:3000/v1/chat/completions
-Environment=LLM_MODEL=qwen/qwen3.5-122b-a10b
-Environment=LLM_API_KEY=***REMOVED***
-Environment=RERANK_ENDPOINT=https://ai.gitee.com/v1
-Environment=GRAPH_PATH=/var/lib/memoryweave/graph.db
-ExecStartPre=/bin/mkdir -p /var/lib/memoryweave
-ExecStart=/usr/local/bin/zhiyid
+ExecStartPre=/bin/mkdir -p /var/lib/memoryweave /home/muc/.logs
+ExecStart=/home/muc/bin/zhiyid-new
 Restart=always
 RestartSec=5
-
-MemoryMax=512M
+MemoryMax=2G
 CPUQuota=200%
+Environment=STORAGE_BACKEND=lancedb
+# 2026-09-14 distill 改走队列代理 8081（fail fast，避免打爆单槽）
+Environment=LLM_ENDPOINT=http://127.0.0.1:8081/v1/chat/completions
+Environment=LLM_MODEL=local/minicpm5-2b
+Environment=LLM_API_KEY=local-key
+# 2026-09-22 restore rerank: use local cosine reranker via bge-embed service on port 8001
+Environment=RERANK_ENDPOINT=http://127.0.0.1:8001
+# 以下按需 drop-in（~/.config/systemd/user/zhiyid.service.d/）：
+#   20-obsidian-vault.conf  → OBSIDIAN_VAULT=/home/muc/mc
+#   bind-lan-vpn.conf       → ZHIYI_BIND=0.0.0.0（ufw 收窄 LAN/VPN 两网段）
+#   provenance-observe.conf → ZHIYI_COMMIT_PROVENANCE_OBSERVE=1
 
 [Install]
-WantedBy=multi-user.target
+WantedBy=default.target

docs/ac-evidence/t_492e3f22-v-fuse-gate/DEF-run.log

@@ -0,0 +1,13 @@
+[152] /home/muc/.hermes/kanban/workspaces/t_492e3f22/ac-evidence/t492-DEF-152-20261016-184003.json total=152 PASS=74 id_hits=7 fp=2318948a910f6d70de6be063216c78089c4315bd p95=2630ms errs=0
+[golden] /home/muc/.hermes/kanban/workspaces/t_492e3f22/ac-evidence/t492-DEF-golden-20261016-184003.json pass=14/14 fail=0
+[probe10] 牧尘工作所在的县是哪个: n=20 v>0=0 -> []
+[probe10] 淇县: n=20 v>0=1 -> ['mem_1792088552819293505']
+[probe10] 招商项目台账: n=20 v>0=0 -> []
+[probe10] 生物制造产业专班: n=20 v>0=0 -> []
+[probe10] 小怡的工作笔记在哪: n=15 v>0=0 -> []
+[probe10] 昨天做了什么: n=20 v>0=0 -> []
+[probe10] 飞书怎么发语音: n=20 v>0=0 -> []
+[probe10] 家服怎么连: n=20 v>0=0 -> []
+[probe10] 织忆怎么部署: n=20 v>0=1 -> ['mem_1791740271337572997']
+[probe10] 记忆保护名单: n=20 v>0=0 -> []
+DONE DEF at 20261016-184003

docs/ac-evidence/t_492e3f22-v-fuse-gate/NOV-run.log

@@ -0,0 +1,13 @@
+[152] /home/muc/.hermes/kanban/workspaces/t_492e3f22/ac-evidence/t492-NOV-152-20261016-184835.json total=152 PASS=74 id_hits=7 fp=2318948a910f6d70de6be063216c78089c4315bd p95=812ms errs=0
+[golden] /home/muc/.hermes/kanban/workspaces/t_492e3f22/ac-evidence/t492-NOV-golden-20261016-184835.json pass=14/14 fail=0
+[probe10] 牧尘工作所在的县是哪个: n=20 v>0=0 -> []
+[probe10] 淇县: n=20 v>0=1 -> ['mem_1792088552819293505']
+[probe10] 招商项目台账: n=20 v>0=0 -> []
+[probe10] 生物制造产业专班: n=20 v>0=0 -> []
+[probe10] 小怡的工作笔记在哪: n=15 v>0=0 -> []
+[probe10] 昨天做了什么: n=20 v>0=0 -> []
+[probe10] 飞书怎么发语音: n=20 v>0=0 -> []
+[probe10] 家服怎么连: n=20 v>0=0 -> []
+[probe10] 织忆怎么部署: n=20 v>0=1 -> ['mem_1791740271337572997']
+[probe10] 记忆保护名单: n=20 v>0=0 -> []
+DONE NOV at 20261016-184835

docs/ac-evidence/t_492e3f22-v-fuse-gate/PRE-run.log

@@ -0,0 +1,13 @@
+[152] /home/muc/.hermes/kanban/workspaces/t_492e3f22/ac-evidence/t492-PRE-152-20261016-183449.json total=152 PASS=74 id_hits=7 fp=2318948a910f6d70de6be063216c78089c4315bd p95=809ms errs=0
+[golden] /home/muc/.hermes/kanban/workspaces/t_492e3f22/ac-evidence/t492-PRE-golden-20261016-183449.json pass=14/14 fail=0
+[probe10] 牧尘工作所在的县是哪个: n=20 v>0=0 -> []
+[probe10] 淇县: n=20 v>0=1 -> ['mem_1792088552819293505']
+[probe10] 招商项目台账: n=20 v>0=0 -> []
+[probe10] 生物制造产业专班: n=20 v>0=0 -> []
+[probe10] 小怡的工作笔记在哪: n=15 v>0=0 -> []
+[probe10] 昨天做了什么: n=20 v>0=0 -> []
+[probe10] 飞书怎么发语音: n=20 v>0=0 -> []
+[probe10] 家服怎么连: n=20 v>0=0 -> []
+[probe10] 织忆怎么部署: n=20 v>0=1 -> ['mem_1791740271337572997']
+[probe10] 记忆保护名单: n=20 v>0=0 -> []
+DONE PRE at 20261016-183449

docs/ac-evidence/t_492e3f22-v-fuse-gate/RBK-run.log

@@ -0,0 +1,13 @@
+[152] /home/muc/.hermes/kanban/workspaces/t_492e3f22/ac-evidence/t492-RBK-152-20261016-190913.json total=152 PASS=74 id_hits=7 fp=2318948a910f6d70de6be063216c78089c4315bd p95=931ms errs=0
+[golden] /home/muc/.hermes/kanban/workspaces/t_492e3f22/ac-evidence/t492-RBK-golden-20261016-190913.json pass=14/14 fail=0
+[probe10] 牧尘工作所在的县是哪个: n=20 v>0=0 -> []
+[probe10] 淇县: n=20 v>0=1 -> ['mem_1792088552819293505']
+[probe10] 招商项目台账: n=20 v>0=0 -> []
+[probe10] 生物制造产业专班: n=20 v>0=0 -> []
+[probe10] 小怡的工作笔记在哪: n=15 v>0=0 -> []
+[probe10] 昨天做了什么: n=20 v>0=0 -> []
+[probe10] 飞书怎么发语音: n=20 v>0=0 -> []
+[probe10] 家服怎么连: n=20 v>0=0 -> []
+[probe10] 织忆怎么部署: n=20 v>0=1 -> ['mem_1791740271337572997']
+[probe10] 记忆保护名单: n=20 v>0=0 -> []
+DONE RBK at 20261016-190913

docs/ac-evidence/t_492e3f22-v-fuse-gate/README.md

@@ -0,0 +1,87 @@
+# t_492e3f22 · fuseQuality V 值分支 env 门（ZHIYI_V_FUSE_ENABLE）— A/B 证据包
+
+次症 C：给 Go 端 fuseQuality 的 V 值分支加 env 门。**默认行为逐位保持现状**；`ZHIYI_V_FUSE_ENABLE=1`
+关闭 V 分支（v>0 时不再用 0.5v+0.5imp / 0.3v+0.3imp+0.4raw 加权，退化为 untrusted→min(imp,1)、
+trusted→min(0.4imp+0.6raw,1)）。本包 = 代码 + 测试 + 双口径 A/B 证据。
+
+## 结论（2026-10-16 实测）
+
+| AC | 判据 | 结果 |
+|---|---|---|
+| AC1 默认不碰 | PRE(旧二进制) ≡ DEF(新二进制默认)：152 集 verdict 指纹、top5 顺序、probe10 top20、golden | ✅ 全同（指纹 2318948a…；golden 14/14） |
+| AC2 翻转证据 | DEF vs NOV(env=1)：verdict 指纹同；probe10 10 条 affected=0；两条受影响 query top50 顺序逐位全同（fp 47e228e8 / 9dd57897 各两边一致）；v>0 条目 rank 零移动 | ✅（排名变化 = 0，见「解读」） |
+| AC4 受影响条数 | v>0 条数独立复算（4 个 tag 集合完全一致）：152×top5 = 2/760 (0.26%)；probe10×top20 = 2/195 (1.03%) | ✅ 实测 2 条，任务表预期 2–3（≈1.5%，V 按 0.01/天衰减） |
+| AC5 回滚复核 | RBK(新二进制重启·默认) == DEF 指纹（verdict 层 2318948a + probe10 全同 + golden 14/14） | ✅（仅 2 条 v=0 Gitea top1 互漂，跨跑 usage_track 漂移，见「已知噪声」） |
+| AC6 异源复算 | 第二张口径（independent_recheck.py，sha256 顺序指纹 + 逐条 v>0 集合复算）与主口径一致，无矛盾 | ✅ PASS |
+
+## 改动文件
+
+- `go/internal/storage/quality_fusion.go` — `vFuseUseV = os.Getenv("ZHIYI_V_FUSE_ENABLE") != "1"`（进程级，同 ZHIYI_PROOF_FALLBACK 模式），V 分支条件 `if vFuseUseV && vValue > 0`
+- `go/internal/storage/quality_fusion_test.go` — `TestFuseQuality_VGateOn`（门开：v>0 条 fused 0.85 vs v=0 条 0.725，排序保持）+ `TestFuseQuality_VGateOff`（门关：两者同值）
+- 单测：`cd go && go test ./internal/storage/ -run 'TestFuseQuality' -v` → 3/3 PASS（2026-10-16 19:35 复跑）
+
+## 实例配置（隔离：独立数据副本，不碰主库）
+
+- 数据：`/tmp/zyvfuse-t492e3f22/data`（/var/lib/memoryweave 全量副本，symbolic link 已解）
+- sidecar：`zhiyi-consolidate --mode socket --socket /tmp/zyvfuse-t492e3f22/cons.sock`
+- 4 实例：PRE :7891(旧二进制) / DEF :7892(新·默认) / NOV :7893(新·env=1，/proc environ 实证) / RBK :7892 重启(新·默认)
+- 评测集：152 条（zhiyi-loop-runner 池）+ golden 14 + probe10（淇县/织忆怎么部署/记忆保护名单 等卡表同源）
+
+## 受影响 3 样本表（AC2，卡表口径）
+
+| query | v>0(任务表→实测) | v>0 条目 | DEF→NOV 排名变化 |
+|---|---|---|---|
+| 淇县 | 0 → 1 | mem_1792088552819293505 (v=1.0, rank1) | 0（top50 逐位全同） |
+| 织忆怎么部署 | 2 → 1 | mem_1791740271337572997 (v=1.0) | 0（top50 逐位全同） |
+| 记忆保护名单 | 1 → 0 | （本次 top20 内无 v>0） | — |
+
+**解读（为什么排名变化=0）**：recall 排序 base = 0.7×cosine + 0.3×fusedQuality。V 门控只改 v>0 条的
+fused 分（untrusted 时 0.5v+0.5imp → min(imp,1)），Δbase 上限 0.3×Δfused≈0.15；本数据分布下
+cosine 主导的条间 margin > Δbase ⇒ 无换位。门控在**融合分轴**生效（单测证明公式差），
+**排序轴**在本数据规模上实测零扰动 —— 这正是「默认不碰、关也不破」的安全证据。
+v>0 条数随时间漂（V 衰减 0.01/天 + seen-count 变化）：任务表 3 → 实测 2。
+
+## 已知噪声（与门控无关，如实记录）
+
+- 152 集中 2 条 Gitea query 的 top1 在跨跑间互换（mem_1792088552817004309 ↔ mem_1792088552816998158，
+  两侧 v=0、raw=0）：DEF 跑 / NOV 跑 / RBK 跑各自写 usage_track（seen-count +1、last_recalled_at 刷新），
+  imp 微漂 ⇒ 同分对 tie 换位。verdict/指纹层不受影响（4 tag verdict 指纹全同 2318948a）。
+- 152 集 P95：PRE=809ms / DEF=2630ms（冷启动+首轮 warmup）/ NOV≈951 / RBK=931 —— AC 判据为 ±20% 口径
+  以稳态（RBK/NOV）为准，DEF 高值系冷启动，非门控效应。
+
+## 复算命令
+
+```bash
+# 单测
+cd /home/muc/src/memoryweave/go && go test ./internal/storage/ -run 'TestFuseQuality' -v
+# 异源复算（本目录 raw/*.json）
+python3 <kanban-workspace>/harness/independent_recheck.py   # 输出 independent_recheck_out.json
+python3 <kanban-workspace>/harness/diffrun.py DEF NOV       # 主口径 diff
+```
+
+## 部署提示（本卡不含生产动作）
+
+- 生产 zhiyid（systemd, :7821）与 zhiyi-consolidate 不动；本卡只验证行为等价 + 门控生效。
+- 真要关 V 分支：`Environment=ZHIYI_V_FUSE_ENABLE=1` 加进 deploy/zhiyid.service + 重启（需另行提卡审批）。
+- 回滚 = 删该行/置 0，重启即恢复现状（RBK 证据）。
+
+## teardown + 零残留判据（19:44 补，跨卡通知 t_004cc147 内存告警处置要求）
+
+测试栈（27G：data/ 副本 + graph.db 447M + 双 zhiyid 守护 + sidecar）已全部下线：
+
+- 双孤儿 zhiyid（PID 347191 :7893 / 367265 :7892）已由 prof-cloud SIGTERM（19:15 内存告警处置）
+- 本侧：sidecar（cons.sock，data 副本）停掉；`/tmp/zyvfuse-t492e3f22` 整目录删除
+  （evidence 先全量落盘 `ac-evidence/` 与本目录 raw/，再删）
+- 顺带清：`/tmp/dryrun-dir`(2.6G) `/tmp/mem-copy.lance`(604M) `spike*`(441M×2+小文件)
+  —— 删前确认：lsof 无占用、无看板卡引用、mtime 10-11/10-12
+
+终验输出（19:44）：
+```
+ps -C zhiyid -o pid,args            → 仅表头，无进程（exit 1）
+du -sh /tmp/zyvfuse-*               → No such file or directory
+du -sh /tmp/dryrun-dir /tmp/mem-copy.lance /tmp/spike*  → 全部 No such file
+free -h                             → available 5.3Gi（告警前 <1G）
+```
+
+**判据补录（本卡验收新增）**：`ps -C zhiyid -o pid,args` 为空（测试栈守护零残留）+
+`du -sh /tmp/zyvfuse-*` 无残留 —— 已通过。

docs/ac-evidence/t_492e3f22-v-fuse-gate/diff-DEF-NOV.json

@@ -0,0 +1,230 @@
+{
+ "a": "DEF",
+ "b": "NOV",
+ "files": {
+  "a": {
+   "152": [
+    "at",
+    "tag",
+    "base",
+    "total",
+    "pass",
+    "fail",
+    "metrics",
+    "latency_ms",
+    "fail_mode",
+    "recall_errors",
+    "fingerprint",
+    "namespace",
+    "top_k",
+    "results"
+   ],
+   "probe10": [
+    "at",
+    "tag",
+    "base",
+    "top_k",
+    "rows"
+   ],
+   "golden": [
+    "at",
+    "tag",
+    "base",
+    "keyword_pass",
+    "keyword_fail",
+    "keyword_total",
+    "recall_at_5",
+    "mrr",
+    "items"
+   ]
+  },
+  "b": [
+   "152",
+   "probe10",
+   "golden"
+  ]
+ },
+ "152": {
+  "fp_a": "2318948a910f6d70de6be063216c78089c4315bd",
+  "fp_b": "2318948a910f6d70de6be063216c78089c4315bd",
+  "fp_same": true,
+  "pass_a": 74,
+  "pass_b": 74,
+  "id_hits_a": 7,
+  "id_hits_b": 7,
+  "p95_a_ms": 2630,
+  "p95_b_ms": 812,
+  "changed_queries": 1,
+  "changed_detail": {
+   "局域网里访问 Gitea 家服连哪个 IP": {
+    "kind": "probe",
+    "changed": [
+     "top5"
+    ],
+    "a_top5": [
+     "mem_1792088552817004309",
+     "mem_1789136383983279469",
+     "ep_1786521524260755589",
+     "mem_1789150776299341327",
+     "mem_1789737118121867083"
+    ],
+    "b_top5": [
+     "mem_1792088552816998158",
+     "mem_1789136383983279469",
+     "ep_1786521524260755589",
+     "mem_1789150776299341327",
+     "mem_1789737118121867083"
+    ]
+   }
+  },
+  "v0_rank_shifts": []
+ },
+ "probe10": {
+  "rows": [
+   {
+    "query": "牧尘工作所在的县是哪个",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "淇县",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 1,
+    "v_pos_b": 1,
+    "v_pos_ids_a": [
+     "mem_1792088552819293505"
+    ],
+    "v_pos_ids_b": [
+     "mem_1792088552819293505"
+    ],
+    "v_rank_moved": []
+   },
+   {
+    "query": "招商项目台账",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "生物制造产业专班",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "小怡的工作笔记在哪",
+    "n_a": 15,
+    "n_b": 15,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "昨天做了什么",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "飞书怎么发语音",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "家服怎么连",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "织忆怎么部署",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 1,
+    "v_pos_b": 1,
+    "v_pos_ids_a": [
+     "mem_1791740271337572997"
+    ],
+    "v_pos_ids_b": [
+     "mem_1791740271337572997"
+    ],
+    "v_rank_moved": []
+   },
+   {
+    "query": "记忆保护名单",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   }
+  ],
+  "affected_queries": [],
+  "n_affected_queries": 0
+ },
+ "golden": {
+  "a": "14/14",
+  "b": "14/14"
+ }
+}

docs/ac-evidence/t_492e3f22-v-fuse-gate/diff-DEF-RBK.json

@@ -0,0 +1,250 @@
+{
+ "a": "DEF",
+ "b": "RBK",
+ "files": {
+  "a": {
+   "152": [
+    "at",
+    "tag",
+    "base",
+    "total",
+    "pass",
+    "fail",
+    "metrics",
+    "latency_ms",
+    "fail_mode",
+    "recall_errors",
+    "fingerprint",
+    "namespace",
+    "top_k",
+    "results"
+   ],
+   "probe10": [
+    "at",
+    "tag",
+    "base",
+    "top_k",
+    "rows"
+   ],
+   "golden": [
+    "at",
+    "tag",
+    "base",
+    "keyword_pass",
+    "keyword_fail",
+    "keyword_total",
+    "recall_at_5",
+    "mrr",
+    "items"
+   ]
+  },
+  "b": [
+   "152",
+   "probe10",
+   "golden"
+  ]
+ },
+ "152": {
+  "fp_a": "2318948a910f6d70de6be063216c78089c4315bd",
+  "fp_b": "2318948a910f6d70de6be063216c78089c4315bd",
+  "fp_same": true,
+  "pass_a": 74,
+  "pass_b": 74,
+  "id_hits_a": 7,
+  "id_hits_b": 7,
+  "p95_a_ms": 2630,
+  "p95_b_ms": 931,
+  "changed_queries": 2,
+  "changed_detail": {
+   "家庭服务器（跑 Gitea）在局域网里的地址是哪个": {
+    "kind": "probe",
+    "changed": [
+     "top5"
+    ],
+    "a_top5": [
+     "mem_1792088552816998158",
+     "mem_1789136383983279469",
+     "ep_1785695216419714474",
+     "ep_1786336919812827426",
+     "ep_1786452727331645773"
+    ],
+    "b_top5": [
+     "mem_1792088552817004309",
+     "mem_1789136383983279469",
+     "ep_1785695216419714474",
+     "ep_1786336919812827426",
+     "ep_1786452727331645773"
+    ]
+   },
+   "局域网里访问 Gitea 家服连哪个 IP": {
+    "kind": "probe",
+    "changed": [
+     "top5"
+    ],
+    "a_top5": [
+     "mem_1792088552817004309",
+     "mem_1789136383983279469",
+     "ep_1786521524260755589",
+     "mem_1789150776299341327",
+     "mem_1789737118121867083"
+    ],
+    "b_top5": [
+     "mem_1792088552816998158",
+     "mem_1789136383983279469",
+     "ep_1786521524260755589",
+     "mem_1789150776299341327",
+     "mem_1789737118121867083"
+    ]
+   }
+  },
+  "v0_rank_shifts": []
+ },
+ "probe10": {
+  "rows": [
+   {
+    "query": "牧尘工作所在的县是哪个",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "淇县",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 1,
+    "v_pos_b": 1,
+    "v_pos_ids_a": [
+     "mem_1792088552819293505"
+    ],
+    "v_pos_ids_b": [
+     "mem_1792088552819293505"
+    ],
+    "v_rank_moved": []
+   },
+   {
+    "query": "招商项目台账",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "生物制造产业专班",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "小怡的工作笔记在哪",
+    "n_a": 15,
+    "n_b": 15,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "昨天做了什么",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "飞书怎么发语音",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "家服怎么连",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   },
+   {
+    "query": "织忆怎么部署",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 1,
+    "v_pos_b": 1,
+    "v_pos_ids_a": [
+     "mem_1791740271337572997"
+    ],
+    "v_pos_ids_b": [
+     "mem_1791740271337572997"
+    ],
+    "v_rank_moved": []
+   },
+   {
+    "query": "记忆保护名单",
+    "n_a": 20,
+    "n_b": 20,
+    "set_added": [],
+    "set_removed": [],
+    "order_diff_positions": 0,
+    "v_pos_a": 0,
+    "v_pos_b": 0,
+    "v_pos_ids_a": [],
+    "v_pos_ids_b": [],
+    "v_rank_moved": []
+   }
+  ],
+  "affected_queries": [],
+  "n_affected_queries": 0
+ },
+ "golden": {
+  "a": "14/14",
+  "b": "14/14"
+ }
+}

docs/ac-evidence/t_492e3f22-v-fuse-gate/raw/t492-DEF-golden-20261016-184003.json

@@ -0,0 +1,82 @@
+{
+ "at": "20261016-184343",
+ "tag": "DEF",
+ "base": "http://127.0.0.1:7892",
+ "keyword_pass": 14,
+ "keyword_fail": 0,
+ "keyword_total": 14,
+ "recall_at_5": 0.42857142857142855,
+ "mrr": 0.2380952380952381,
+ "items": [
+  {
+   "query": "织忆是什么",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆的图谱扩展机制是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "织忆如何通过 E1 提取和 MMR 实现多样去重",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "牧尘的偏好是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "牧尘喜欢什么样的工作方式",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "如何根据牧尘的偏好调整记忆检索策略",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "当前项目的技术栈是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "织忆项目有哪些核心组件",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆和其他记忆系统相比有什么架构优势",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "如何使用 recall 接口",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆有哪些管理工具",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "如何通过 API 扩展织忆功能",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "本地 llama 服务 8080 端口",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "session-to-zhiyi.py 脚本名",
+   "keyword_ok": true,
+   "id_recall": 1
+  }
+ ]
+}

docs/ac-evidence/t_492e3f22-v-fuse-gate/raw/t492-DEF-probe10-20261016-184003.json

@@ -0,0 +1,932 @@
+{
+ "at": "20261016-184350",
+ "tag": "DEF",
+ "base": "http://127.0.0.1:7892",
+ "top_k": 20,
+ "rows": [
+  {
+   "query": "牧尘工作所在的县是哪个",
+   "latency_ms": 689,
+   "top_ids": [
+    "mem_1792088552818864597",
+    "mem_1792144084503554069",
+    "mem_1792144084726385461",
+    "mem_1790349546889496890",
+    "mem_1787450890745924038",
+    "mem_1790448247125520424",
+    "mem_1789063252280657787",
+    "mem_1792088552817383356",
+    "ep_1786293372952742428",
+    "mem_1792144084989342595",
+    "mem_1789815923678472264",
+    "mem_1788006341389931001",
+    "mem_1790425119425241515",
+    "mem_1791556228356038757",
+    "mem_1789230281180128695",
+    "mem_1792088552824873585",
+    "mem_1792088552817930263",
+    "mem_1790424260127845812",
+    "mem_1789737493969591957",
+    "mem_1787926246493058856"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.5,
+    0.59,
+    0.59,
+    0.755,
+    0.255,
+    0.755,
+    0.755,
+    0.5,
+    0,
+    0.59,
+    0.83,
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.5,
+    0.5,
+    0.753,
+    0.755,
+    0.775
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "淇县",
+   "latency_ms": 717,
+   "top_ids": [
+    "mem_1792144084726385461",
+    "mem_1792088552818354175",
+    "mem_1792088552818868548",
+    "mem_1792088552818381759",
+    "mem_1792144084503554069",
+    "mem_1790938090098284542",
+    "mem_1790938089297531705",
+    "mem_1790603669933523968",
+    "mem_1792088552818672993",
+    "mem_1790425189400296970",
+    "mem_1791564884666490485",
+    "mem_1790848976254029727",
+    "mem_1792088552819307750",
+    "mem_1790657384251944657",
+    "mem_1789328667925309581",
+    "mem_1792116219795602288",
+    "mem_1790425213354467843",
+    "mem_1791395173327211388",
+    "mem_1792141689103478828",
+    "mem_1792088552819293505"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1
+   ],
+   "quality_scores": [
+    0.59,
+    0.5,
+    0.5,
+    0.5,
+    0.59,
+    0.83,
+    0.83,
+    0,
+    0.5,
+    0.755,
+    0,
+    0.755,
+    0.5,
+    0.755,
+    0.83,
+    0.63,
+    0.755,
+    0.675,
+    0.525,
+    0.99
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 1,
+   "v_pos_ids": [
+    "mem_1792088552819293505"
+   ]
+  },
+  {
+   "query": "招商项目台账",
+   "latency_ms": 751,
+   "top_ids": [
+    "mem_1790657711315786810",
+    "mem_1790425023223201421",
+    "mem_1792088552819227083",
+    "mem_1792088552817389514",
+    "mem_1791194371133842796",
+    "mem_1790565541250175254",
+    "mem_1791387362920806120",
+    "mem_1791899910885939521",
+    "mem_1789694252600441660",
+    "mem_1791665519266094329",
+    "mem_1790309303666562764",
+    "mem_1790581054400784704",
+    "mem_1792088552818016699",
+    "mem_1792088552818037737",
+    "mem_1790590528061314469",
+    "mem_1790187043969918230",
+    "mem_1790448247323881504",
+    "mem_1790329285912241710",
+    "mem_1792088552817184125",
+    "mem_1789327284892773149"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.755,
+    0.853,
+    0.5,
+    0.5,
+    0.755,
+    0.755,
+    0.765,
+    0.59,
+    0.755,
+    0,
+    0.755,
+    0.755,
+    0.5,
+    0.5,
+    0.755,
+    0.755,
+    0.755,
+    0.755,
+    0.5,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "生物制造产业专班",
+   "latency_ms": 722,
+   "top_ids": [
+    "mem_1792144084989342595",
+    "mem_1792107514588524675",
+    "mem_1791567307185364992",
+    "mem_1790051867711758989",
+    "mem_1789983973135821157",
+    "mem_1792088552816455607",
+    "mem_1791380468697720355",
+    "mem_1792088552817215072",
+    "mem_1790658098063956127",
+    "mem_1792088552825118896",
+    "mem_1790656401441522455",
+    "mem_1789755524758238574",
+    "mem_1789741519238314161",
+    "mem_1789736468534024216",
+    "mem_1789731934167097953",
+    "mem_1791916787313401423",
+    "mem_1787336595731500272",
+    "mem_1791794756620086535",
+    "mem_1791811143462384371",
+    "mem_1791735819718797684"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.59,
+    0.63,
+    0,
+    0.755,
+    0.755,
+    0.5,
+    0.78,
+    0.5,
+    0.755,
+    0.5,
+    0.755,
+    0.755,
+    0.755,
+    0.755,
+    0.798,
+    0.395,
+    0.695,
+    0.63,
+    0.315,
+    0.6
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "小怡的工作笔记在哪",
+   "latency_ms": 614,
+   "top_ids": [
+    "mem_1791365211038188118",
+    "mem_1792088552817459951",
+    "mem_1789155186005621317",
+    "mem_1790441160106001927",
+    "mem_1791736002933120377",
+    "mem_1792118701572499829",
+    "mem_1792088552818645685",
+    "mem_1791915848530338805",
+    "mem_1791383333319512288",
+    "mem_1792088552818535755",
+    "mem_1790187044046223493",
+    "mem_1792088552824837822",
+    "mem_1792088552817461960",
+    "mem_1791423103974079882",
+    "mem_1792088552825338800"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.8,
+    0.5,
+    0.755,
+    0.755,
+    0.39,
+    0.63,
+    0.5,
+    0.5,
+    0.8,
+    0.5,
+    0.755,
+    0.5,
+    0.5,
+    0.6,
+    0.5
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "昨天做了什么",
+   "latency_ms": 692,
+   "top_ids": [
+    "mem_1788341686813816985",
+    "mem_1790425119425241515",
+    "mem_1791633893565216844",
+    "mem_1790455525438011532",
+    "mem_1792083942223793654",
+    "mem_1792088770119950733",
+    "mem_1790805789157124891",
+    "mem_1792107513668237224",
+    "mem_1789431180512315186",
+    "mem_1790425238816494139",
+    "mem_1790424256264891954",
+    "mem_1787926246433496550",
+    "mem_1790425017614420145",
+    "mem_1790581068974704280",
+    "mem_1789327284892773149",
+    "mem_1790226520252986178",
+    "mem_1790269130144600786",
+    "mem_1791819411153164256",
+    "mem_1789827897771148953",
+    "mem_1788166192802637740"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.395,
+    0.63,
+    0.755,
+    0.63,
+    0.515,
+    0.755,
+    0.755,
+    0.775,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.395,
+    0.755,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "飞书怎么发语音",
+   "latency_ms": 783,
+   "top_ids": [
+    "mem_1790345721671326922",
+    "mem_1790350480105660984",
+    "mem_1790425022831993728",
+    "mem_1791118583544396904",
+    "mem_1791533226241026355",
+    "mem_1790455570122981543",
+    "mem_1791445601992763027",
+    "mem_1790496848751501955",
+    "mem_1789273123503223085",
+    "mem_1791666641349257940",
+    "mem_1791666631693960928",
+    "mem_1792088552825510526",
+    "mem_1791666639652621391",
+    "mem_1790265788637717626",
+    "mem_1790428068416795396",
+    "mem_1791118583617248844",
+    "mem_1791666711653684211",
+    "mem_1789557079608816710",
+    "mem_1789327693618007907",
+    "mem_1789273123339483168"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.803,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.803,
+    0,
+    0.83,
+    0.755,
+    0,
+    0.789,
+    0.5,
+    0,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.805,
+    0.755,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "家服怎么连",
+   "latency_ms": 700,
+   "top_ids": [
+    "mem_1791556229473565240",
+    "mem_1792084245285474959",
+    "mem_1792088552818289574",
+    "mem_1790180705032672204",
+    "mem_1792100307763093595",
+    "mem_1791665516356975992",
+    "mem_1792084245095017393",
+    "mem_1792082720405068478",
+    "mem_1791440639711180469",
+    "mem_1792114886051333715",
+    "mem_1792088552825127031",
+    "mem_1791642436962837771",
+    "mem_1792088552819438066",
+    "mem_1792088552816808997",
+    "mem_1790187091203506519",
+    "mem_1792126203805297968",
+    "mem_1792088552819539001",
+    "mem_1791904832104972267",
+    "mem_1792088552818662517",
+    "mem_1792088552824945417"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0,
+    0.63,
+    0.5,
+    0.755,
+    0.63,
+    0.8,
+    0.63,
+    0.63,
+    0,
+    0.63,
+    0.5,
+    0.65,
+    0.5,
+    0.5,
+    0.755,
+    0.395,
+    0.5,
+    0.43,
+    0.5,
+    0.5
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "织忆怎么部署",
+   "latency_ms": 724,
+   "top_ids": [
+    "mem_1788623666023332134",
+    "mem_1789793765940732350",
+    "mem_1791479003228249623",
+    "mem_1789149704491770048",
+    "mem_1789154058288011112",
+    "mem_1791818814047111008",
+    "mem_1792144557208382006",
+    "mem_1791402157932171759",
+    "mem_1791899478077693848",
+    "mem_1789411888019473186",
+    "mem_1792088552825130331",
+    "mem_1792088552817959556",
+    "mem_1791912042293867384",
+    "mem_1788657176712786530",
+    "mem_1790424244350350337",
+    "mem_1791740271337572997",
+    "mem_1792088552818065436",
+    "mem_1789132107404617336",
+    "mem_1788340352324421455",
+    "mem_1791431677583759030"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.911,
+    0.755,
+    0,
+    0.99,
+    0.695,
+    0.63,
+    0.565,
+    0.99,
+    0.644,
+    0.865,
+    0.5,
+    0.5,
+    0.59,
+    0.755,
+    0.99,
+    0,
+    0.5,
+    0.755,
+    0.755,
+    0.675
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 1,
+   "v_pos_ids": [
+    "mem_1791740271337572997"
+   ]
+  },
+  {
+   "query": "记忆保护名单",
+   "latency_ms": 708,
+   "top_ids": [
+    "mem_1790424365997011289",
+    "mem_1792126499467223243",
+    "mem_1788602693814224919",
+    "mem_1791725846574107784",
+    "mem_1789222925018439889",
+    "mem_1791666995981185275",
+    "mem_1792088552818471375",
+    "mem_1791686097363885894",
+    "mem_1789554588799449124",
+    "mem_1791420422726554280",
+    "mem_1791686088996261165",
+    "mem_1789760781378060592",
+    "mem_1790427668905832166",
+    "mem_1791111607517908194",
+    "mem_1791818212646773755",
+    "mem_1790051973384869549",
+    "mem_1790564623527240562",
+    "mem_1789992010025213786",
+    "mem_1792088552819583406",
+    "mem_1791685772880116219"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.99,
+    0.5,
+    0.285,
+    0,
+    0.755,
+    0,
+    0.5,
+    0,
+    0.755,
+    0.914629,
+    0,
+    0.755,
+    0.755,
+    0.755,
+    0.63,
+    0.755,
+    0.755,
+    0.99,
+    0.5,
+    0
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  }
+ ]
+}

docs/ac-evidence/t_492e3f22-v-fuse-gate/raw/t492-NOV-golden-20261016-184835.json

@@ -0,0 +1,82 @@
+{
+ "at": "20261016-185023",
+ "tag": "NOV",
+ "base": "http://127.0.0.1:7893",
+ "keyword_pass": 14,
+ "keyword_fail": 0,
+ "keyword_total": 14,
+ "recall_at_5": 0.42857142857142855,
+ "mrr": 0.2380952380952381,
+ "items": [
+  {
+   "query": "织忆是什么",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆的图谱扩展机制是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "织忆如何通过 E1 提取和 MMR 实现多样去重",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "牧尘的偏好是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "牧尘喜欢什么样的工作方式",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "如何根据牧尘的偏好调整记忆检索策略",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "当前项目的技术栈是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "织忆项目有哪些核心组件",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆和其他记忆系统相比有什么架构优势",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "如何使用 recall 接口",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆有哪些管理工具",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "如何通过 API 扩展织忆功能",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "本地 llama 服务 8080 端口",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "session-to-zhiyi.py 脚本名",
+   "keyword_ok": true,
+   "id_recall": 1
+  }
+ ]
+}

docs/ac-evidence/t_492e3f22-v-fuse-gate/raw/t492-NOV-probe10-20261016-184835.json

@@ -0,0 +1,932 @@
+{
+ "at": "20261016-185030",
+ "tag": "NOV",
+ "base": "http://127.0.0.1:7893",
+ "top_k": 20,
+ "rows": [
+  {
+   "query": "牧尘工作所在的县是哪个",
+   "latency_ms": 828,
+   "top_ids": [
+    "mem_1792088552818864597",
+    "mem_1792144084503554069",
+    "mem_1792144084726385461",
+    "mem_1790349546889496890",
+    "mem_1787450890745924038",
+    "mem_1790448247125520424",
+    "mem_1789063252280657787",
+    "mem_1792088552817383356",
+    "ep_1786293372952742428",
+    "mem_1792144084989342595",
+    "mem_1789815923678472264",
+    "mem_1788006341389931001",
+    "mem_1790425119425241515",
+    "mem_1791556228356038757",
+    "mem_1789230281180128695",
+    "mem_1792088552824873585",
+    "mem_1792088552817930263",
+    "mem_1790424260127845812",
+    "mem_1789737493969591957",
+    "mem_1787926246493058856"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.5,
+    0.59,
+    0.59,
+    0.755,
+    0.255,
+    0.755,
+    0.755,
+    0.5,
+    0,
+    0.59,
+    0.83,
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.5,
+    0.5,
+    0.753,
+    0.755,
+    0.775
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "淇县",
+   "latency_ms": 748,
+   "top_ids": [
+    "mem_1792144084726385461",
+    "mem_1792088552818354175",
+    "mem_1792088552818868548",
+    "mem_1792088552818381759",
+    "mem_1792144084503554069",
+    "mem_1790938090098284542",
+    "mem_1790938089297531705",
+    "mem_1790603669933523968",
+    "mem_1792088552818672993",
+    "mem_1790425189400296970",
+    "mem_1791564884666490485",
+    "mem_1790848976254029727",
+    "mem_1792088552819307750",
+    "mem_1790657384251944657",
+    "mem_1789328667925309581",
+    "mem_1792116219795602288",
+    "mem_1790425213354467843",
+    "mem_1791395173327211388",
+    "mem_1792141689103478828",
+    "mem_1792088552819293505"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1
+   ],
+   "quality_scores": [
+    0.59,
+    0.5,
+    0.5,
+    0.5,
+    0.59,
+    0.83,
+    0.83,
+    0,
+    0.5,
+    0.755,
+    0,
+    0.755,
+    0.5,
+    0.755,
+    0.83,
+    0.63,
+    0.755,
+    0.675,
+    0.525,
+    0.99
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 1,
+   "v_pos_ids": [
+    "mem_1792088552819293505"
+   ]
+  },
+  {
+   "query": "招商项目台账",
+   "latency_ms": 692,
+   "top_ids": [
+    "mem_1790657711315786810",
+    "mem_1790425023223201421",
+    "mem_1792088552819227083",
+    "mem_1792088552817389514",
+    "mem_1791194371133842796",
+    "mem_1790565541250175254",
+    "mem_1791387362920806120",
+    "mem_1791899910885939521",
+    "mem_1789694252600441660",
+    "mem_1791665519266094329",
+    "mem_1790309303666562764",
+    "mem_1790581054400784704",
+    "mem_1792088552818016699",
+    "mem_1792088552818037737",
+    "mem_1790590528061314469",
+    "mem_1790187043969918230",
+    "mem_1790448247323881504",
+    "mem_1790329285912241710",
+    "mem_1792088552817184125",
+    "mem_1789327284892773149"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.755,
+    0.853,
+    0.5,
+    0.5,
+    0.755,
+    0.755,
+    0.765,
+    0.59,
+    0.755,
+    0,
+    0.755,
+    0.755,
+    0.5,
+    0.5,
+    0.755,
+    0.755,
+    0.755,
+    0.755,
+    0.5,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "生物制造产业专班",
+   "latency_ms": 708,
+   "top_ids": [
+    "mem_1792144084989342595",
+    "mem_1792107514588524675",
+    "mem_1791567307185364992",
+    "mem_1790051867711758989",
+    "mem_1789983973135821157",
+    "mem_1792088552816455607",
+    "mem_1791380468697720355",
+    "mem_1792088552817215072",
+    "mem_1790658098063956127",
+    "mem_1792088552825118896",
+    "mem_1790656401441522455",
+    "mem_1789755524758238574",
+    "mem_1789741519238314161",
+    "mem_1789736468534024216",
+    "mem_1789731934167097953",
+    "mem_1791916787313401423",
+    "mem_1787336595731500272",
+    "mem_1791794756620086535",
+    "mem_1791811143462384371",
+    "mem_1791735819718797684"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.59,
+    0.63,
+    0,
+    0.755,
+    0.755,
+    0.5,
+    0.78,
+    0.5,
+    0.755,
+    0.5,
+    0.755,
+    0.755,
+    0.755,
+    0.755,
+    0.798,
+    0.395,
+    0.695,
+    0.63,
+    0.315,
+    0.6
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "小怡的工作笔记在哪",
+   "latency_ms": 736,
+   "top_ids": [
+    "mem_1791365211038188118",
+    "mem_1792088552817459951",
+    "mem_1789155186005621317",
+    "mem_1790441160106001927",
+    "mem_1791736002933120377",
+    "mem_1792118701572499829",
+    "mem_1792088552818645685",
+    "mem_1791915848530338805",
+    "mem_1791383333319512288",
+    "mem_1792088552818535755",
+    "mem_1790187044046223493",
+    "mem_1792088552824837822",
+    "mem_1792088552817461960",
+    "mem_1791423103974079882",
+    "mem_1792088552825338800"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.8,
+    0.5,
+    0.755,
+    0.755,
+    0.39,
+    0.63,
+    0.5,
+    0.5,
+    0.8,
+    0.5,
+    0.755,
+    0.5,
+    0.5,
+    0.6,
+    0.5
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "昨天做了什么",
+   "latency_ms": 791,
+   "top_ids": [
+    "mem_1788341686813816985",
+    "mem_1790425119425241515",
+    "mem_1791633893565216844",
+    "mem_1790455525438011532",
+    "mem_1792083942223793654",
+    "mem_1792088770119950733",
+    "mem_1790805789157124891",
+    "mem_1792107513668237224",
+    "mem_1789431180512315186",
+    "mem_1790425238816494139",
+    "mem_1790424256264891954",
+    "mem_1787926246433496550",
+    "mem_1790425017614420145",
+    "mem_1790581068974704280",
+    "mem_1789327284892773149",
+    "mem_1790226520252986178",
+    "mem_1790269130144600786",
+    "mem_1791819411153164256",
+    "mem_1789827897771148953",
+    "mem_1788166192802637740"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.395,
+    0.63,
+    0.755,
+    0.63,
+    0.515,
+    0.755,
+    0.755,
+    0.775,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.395,
+    0.755,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "飞书怎么发语音",
+   "latency_ms": 723,
+   "top_ids": [
+    "mem_1790345721671326922",
+    "mem_1790350480105660984",
+    "mem_1790425022831993728",
+    "mem_1791118583544396904",
+    "mem_1791533226241026355",
+    "mem_1790455570122981543",
+    "mem_1791445601992763027",
+    "mem_1790496848751501955",
+    "mem_1789273123503223085",
+    "mem_1791666641349257940",
+    "mem_1791666631693960928",
+    "mem_1792088552825510526",
+    "mem_1791666639652621391",
+    "mem_1790265788637717626",
+    "mem_1790428068416795396",
+    "mem_1791118583617248844",
+    "mem_1791666711653684211",
+    "mem_1789557079608816710",
+    "mem_1789327693618007907",
+    "mem_1789273123339483168"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.803,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.803,
+    0,
+    0.83,
+    0.755,
+    0,
+    0.789,
+    0.5,
+    0,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.805,
+    0.755,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "家服怎么连",
+   "latency_ms": 722,
+   "top_ids": [
+    "mem_1791556229473565240",
+    "mem_1792084245285474959",
+    "mem_1792088552818289574",
+    "mem_1790180705032672204",
+    "mem_1792100307763093595",
+    "mem_1791665516356975992",
+    "mem_1792084245095017393",
+    "mem_1792082720405068478",
+    "mem_1791440639711180469",
+    "mem_1792114886051333715",
+    "mem_1792088552825127031",
+    "mem_1791642436962837771",
+    "mem_1792088552819438066",
+    "mem_1792088552816808997",
+    "mem_1790187091203506519",
+    "mem_1792126203805297968",
+    "mem_1792088552819539001",
+    "mem_1791904832104972267",
+    "mem_1792088552818662517",
+    "mem_1792088552824945417"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0,
+    0.63,
+    0.5,
+    0.755,
+    0.63,
+    0.8,
+    0.63,
+    0.63,
+    0,
+    0.63,
+    0.5,
+    0.65,
+    0.5,
+    0.5,
+    0.755,
+    0.395,
+    0.5,
+    0.43,
+    0.5,
+    0.5
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "织忆怎么部署",
+   "latency_ms": 682,
+   "top_ids": [
+    "mem_1788623666023332134",
+    "mem_1789793765940732350",
+    "mem_1791479003228249623",
+    "mem_1789149704491770048",
+    "mem_1789154058288011112",
+    "mem_1791818814047111008",
+    "mem_1792144557208382006",
+    "mem_1791402157932171759",
+    "mem_1791899478077693848",
+    "mem_1789411888019473186",
+    "mem_1792088552825130331",
+    "mem_1792088552817959556",
+    "mem_1791912042293867384",
+    "mem_1788657176712786530",
+    "mem_1790424244350350337",
+    "mem_1791740271337572997",
+    "mem_1792088552818065436",
+    "mem_1789132107404617336",
+    "mem_1788340352324421455",
+    "mem_1791431677583759030"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.911,
+    0.755,
+    0,
+    0.99,
+    0.695,
+    0.63,
+    0.565,
+    0.99,
+    0.644,
+    0.865,
+    0.5,
+    0.5,
+    0.59,
+    0.755,
+    0.99,
+    0,
+    0.5,
+    0.755,
+    0.755,
+    0.675
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 1,
+   "v_pos_ids": [
+    "mem_1791740271337572997"
+   ]
+  },
+  {
+   "query": "记忆保护名单",
+   "latency_ms": 664,
+   "top_ids": [
+    "mem_1790424365997011289",
+    "mem_1792126499467223243",
+    "mem_1788602693814224919",
+    "mem_1791725846574107784",
+    "mem_1789222925018439889",
+    "mem_1791666995981185275",
+    "mem_1792088552818471375",
+    "mem_1791686097363885894",
+    "mem_1789554588799449124",
+    "mem_1791420422726554280",
+    "mem_1791686088996261165",
+    "mem_1789760781378060592",
+    "mem_1790427668905832166",
+    "mem_1791111607517908194",
+    "mem_1791818212646773755",
+    "mem_1790051973384869549",
+    "mem_1790564623527240562",
+    "mem_1789992010025213786",
+    "mem_1792088552819583406",
+    "mem_1791685772880116219"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.99,
+    0.5,
+    0.285,
+    0,
+    0.755,
+    0,
+    0.5,
+    0,
+    0.755,
+    0.914629,
+    0,
+    0.755,
+    0.755,
+    0.755,
+    0.63,
+    0.755,
+    0.755,
+    0.99,
+    0.5,
+    0
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  }
+ ]
+}

docs/ac-evidence/t_492e3f22-v-fuse-gate/raw/t492-PRE-golden-20261016-183449.json

@@ -0,0 +1,82 @@
+{
+ "at": "20261016-183633",
+ "tag": "PRE",
+ "base": "http://127.0.0.1:7891",
+ "keyword_pass": 14,
+ "keyword_fail": 0,
+ "keyword_total": 14,
+ "recall_at_5": 0.42857142857142855,
+ "mrr": 0.2380952380952381,
+ "items": [
+  {
+   "query": "织忆是什么",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆的图谱扩展机制是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "织忆如何通过 E1 提取和 MMR 实现多样去重",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "牧尘的偏好是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "牧尘喜欢什么样的工作方式",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "如何根据牧尘的偏好调整记忆检索策略",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "当前项目的技术栈是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "织忆项目有哪些核心组件",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆和其他记忆系统相比有什么架构优势",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "如何使用 recall 接口",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆有哪些管理工具",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "如何通过 API 扩展织忆功能",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "本地 llama 服务 8080 端口",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "session-to-zhiyi.py 脚本名",
+   "keyword_ok": true,
+   "id_recall": 1
+  }
+ ]
+}

docs/ac-evidence/t_492e3f22-v-fuse-gate/raw/t492-PRE-probe10-20261016-183449.json

@@ -0,0 +1,932 @@
+{
+ "at": "20261016-183639",
+ "tag": "PRE",
+ "base": "http://127.0.0.1:7891",
+ "top_k": 20,
+ "rows": [
+  {
+   "query": "牧尘工作所在的县是哪个",
+   "latency_ms": 670,
+   "top_ids": [
+    "mem_1792088552818864597",
+    "mem_1792144084503554069",
+    "mem_1792144084726385461",
+    "mem_1790349546889496890",
+    "mem_1787450890745924038",
+    "mem_1790448247125520424",
+    "mem_1789063252280657787",
+    "mem_1792088552817383356",
+    "ep_1786293372952742428",
+    "mem_1792144084989342595",
+    "mem_1789815923678472264",
+    "mem_1788006341389931001",
+    "mem_1790425119425241515",
+    "mem_1791556228356038757",
+    "mem_1789230281180128695",
+    "mem_1792088552824873585",
+    "mem_1792088552817930263",
+    "mem_1790424260127845812",
+    "mem_1789737493969591957",
+    "mem_1787926246493058856"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.5,
+    0.59,
+    0.59,
+    0.755,
+    0.255,
+    0.755,
+    0.755,
+    0.5,
+    0,
+    0.59,
+    0.83,
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.5,
+    0.5,
+    0.753,
+    0.755,
+    0.775
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "淇县",
+   "latency_ms": 699,
+   "top_ids": [
+    "mem_1792144084726385461",
+    "mem_1792088552818354175",
+    "mem_1792088552818868548",
+    "mem_1792088552818381759",
+    "mem_1792144084503554069",
+    "mem_1790938090098284542",
+    "mem_1790938089297531705",
+    "mem_1790603669933523968",
+    "mem_1792088552818672993",
+    "mem_1790425189400296970",
+    "mem_1791564884666490485",
+    "mem_1790848976254029727",
+    "mem_1792088552819307750",
+    "mem_1790657384251944657",
+    "mem_1789328667925309581",
+    "mem_1792116219795602288",
+    "mem_1790425213354467843",
+    "mem_1791395173327211388",
+    "mem_1792141689103478828",
+    "mem_1792088552819293505"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1
+   ],
+   "quality_scores": [
+    0.59,
+    0.5,
+    0.5,
+    0.5,
+    0.59,
+    0.83,
+    0.83,
+    0,
+    0.5,
+    0.755,
+    0,
+    0.755,
+    0.5,
+    0.755,
+    0.83,
+    0.63,
+    0.755,
+    0.675,
+    0.525,
+    0.99
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 1,
+   "v_pos_ids": [
+    "mem_1792088552819293505"
+   ]
+  },
+  {
+   "query": "招商项目台账",
+   "latency_ms": 682,
+   "top_ids": [
+    "mem_1790657711315786810",
+    "mem_1790425023223201421",
+    "mem_1792088552819227083",
+    "mem_1792088552817389514",
+    "mem_1791194371133842796",
+    "mem_1790565541250175254",
+    "mem_1791387362920806120",
+    "mem_1791899910885939521",
+    "mem_1789694252600441660",
+    "mem_1791665519266094329",
+    "mem_1790309303666562764",
+    "mem_1790581054400784704",
+    "mem_1792088552818016699",
+    "mem_1792088552818037737",
+    "mem_1790590528061314469",
+    "mem_1790187043969918230",
+    "mem_1790448247323881504",
+    "mem_1790329285912241710",
+    "mem_1792088552817184125",
+    "mem_1789327284892773149"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.755,
+    0.853,
+    0.5,
+    0.5,
+    0.755,
+    0.755,
+    0.765,
+    0.59,
+    0.755,
+    0,
+    0.755,
+    0.755,
+    0.5,
+    0.5,
+    0.755,
+    0.755,
+    0.755,
+    0.755,
+    0.5,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "生物制造产业专班",
+   "latency_ms": 675,
+   "top_ids": [
+    "mem_1792144084989342595",
+    "mem_1792107514588524675",
+    "mem_1791567307185364992",
+    "mem_1790051867711758989",
+    "mem_1789983973135821157",
+    "mem_1792088552816455607",
+    "mem_1791380468697720355",
+    "mem_1792088552817215072",
+    "mem_1790658098063956127",
+    "mem_1792088552825118896",
+    "mem_1790656401441522455",
+    "mem_1789755524758238574",
+    "mem_1789741519238314161",
+    "mem_1789736468534024216",
+    "mem_1789731934167097953",
+    "mem_1791916787313401423",
+    "mem_1787336595731500272",
+    "mem_1791794756620086535",
+    "mem_1791811143462384371",
+    "mem_1791735819718797684"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.59,
+    0.63,
+    0,
+    0.755,
+    0.755,
+    0.5,
+    0.78,
+    0.5,
+    0.755,
+    0.5,
+    0.755,
+    0.755,
+    0.755,
+    0.755,
+    0.798,
+    0.395,
+    0.695,
+    0.63,
+    0.315,
+    0.6
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "小怡的工作笔记在哪",
+   "latency_ms": 583,
+   "top_ids": [
+    "mem_1791365211038188118",
+    "mem_1792088552817459951",
+    "mem_1789155186005621317",
+    "mem_1790441160106001927",
+    "mem_1791736002933120377",
+    "mem_1792118701572499829",
+    "mem_1792088552818645685",
+    "mem_1791915848530338805",
+    "mem_1791383333319512288",
+    "mem_1792088552818535755",
+    "mem_1790187044046223493",
+    "mem_1792088552824837822",
+    "mem_1792088552817461960",
+    "mem_1791423103974079882",
+    "mem_1792088552825338800"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.8,
+    0.5,
+    0.755,
+    0.755,
+    0.39,
+    0.63,
+    0.5,
+    0.5,
+    0.8,
+    0.5,
+    0.755,
+    0.5,
+    0.5,
+    0.6,
+    0.5
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "昨天做了什么",
+   "latency_ms": 737,
+   "top_ids": [
+    "mem_1788341686813816985",
+    "mem_1790425119425241515",
+    "mem_1791633893565216844",
+    "mem_1790455525438011532",
+    "mem_1792083942223793654",
+    "mem_1792088770119950733",
+    "mem_1790805789157124891",
+    "mem_1792107513668237224",
+    "mem_1789431180512315186",
+    "mem_1790425238816494139",
+    "mem_1790424256264891954",
+    "mem_1787926246433496550",
+    "mem_1790425017614420145",
+    "mem_1790581068974704280",
+    "mem_1789327284892773149",
+    "mem_1790226520252986178",
+    "mem_1790269130144600786",
+    "mem_1791819411153164256",
+    "mem_1789827897771148953",
+    "mem_1788166192802637740"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.395,
+    0.63,
+    0.755,
+    0.63,
+    0.515,
+    0.755,
+    0.755,
+    0.775,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.395,
+    0.755,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "飞书怎么发语音",
+   "latency_ms": 710,
+   "top_ids": [
+    "mem_1790345721671326922",
+    "mem_1790350480105660984",
+    "mem_1790425022831993728",
+    "mem_1791118583544396904",
+    "mem_1791533226241026355",
+    "mem_1790455570122981543",
+    "mem_1791445601992763027",
+    "mem_1790496848751501955",
+    "mem_1789273123503223085",
+    "mem_1791666641349257940",
+    "mem_1791666631693960928",
+    "mem_1792088552825510526",
+    "mem_1791666639652621391",
+    "mem_1790265788637717626",
+    "mem_1790428068416795396",
+    "mem_1791118583617248844",
+    "mem_1791666711653684211",
+    "mem_1789557079608816710",
+    "mem_1789327693618007907",
+    "mem_1789273123339483168"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.803,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.803,
+    0,
+    0.83,
+    0.755,
+    0,
+    0.789,
+    0.5,
+    0,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.805,
+    0.755,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "家服怎么连",
+   "latency_ms": 718,
+   "top_ids": [
+    "mem_1791556229473565240",
+    "mem_1792084245285474959",
+    "mem_1792088552818289574",
+    "mem_1790180705032672204",
+    "mem_1792100307763093595",
+    "mem_1791665516356975992",
+    "mem_1792084245095017393",
+    "mem_1792082720405068478",
+    "mem_1791440639711180469",
+    "mem_1792114886051333715",
+    "mem_1792088552825127031",
+    "mem_1791642436962837771",
+    "mem_1792088552819438066",
+    "mem_1792088552816808997",
+    "mem_1790187091203506519",
+    "mem_1792126203805297968",
+    "mem_1792088552819539001",
+    "mem_1791904832104972267",
+    "mem_1792088552818662517",
+    "mem_1792088552824945417"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0,
+    0.63,
+    0.5,
+    0.755,
+    0.63,
+    0.8,
+    0.63,
+    0.63,
+    0,
+    0.63,
+    0.5,
+    0.65,
+    0.5,
+    0.5,
+    0.755,
+    0.395,
+    0.5,
+    0.43,
+    0.5,
+    0.5
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "织忆怎么部署",
+   "latency_ms": 732,
+   "top_ids": [
+    "mem_1788623666023332134",
+    "mem_1789793765940732350",
+    "mem_1791479003228249623",
+    "mem_1789149704491770048",
+    "mem_1789154058288011112",
+    "mem_1791818814047111008",
+    "mem_1792144557208382006",
+    "mem_1791402157932171759",
+    "mem_1791899478077693848",
+    "mem_1789411888019473186",
+    "mem_1792088552825130331",
+    "mem_1792088552817959556",
+    "mem_1791912042293867384",
+    "mem_1788657176712786530",
+    "mem_1790424244350350337",
+    "mem_1791740271337572997",
+    "mem_1792088552818065436",
+    "mem_1789132107404617336",
+    "mem_1788340352324421455",
+    "mem_1791431677583759030"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.911,
+    0.755,
+    0,
+    0.99,
+    0.695,
+    0.63,
+    0.565,
+    0.99,
+    0.644,
+    0.865,
+    0.5,
+    0.5,
+    0.59,
+    0.755,
+    0.99,
+    0,
+    0.5,
+    0.755,
+    0.755,
+    0.675
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 1,
+   "v_pos_ids": [
+    "mem_1791740271337572997"
+   ]
+  },
+  {
+   "query": "记忆保护名单",
+   "latency_ms": 694,
+   "top_ids": [
+    "mem_1790424365997011289",
+    "mem_1792126499467223243",
+    "mem_1788602693814224919",
+    "mem_1791725846574107784",
+    "mem_1789222925018439889",
+    "mem_1791666995981185275",
+    "mem_1792088552818471375",
+    "mem_1791686097363885894",
+    "mem_1789554588799449124",
+    "mem_1791420422726554280",
+    "mem_1791686088996261165",
+    "mem_1789760781378060592",
+    "mem_1790427668905832166",
+    "mem_1791111607517908194",
+    "mem_1791818212646773755",
+    "mem_1790051973384869549",
+    "mem_1790564623527240562",
+    "mem_1789992010025213786",
+    "mem_1792088552819583406",
+    "mem_1791685772880116219"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.99,
+    0.5,
+    0.285,
+    0,
+    0.755,
+    0,
+    0.5,
+    0,
+    0.755,
+    0.914629,
+    0,
+    0.755,
+    0.755,
+    0.755,
+    0.63,
+    0.755,
+    0.755,
+    0.99,
+    0.5,
+    0
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  }
+ ]
+}

docs/ac-evidence/t_492e3f22-v-fuse-gate/raw/t492-RBK-golden-20261016-190913.json

@@ -0,0 +1,82 @@
+{
+ "at": "20261016-191108",
+ "tag": "RBK",
+ "base": "http://127.0.0.1:7892",
+ "keyword_pass": 14,
+ "keyword_fail": 0,
+ "keyword_total": 14,
+ "recall_at_5": 0.42857142857142855,
+ "mrr": 0.2380952380952381,
+ "items": [
+  {
+   "query": "织忆是什么",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆的图谱扩展机制是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "织忆如何通过 E1 提取和 MMR 实现多样去重",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "牧尘的偏好是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "牧尘喜欢什么样的工作方式",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "如何根据牧尘的偏好调整记忆检索策略",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "当前项目的技术栈是什么",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "织忆项目有哪些核心组件",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆和其他记忆系统相比有什么架构优势",
+   "keyword_ok": true,
+   "id_recall": 1
+  },
+  {
+   "query": "如何使用 recall 接口",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "织忆有哪些管理工具",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "如何通过 API 扩展织忆功能",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "本地 llama 服务 8080 端口",
+   "keyword_ok": true,
+   "id_recall": 0
+  },
+  {
+   "query": "session-to-zhiyi.py 脚本名",
+   "keyword_ok": true,
+   "id_recall": 1
+  }
+ ]
+}

docs/ac-evidence/t_492e3f22-v-fuse-gate/raw/t492-RBK-probe10-20261016-190913.json

@@ -0,0 +1,932 @@
+{
+ "at": "20261016-191116",
+ "tag": "RBK",
+ "base": "http://127.0.0.1:7892",
+ "top_k": 20,
+ "rows": [
+  {
+   "query": "牧尘工作所在的县是哪个",
+   "latency_ms": 836,
+   "top_ids": [
+    "mem_1792088552818864597",
+    "mem_1792144084503554069",
+    "mem_1792144084726385461",
+    "mem_1790349546889496890",
+    "mem_1787450890745924038",
+    "mem_1790448247125520424",
+    "mem_1789063252280657787",
+    "mem_1792088552817383356",
+    "ep_1786293372952742428",
+    "mem_1792144084989342595",
+    "mem_1789815923678472264",
+    "mem_1788006341389931001",
+    "mem_1790425119425241515",
+    "mem_1791556228356038757",
+    "mem_1789230281180128695",
+    "mem_1792088552824873585",
+    "mem_1792088552817930263",
+    "mem_1790424260127845812",
+    "mem_1789737493969591957",
+    "mem_1787926246493058856"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.5,
+    0.59,
+    0.59,
+    0.755,
+    0.255,
+    0.755,
+    0.755,
+    0.5,
+    0,
+    0.59,
+    0.83,
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.5,
+    0.5,
+    0.753,
+    0.755,
+    0.775
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "淇县",
+   "latency_ms": 789,
+   "top_ids": [
+    "mem_1792144084726385461",
+    "mem_1792088552818354175",
+    "mem_1792088552818868548",
+    "mem_1792088552818381759",
+    "mem_1792144084503554069",
+    "mem_1790938090098284542",
+    "mem_1790938089297531705",
+    "mem_1790603669933523968",
+    "mem_1792088552818672993",
+    "mem_1790425189400296970",
+    "mem_1791564884666490485",
+    "mem_1790848976254029727",
+    "mem_1792088552819307750",
+    "mem_1790657384251944657",
+    "mem_1789328667925309581",
+    "mem_1792116219795602288",
+    "mem_1790425213354467843",
+    "mem_1791395173327211388",
+    "mem_1792141689103478828",
+    "mem_1792088552819293505"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1
+   ],
+   "quality_scores": [
+    0.59,
+    0.5,
+    0.5,
+    0.5,
+    0.59,
+    0.83,
+    0.83,
+    0,
+    0.5,
+    0.755,
+    0,
+    0.755,
+    0.5,
+    0.755,
+    0.83,
+    0.63,
+    0.755,
+    0.675,
+    0.525,
+    0.99
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 1,
+   "v_pos_ids": [
+    "mem_1792088552819293505"
+   ]
+  },
+  {
+   "query": "招商项目台账",
+   "latency_ms": 730,
+   "top_ids": [
+    "mem_1790657711315786810",
+    "mem_1790425023223201421",
+    "mem_1792088552819227083",
+    "mem_1792088552817389514",
+    "mem_1791194371133842796",
+    "mem_1790565541250175254",
+    "mem_1791387362920806120",
+    "mem_1791899910885939521",
+    "mem_1789694252600441660",
+    "mem_1791665519266094329",
+    "mem_1790309303666562764",
+    "mem_1790581054400784704",
+    "mem_1792088552818016699",
+    "mem_1792088552818037737",
+    "mem_1790590528061314469",
+    "mem_1790187043969918230",
+    "mem_1790448247323881504",
+    "mem_1790329285912241710",
+    "mem_1792088552817184125",
+    "mem_1789327284892773149"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.755,
+    0.853,
+    0.5,
+    0.5,
+    0.755,
+    0.755,
+    0.765,
+    0.59,
+    0.755,
+    0,
+    0.755,
+    0.755,
+    0.5,
+    0.5,
+    0.755,
+    0.755,
+    0.755,
+    0.755,
+    0.5,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "生物制造产业专班",
+   "latency_ms": 753,
+   "top_ids": [
+    "mem_1792144084989342595",
+    "mem_1792107514588524675",
+    "mem_1791567307185364992",
+    "mem_1790051867711758989",
+    "mem_1789983973135821157",
+    "mem_1792088552816455607",
+    "mem_1791380468697720355",
+    "mem_1792088552817215072",
+    "mem_1790658098063956127",
+    "mem_1792088552825118896",
+    "mem_1790656401441522455",
+    "mem_1789755524758238574",
+    "mem_1789741519238314161",
+    "mem_1789736468534024216",
+    "mem_1789731934167097953",
+    "mem_1791916787313401423",
+    "mem_1787336595731500272",
+    "mem_1791794756620086535",
+    "mem_1791811143462384371",
+    "mem_1791735819718797684"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.59,
+    0.63,
+    0,
+    0.755,
+    0.755,
+    0.5,
+    0.78,
+    0.5,
+    0.755,
+    0.5,
+    0.755,
+    0.755,
+    0.755,
+    0.755,
+    0.798,
+    0.395,
+    0.695,
+    0.63,
+    0.315,
+    0.6
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "小怡的工作笔记在哪",
+   "latency_ms": 653,
+   "top_ids": [
+    "mem_1791365211038188118",
+    "mem_1792088552817459951",
+    "mem_1789155186005621317",
+    "mem_1790441160106001927",
+    "mem_1791736002933120377",
+    "mem_1792118701572499829",
+    "mem_1792088552818645685",
+    "mem_1791915848530338805",
+    "mem_1791383333319512288",
+    "mem_1792088552818535755",
+    "mem_1790187044046223493",
+    "mem_1792088552824837822",
+    "mem_1792088552817461960",
+    "mem_1791423103974079882",
+    "mem_1792088552825338800"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.8,
+    0.5,
+    0.755,
+    0.755,
+    0.39,
+    0.63,
+    0.5,
+    0.5,
+    0.8,
+    0.5,
+    0.755,
+    0.5,
+    0.5,
+    0.6,
+    0.5
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "昨天做了什么",
+   "latency_ms": 739,
+   "top_ids": [
+    "mem_1788341686813816985",
+    "mem_1790425119425241515",
+    "mem_1791633893565216844",
+    "mem_1790455525438011532",
+    "mem_1792083942223793654",
+    "mem_1792088770119950733",
+    "mem_1790805789157124891",
+    "mem_1792107513668237224",
+    "mem_1789431180512315186",
+    "mem_1790425238816494139",
+    "mem_1790424256264891954",
+    "mem_1787926246433496550",
+    "mem_1790425017614420145",
+    "mem_1790581068974704280",
+    "mem_1789327284892773149",
+    "mem_1790226520252986178",
+    "mem_1790269130144600786",
+    "mem_1791819411153164256",
+    "mem_1789827897771148953",
+    "mem_1788166192802637740"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.395,
+    0.63,
+    0.755,
+    0.63,
+    0.515,
+    0.755,
+    0.755,
+    0.775,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.755,
+    0.395,
+    0.755,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "飞书怎么发语音",
+   "latency_ms": 810,
+   "top_ids": [
+    "mem_1790345721671326922",
+    "mem_1790350480105660984",
+    "mem_1790425022831993728",
+    "mem_1791118583544396904",
+    "mem_1791533226241026355",
+    "mem_1790455570122981543",
+    "mem_1791445601992763027",
+    "mem_1790496848751501955",
+    "mem_1789273123503223085",
+    "mem_1791666641349257940",
+    "mem_1791666631693960928",
+    "mem_1792088552825510526",
+    "mem_1791666639652621391",
+    "mem_1790265788637717626",
+    "mem_1790428068416795396",
+    "mem_1791118583617248844",
+    "mem_1791666711653684211",
+    "mem_1789557079608816710",
+    "mem_1789327693618007907",
+    "mem_1789273123339483168"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.803,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.803,
+    0,
+    0.83,
+    0.755,
+    0,
+    0.789,
+    0.5,
+    0,
+    0.755,
+    0.755,
+    0.755,
+    0,
+    0.805,
+    0.755,
+    0.755
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "家服怎么连",
+   "latency_ms": 787,
+   "top_ids": [
+    "mem_1791556229473565240",
+    "mem_1792084245285474959",
+    "mem_1792088552818289574",
+    "mem_1790180705032672204",
+    "mem_1792100307763093595",
+    "mem_1791665516356975992",
+    "mem_1792084245095017393",
+    "mem_1792082720405068478",
+    "mem_1791440639711180469",
+    "mem_1792114886051333715",
+    "mem_1792088552825127031",
+    "mem_1791642436962837771",
+    "mem_1792088552819438066",
+    "mem_1792088552816808997",
+    "mem_1790187091203506519",
+    "mem_1792126203805297968",
+    "mem_1792088552819539001",
+    "mem_1791904832104972267",
+    "mem_1792088552818662517",
+    "mem_1792088552824945417"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0,
+    0.63,
+    0.5,
+    0.755,
+    0.63,
+    0.8,
+    0.63,
+    0.63,
+    0,
+    0.63,
+    0.5,
+    0.65,
+    0.5,
+    0.5,
+    0.755,
+    0.395,
+    0.5,
+    0.43,
+    0.5,
+    0.5
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  },
+  {
+   "query": "织忆怎么部署",
+   "latency_ms": 936,
+   "top_ids": [
+    "mem_1788623666023332134",
+    "mem_1789793765940732350",
+    "mem_1791479003228249623",
+    "mem_1789149704491770048",
+    "mem_1789154058288011112",
+    "mem_1791818814047111008",
+    "mem_1792144557208382006",
+    "mem_1791402157932171759",
+    "mem_1791899478077693848",
+    "mem_1789411888019473186",
+    "mem_1792088552825130331",
+    "mem_1792088552817959556",
+    "mem_1791912042293867384",
+    "mem_1788657176712786530",
+    "mem_1790424244350350337",
+    "mem_1791740271337572997",
+    "mem_1792088552818065436",
+    "mem_1789132107404617336",
+    "mem_1788340352324421455",
+    "mem_1791431677583759030"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    1,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.911,
+    0.755,
+    0,
+    0.99,
+    0.695,
+    0.63,
+    0.565,
+    0.99,
+    0.644,
+    0.865,
+    0.5,
+    0.5,
+    0.59,
+    0.755,
+    0.99,
+    0,
+    0.5,
+    0.755,
+    0.755,
+    0.675
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 1,
+   "v_pos_ids": [
+    "mem_1791740271337572997"
+   ]
+  },
+  {
+   "query": "记忆保护名单",
+   "latency_ms": 968,
+   "top_ids": [
+    "mem_1790424365997011289",
+    "mem_1792126499467223243",
+    "mem_1788602693814224919",
+    "mem_1791725846574107784",
+    "mem_1789222925018439889",
+    "mem_1791666995981185275",
+    "mem_1792088552818471375",
+    "mem_1791686097363885894",
+    "mem_1789554588799449124",
+    "mem_1791420422726554280",
+    "mem_1791686088996261165",
+    "mem_1789760781378060592",
+    "mem_1790427668905832166",
+    "mem_1791111607517908194",
+    "mem_1791818212646773755",
+    "mem_1790051973384869549",
+    "mem_1790564623527240562",
+    "mem_1789992010025213786",
+    "mem_1792088552819583406",
+    "mem_1791685772880116219"
+   ],
+   "v_values": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "quality_scores": [
+    0.99,
+    0.5,
+    0.285,
+    0,
+    0.755,
+    0,
+    0.5,
+    0,
+    0.755,
+    0.914629,
+    0,
+    0.755,
+    0.755,
+    0.755,
+    0.63,
+    0.755,
+    0.755,
+    0.99,
+    0.5,
+    0
+   ],
+   "raw_quality_scores": [
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0,
+    0
+   ],
+   "v_pos": 0,
+   "v_pos_ids": []
+  }
+ ]
+}

docs/ac-evidence/t_5ad3f843-metanote-chokepoint/DELIVERY.md

@@ -0,0 +1,116 @@
+# 交付说明 · t_5ad3f843【织忆·主症B 返修】元笔记分类器 + 咽喉点收口
+
+- **卡号**：`t_5ad3f843` ｜ **执行者**：prof-dev（2 次 run 均 `rc=1` 崩于"写交付说明"步骤）
+- **终审收口**：小怡（default）—— 逐条**亲跑**验证，不采信 worker 自报
+- **日期**：2026-10-16 ｜ **commit**：`e85c9db`（代码）+ 本文件（证据）
+- **结论**：AC1–AC8 **全绿**；**生产未部署**（源码级交付）⇒ **可申请激活**（需另提卡 + 牧尘放行）
+
+---
+
+## ① 判据脚本原始输出（AC1 · exit 0）
+
+```
+$ python3 ~/.hermes/scripts/verify-meta-filter.py
+正例 14 条 / 负例 11 条；不符 0 条
+
+[PASS] 分类器达标
+$ echo $?
+0
+```
+
+> 基线（返修前）：exit=1，**10 条不符（7 漏拦 + 3 误拦）**。
+
+## ② 逐条 before / after（AC7 要求）
+
+### 病根一 · 漏拦（7 条）：判据只认"写下来的那几种写法"
+
+| 案例（生产原文） | before | after | 命中规则 |
+|---|---|---|---|
+| `实测 淇县 ANN 731→15、鹤壁 6536→18（都进池50）` | **放行 ✗** | 拦 ✓ | `pool_in+rank_num+num_arrow+pool` |
+| `实测 淇县 ANN 731→15（都进池50）` | **放行 ✗** | 拦 ✓ | `pool_in+rank_num+num_arrow+pool` |
+| `短问法 MISS 复现：牧尘工作所在的县是哪个` | 放行 ✗ | 拦 ✓ | `repro` |
+| `淇县库里 21 条、鹤壁 10 条、河南 15 条语料可用。` | 放行 ✗ | 拦 ✓ | `data_inv` |
+| `实测（只读，未改任何库/代码）：目标「用户在招商局工作」在 query…` | 放行 ✗ | 拦 ✓ | `data_inv+pool+rank` |
+| `memory_write 落 episodes 表 + vector 全 0 —— 写入…` | 放行 ✗ | 拦 ✓ | `write_bug` |
+| `织忆 golden 14/14 不退化，recall@5 = 0.4286` | 放行 ✗ | 拦 ✓ | `eval_metric` |
+
+**新增规则原文**（形态匹配，非加字面词 —— 满足 AC3）：
+```
+_RANK_NUM   = \bANN\s*\d+\s*→\s*\d+
+_NUM_ARROW  = \d+\s*→\s*\d+
+_POOL_IN    = 进\s*池\s*\d+
+_REPRO      = (MISS\s*复现|复现[:：（(]|复现\s*(步骤|方法|过程|流程|记录)\s*[:：（(])
+_DATA_INV   = （语料盘点：N 条…可用）
+_EVAL_METRIC= (golden \d+/\d+|recall@\d+)
+_DEBUG_MARK = (\bDEBUG\b|探针|自报|脚本输出)
+_WRITE_BUG  = （写入类故障记录）
+```
+
+### 病根二 · 误拦（3 条）：把"关于系统的事实"当调查记录拦了
+
+| 案例（事实句，金标同族题面） | before | after | 理由 |
+|---|---|---|---|
+| `织忆召回接口 POST /api/v1/recall 接受 top_k 参数。` | **拦 ✗** | 放 ✓ | `_FACT_DECL` 事实声明形豁免 |
+| `召回缓存 recall-cache 每 6 小时刷新一次。` | **拦 ✗** | 放 ✓ | `_FACT_DECL`（周期形） |
+| `织忆主服务进程名是 zhiyid-new，合并服务是 zhiyi-consolidate。` | **拦 ✗** | 放 ✓ | `_FACT_DECL`（进程名形） |
+
+**AC4 区分判据**：`svc_core` 由 STRONG 降级为 **cue 计数**；通用功能词（召回|蒸馏）各计 1 个 cue，需叠加才判；**事实声明形**（`_FACT_DECL`：接受X参数/监听N端口/每N小时…一次/进程名是/地址是/模型是/字体/行距）⇒ **判事实放行**；带 实测/复现/诊断/命中/命名空间/限流/退避/碎片化/FALLBACK 等概念词的调查记录**不受豁免** ⇒ 既放行 3 条误拦，又**不漏原有 16 条自检正例**（两套用例同时绿）。
+
+### 负例 11 条全放行（`[放✓]`）
+牧尘在河南省鹤壁市淇县工作。· 淇县招商项目台账（09-27阅读版）.md 已发布。· 织忆服务的健康检查接口监听 7821 端口。· 织忆召回接口 POST /api/v1/recall 接受 top_k 参数。· 召回缓存 recall-cache 每 6 小时刷新一次。· 织忆主服务进程名是 zhiyid-new，合并服务是 zhiyi-consolidate · 公文排版正文用仿宋_GB2312 字体，行距 28 磅。· 家庭服务器（跑 Gitea）在局域网里的地址是 192.168.123.11。· 本地 2B 模型服务监听 8080 端口。· 牧尘负责生物制造产业专班 + 卫都街道安全管理。· 小怡日常主链模型是 deepseek-v4-flash（官方直连）。
+
+## ③ AC8 咽喉点收口（本卡新增，最关键）
+
+**问题**：B 卡只接了 3 条管线，而污染样本 **5/6 来自未覆盖路径**（27 个写者只有 3 个被覆盖）。
+
+**修法**：接在**唯一共享入口**上 —— 所有 27 个写者都走 `POST /api/v1/commit`：
+
+```go
+// go/internal/api/routes/core.go:108（/commit 处理器，AutoDistillTrigger 之前）
+if skip, reason := metanote.ShouldSkip(req.Content); skip {
+    log.Printf("[metanote] commit rejected (gate ON, meta-note): ...")
+    respondError(w, 422, "meta-note rejected: "+reason)
+    return          // ← 真拒写：不落 episode、不进蒸馏
+}
+// core.go:322（write_memory 处理器）同样接线
+```
+⇒ 覆盖 **27/27 写者**（对比修前 3/27）。
+
+**Python↔Go 同源锁死**：`internal/metanote`（patterns.go / classify.go）与 `~/.hermes/scripts/lib/meta_note_filter.py` 口径同源，由 parity + 差分测试锁死：
+- `TestParity_PythonLive`（Live parity，0.03s）
+- `TestParity_MutatedCopyFails` —— **变异必 FAIL**（负控，0.03s）
+- `TestDifferential_PythonReference` / `TestDifferential_ExpectedSpotChecks`
+
+## ④ 小怡终审亲跑证据（非自报）
+
+| 项 | 命令 | 结果 |
+|---|---|---|
+| 构建 | `go build ./...` | **exit 0** |
+| 测试 | `go test ./internal/metanote/` | **ok 0.130s**（8 项含 parity/mutation） |
+| 判据 | `verify-meta-filter.py` | **exit 0**（正14/负11/不符0） |
+| 判据未被改（AC2） | `verify-meta-filter.py` mtime **20:08:22** 早于 worker 首启 **20:09:36** | ✓ 正/负例逐字未动 |
+| 门默认关（AC6） | gateway env / zhiyid env | **各 0 处** `ZHIYI_META_FILTER` |
+| **未部署** | 生产二进制 mtime `19:02:50` **早于** 提交 `21:59:28`；md5 `95286f63789c7ecd` 未变；`strings` 内 `metanote` **0 处** | ✓ 源码级 |
+| 零行为变化 | `classify.go:91 ShouldSkip` → `if !GateEnabled() { return false, "" }` | ✓ 门关即不参与 |
+
+## ⑤ 改动文件清单
+
+| 文件 | 说明 |
+|---|---|
+| `go/internal/metanote/patterns.go` | 新增（规则表，Python 口径同源） |
+| `go/internal/metanote/classify.go` | 新增（分类 + `GateEnabled`/`ShouldSkip`） |
+| `go/internal/metanote/metanote_parity_test.go` | 新增（parity） |
+| `go/internal/metanote/metanote_live_test.go` | 新增（Live parity + 变异负控） |
+| `go/internal/metanote/metanote_diff_test.go` | 新增（差分对照） |
+| `go/internal/api/routes/core.go` | 咽喉点接线（:108 / :322） |
+| `~/.hermes/scripts/lib/meta_note_filter.py` | 分类器 193 → 326 行（仓库外，`.hermes` 侧） |
+
+## ⑥ 激活状态
+
+**可申请激活**（分类器绿 + 咽喉点覆盖 27/27 + parity/变异测试在 + 默认关 + 未部署）。
+**但激活不在本卡范围**：需 ① 另提卡 ② 重建二进制并部署 ③ 牧尘放行（`prod-write-gate.py` 会拦，别绕）。
+
+## ⑦ 遗留（诚实记录）
+
+1. **两次 run 均 `rc=1` 崩于"写交付说明"步骤**，日志**无任何错误行**（裸退）；代码工作在两轮内已完成并提交 ⇒ 未放行第三次盲跑。**真因未定**（疑与同 profile 并发 worker / 会话结束路径有关，待查）。
+2. **AC8.1 覆盖表**：本文件 ③ 已给出结构性论证（单一入口 ⇒ 27/27），worker 原计划的逐写者分布表未落盘（随交付说明一起丢）。

docs/hit-outcome-signal.md

@@ -0,0 +1,89 @@
+# hits+/hits− outcome 信号定义（GA-P0-4 · t_bb10e8bc）
+
+> 一句话：hits+/hits− 是**独立、语义单一**的「真实 outcome」信号，供 P2 自适应阈值做
+> 有意义的拟合。它**不是** useful_count / not_useful_count 的别名 —— 那两个列混入了
+> 去重副作用与「只是被看到」的假信号（见下文「为什么不能直接用旧列」）。
+
+## 1. 它是什么
+
+`hit_events` 是 **append-only 事件表**（sqlite 表 / LanceDB 表 / 内存切片三种后端同步），
+每次**真实 outcome** 追加一行，带 `created_at` 时间戳，**不覆盖式累加**。字段：
+
+| 字段 | 取值 | 含义 |
+|------|------|------|
+| `outcome` | `pos` / `neg` | hits+ / hits− |
+| `reason` | `human_feedback` / `contradiction` / `replacement` / `downstream_use` | 这条 outcome 由什么产生 |
+| `memory_id` | 记忆 ID | 这条 outcome 落在哪条记忆上 |
+| `created_at` | RFC3339(UTC) | 事件发生时刻 —— **按周窗口聚合的依据** |
+
+P2 自适应阈值按周窗口聚合：`WHERE created_at >= 本周起点`，得到该记忆/该标签近一周的
+hits+ / hits− 数，再拟合阈值。因为事件带时间戳且 append-only，周窗口天然可算（老窗口
+自然滑出，不污染当前拟合）。
+
+## 2. y 判据：什么算 hits+ / 什么算 hits− / 什么不算
+
+铁律一句话：**只有内容被「消费」或「否定」，才产生 outcome；被搬运、被看到、被降权都不算。**
+
+### hits+（`outcome="pos"`）—— 被真实使用/认可
+1. **human_feedback**：用户对召回结果点「有用」（`/feedback/useful`）、或「修正」（`/feedback/correct`）
+2. **downstream_use**：下游流程**实际引用**了这条记忆的内容（写代码/写文档/决策时真正用上）
+3. **contradiction（正面）**：后续对话中它被反复印证、未被推翻
+
+### hits−（`outcome="neg"`）—— 被真实否定/取代
+1. **human_feedback**：用户对召回结果点「没用」（`/feedback/not-useful`）
+2. **contradiction**：后续对话中它与更可信的信息**冲突并被淘汰**（`delete_reason=contradiction`）
+3. **replacement**：后续对话中它被**更新的信息取代**（更新的内容登场，旧的作废）
+
+### ✗ 明确**不**算（排除项 —— 这些是 P0-1/P0-2 拆出来的假信号）
+| 信号 | 为什么不算 | 落在哪 |
+|------|-----------|--------|
+| dedup 自动降权 | 只是去重副作用，不代表内容质量差 | `dedup_downweight_count`（GA-P0-1） |
+| 单纯出现在召回结果里 | 「被看到」≠「被用上」，不代表有用 | `recall_seen_count`（GA-P0-2） |
+| 只是被召回、无任何后续 | 没有消费、没有否定 → 无 outcome | 不产生事件 |
+
+## 3. 为什么不能直接用旧列（等同不成立的证据）
+
+方案文档（`已办.md:16623`）曾写 `hits+ = useful_count`、`hits− = not_useful_count`、
+`V = hits+ + hits−`。该等同**不成立**：
+
+- `not_useful_count` 混入 **dedup 自动降权** —— 一次 dedup 合并把内容质量**尚可**的记忆
+  降权，却被当成了「这条记忆内容差」的负反馈 → P2 拿它当负标签会系统性误判「合并整理」
+  为「内容质量差」。
+- `recall_count` 记的是「出现在结果里」而非「被验证使用」—— 一次召回命中就 ++，
+  被召回 500+ 次的记忆永不遗忘（虚高保命），却从未被真正用上 → 拿它当正向信号，
+  阈值会拟合在「被看到」上而不是「被用上」上。
+
+所以必须新建**只由真实 outcome 写入**的独立信号。本表就是它。
+
+## 4. 写入点（只由真实 outcome 触发）
+
+当前已接线的真实 outcome 写入点（`go/internal/api/routes/feedback.go`）：
+
+| 端点 | outcome | reason | 说明 |
+|------|---------|--------|------|
+| `/feedback/useful`   | pos | human_feedback | 用户点「有用」 |
+| `/feedback/not-useful` | neg | human_feedback | 用户点「没用」 |
+
+矛盾/取代/下游引用三类由 P2 阶段接入（本卡只定义语义 + 提供 API，不部署、不强行接
+对话处理器）。
+
+## 5. 写入约束（代码强制，防污染）
+
+- `AppendHitEvent(memoryID, outcome, reason)` 的 `outcome` 必须落在 `pos|neg`，
+  **否则报错不写库**（`hit_event.go:NormalizeHitOutcome`）—— 非法/混合语义值在源头被拦。
+- `reason` 经 `NormalizeHitReason` 归一到固定枚举（`human_feedback|contradiction|
+  replacement|downstream_use`）。
+- dedup 降权 / 单纯召回出现**禁止**调用本方法（它们在 P0-1/P0-2 已各有专属字段）。
+
+## 6. 实现位置
+
+| 层 | 文件 |
+|----|------|
+| Go 语义定义（枚举+归一化，单一真源） | `go/internal/storage/hit_event.go` |
+| Go 模型 | `go/internal/models/memory.go`（`HitEvent`） |
+| sqlite 后端 | `go/internal/storage/sqlite.go`（`hit_events` 表 + 3 方法） |
+| 内存后端 | `go/internal/storage/memvector.go`（`hitEvents` 切片 + 3 方法） |
+| Rust-LanceDB 后端 | `rust/src/lancedb_ops.rs`（`hit_events` 表 + 3 方法）· `rust/src/main.rs`（IPC） |
+| 真实 outcome 接线 | `go/internal/api/routes/feedback.go` |
+
+**禁部署**：本卡只落代码 + 测试 + 文档，不更新运行中的二进制、不碰 live 库。

docs/plans/2026-10-06-distill-quality-pipeline-fix.md

@@ -0,0 +1,276 @@
+# 织忆蒸馏质量链路修复方案
+
+> **面向实现者**：每任务 2~5 分钟粒度，含精确路径、行号、完整代码、验证命令与期望输出。
+> **日期**：2026-10-06 ｜ **状态**：待拍板 ｜ **拟定**：小怡
+
+---
+
+## 一、目标
+
+切断这条因果链（本次排查发现的真实病灶）：
+
+```
+评分闸门失效（prompt 示例被模型原样 echo）
+        ↓
+只剩「文本长度」一道闸门在扛
+        ↓
+提取器为过闸门 → 给短规则追加固定话术（enrich）
+        ↓
+伪造内容进入长期记忆库（已污染 6 条，含跨技能串味）
+```
+
+并顺带修掉 2 个次生 bug。
+
+**注意**：上一轮我曾建议「蒸馏入口拒收规则块」—— **该建议作废**。证据 E4 表明「规则→episode→蒸馏」是**有意设计且工作正常**的流水线，拒收会砸掉整条技能规则入库链路。
+
+## 二、架构
+
+三处独立改动，互不阻塞：
+
+| 层 | 文件 | 改什么 |
+|---|---|---|
+| 蒸馏（Go） | `go/internal/distill/engine.go` | 去 echo 源 + 加 echo 检测 + 影子模式 |
+| 提取（Python） | `~/.hermes/scripts/skill_rule_extractor_v2.py` | 停用伪造式扩写 |
+| 遗忘（Go） | `go/internal/api/routes/admin.go` · `go/internal/storage/lancedb_ipc.go` · `go/internal/api/server.go` | 空 id 防御 + 计数修正 |
+
+## 三、证据链（为什么这么改）
+
+| # | 事实 | 证据（可复现） |
+|---|---|---|
+| E1 | 评分是 prompt 示例复读 | `engine.go:534` 写死示例 `"is":0.8,"su":0.7,"pa":0.6,"vd":0.9,"ru":0.7`。实测：换内容分数不变；把示例数字改成 0.1~0.5，输出跟着变成 0.3~0.7 |
+| E2 | 评分门槛永不触发 | `engine.go:417` 判据 `overall<0.7 && score.VD<0.8`；echo 值按 `engine.go:89` 权重算得 `overall=0.755`，恒 ≥0.7 |
+| E3 | 扩写话术已入记忆库 | `memories` 表命中 **6 条**，含 3 条纯话术本身、1 条跨技能串味（`hermes-soul-maintenance｜修改纪律：…rclone lsf`） |
+| E4 | 流水线本身是好的 | 台账按规则指纹去重：**入库 1539 / 被丢 0 / 在途 60 = 100% 收敛**（`~/.hermes/skill_rules_ingest_ledger.jsonl`） |
+| E5 | 空 id 根因在数据层 | `memories` 表有 **2 行** id/content/source 全空 ⇒ 命中「<30 字碎片快速道」⇒ `SoftDelete("")` ⇒ Rust `rust/src/main.rs:331` 拒绝 |
+
+## 四、施工铁律
+
+1. **不许 python 直连写 LanceDB**：zhiyid 运行期直连写 = 双写源（与 state.db 同类事故）。**所有删除走 HTTP API**。
+2. **P0-1/P0-2 必须影子模式先行**：修好 echo 会**激活一个此前从不触发的闸门**，可能造成批量丢弃。先 log-only 观察，再决定是否真拦。
+3. **改前备份、改后复验**：Go 改动过 `go vet` + 全量测试；Python 改动跑断言。
+
+---
+
+## 五、任务清单
+
+### P0-3 停用「伪造式扩写」〔建议先做，止血〕
+
+**文件**：`/home/muc/.hermes/scripts/skill_rule_extractor_v2.py:407-409`
+
+**Step 1** 删掉追加固定话术的分支
+
+旧：
+```python
+    if len(r["rule"]) < 45:
+        body += (" ｜执行时须落到具体命令与路径（如 hermes kanban list、pgrep -f、"
+                 "~/.hermes/scripts/、rclone lsf），并在汇报里贴出原始输出片段作为证据。")
+```
+
+新：
+```python
+    # 🔴 禁止追加与规则无关的固定话术（2026-10-06）：那会伪造出技能从未声明过的"事实"，
+    #    已污染 memories 6 条（含跨技能串味）。短规则宁可被丢，也不许编。
+    #    真实扩写只允许用规则自带字段（why/anti/scope，已在上方拼入）。
+```
+
+**Step 2** 验证（必须证明它不再产出那句话）
+```bash
+python3 -c "
+import importlib.util,pathlib
+spec=importlib.util.spec_from_file_location('e', pathlib.Path.home()/'.hermes/scripts/skill_rule_extractor_v2.py')
+m=importlib.util.module_from_spec(spec); spec.loader.exec_module(m)
+out=m.enrich({'skill':'x','section':'y','rule':'短规则','why':'因为所以','anti':'','scope':''})
+print(out)
+assert '执行时须落到具体命令与路径' not in out, '❌ 伪造话术仍在'
+print('✅ 已停用伪造扩写')
+"
+```
+期望：打印 `✅ 已停用伪造扩写`。
+
+**Step 3** 记录代价：改完重跑一轮 `skill_rules_sync.py`，把台账 `landed` 变化作为**代价数据**上报（短规则入库率会下降，这是可接受的真话）。
+
+---
+
+### P0-4 清理已污染的 6 条记忆〔建议先做，止血〕
+
+**🔴 走 HTTP API，不许直连写库。**
+
+**Step 1** 取 id（已取得，见下）→ 先 dry-run 看影响面
+**Step 2** 逐个软删，`reason` 用 `polluted_expansion_20261006`
+
+```
+mem_1791262758250672496  收到牧尘任务…｜依据…｜适用…｜执行时须落到具体命令与路径（…）
+mem_1791261603417879613  适用：cron 连续失败诊断 ｜执行时须落到…
+mem_1791267692697346820  hermes-soul-maintenance｜修改纪律：执行时须落到…   ← 跨技能串味
+mem_1791215971324172233  执行时须落到具体命令与路径（…）                  ← 纯话术
+mem_1791215338803223776  执行时须落到具体命令与路径（…）                  ← 纯话术
+mem_1791270196923533388  使用 --move-slide 时 ｜执行时须落到…              ← 无关规则被串味
+```
+
+**Step 3** 复验：同脚本再查一次，`执行时须落到具体命令与路径` 命中 **0**。
+
+---
+
+### P0-1 去掉评分 prompt 的 echo 源
+
+**文件**：`go/internal/distill/engine.go:533-536`
+
+**Step 1** 把「填好的示例」换成「类型骨架 + 显式禁令」
+
+旧：
+```
+输出格式（严格JSON）:
+{"facts": ["独立事实1", "独立事实2"], "decisions": ["决定1"], "conclusions": ["结论1"], "actions_taken": ["行动1"], "open_questions": ["问题1"], "entities": ["entity1", "entity2"], "is": 0.8, "su": 0.7, "pa": 0.6, "vd": 0.9, "ru": 0.7}
+
+评分说明：is/su/pa/vd/ru 是 0.0 到 1.0 之间的浮点数（越高越好），不要用 0-10 整数。
+```
+
+新：
+```
+输出格式（严格JSON）:
+{"facts": ["..."], "decisions": ["..."], "conclusions": ["..."], "actions_taken": ["..."], "open_questions": ["..."], "entities": ["..."], "is": <小数>, "su": <小数>, "pa": <小数>, "vd": <小数>, "ru": <小数>}
+
+评分说明：is/su/pa/vd/ru 必须是 0.0~1.0 之间的浮点数（越高越好），不得用 0-10 整数。
+🔴 上例中的 <小数> 是类型占位符，不是数值。必须依据**本次内容**独立打分；沿用任何示例数值都算错误答案。
+```
+
+**Step 2** 验证
+```bash
+cd /home/muc/src/memoryweave/go && go vet ./internal/distill/ && go test ./internal/distill/ -count=1
+```
+期望：全 PASS。
+
+**Step 3** liveprobe 实证（沿用 `quality_liveprobe_test.go` 的手法）
+```bash
+go test ./internal/distill/ -tags liveprobe -run TestLiveScoreNotEcho -v -count=1
+```
+期望：6 条不同内容 → **分数不再全部相同**（≥3 种不同向量）。
+
+---
+
+### P0-2 加「echo 检测 + 影子模式」
+
+**Step 1** 新增检测（`engine.go`，用现成评分结构，见 `engine.go:65-80`）
+
+```go
+// echoScoreVectors 是 prompt 历史版本里出现过的示例评分向量。
+// 模型输出与任一向量完全相同 ⇒ 判定为「照抄示例」，评分不可信。
+var echoScoreVectors = [][5]float64{
+	{0.8, 0.7, 0.6, 0.9, 0.7},
+}
+
+// IsEchoScore 判断评分是否在照抄 prompt 示例。
+func IsEchoScore(s Score5D) bool {
+	for _, v := range echoScoreVectors {
+		if s.IS == v[0] && s.SU == v[1] && s.PA == v[2] && s.VD == v[3] && s.RU == v[4] {
+			return true
+		}
+	}
+	return false
+}
+```
+
+**Step 2** 影子开关（默认开=只记录不拦）
+
+```go
+// 默认影子模式：只记录「本该被丢」的样本，不真的丢。
+// 修好 echo 会激活一个此前从不触发的闸门，必须先观察再开拦。
+var scoreShadowMode = os.Getenv("ZHIYI_SCORE_SHADOW") != "0"
+```
+
+在 `engine.go:417` 判据处：
+```go
+	if overall < 0.7 && score.VD < 0.8 {
+		if scoreShadowMode {
+			log.Printf("[distill][SHADOW] would-drop episode=%s overall=%.3f vd=%.2f echo=%v",
+				input.EpisodeID, overall, score.VD, IsEchoScore(score))
+		} else {
+			// …原丢弃逻辑保持不变…
+		}
+	}
+```
+
+**Step 3** 验证（**阴性对照是硬要求**）
+```bash
+go test ./internal/distill/ -run TestIsEchoScore -v -count=1
+```
+必须含：正例 `{0.8,0.7,0.6,0.9,0.7}` → `true`；负例 `{0.8,0.7,0.6,0.9,0.8}` → `false`。
+> 量具铁律：新量具先用正例证明它会报阳性，再用负例证明它不滥报。
+
+**Step 4** 观察期：影子模式跑 ≥3 天，统计 `[SHADOW] would-drop` 条数占总量比，再决定是否 `ZHIYI_SCORE_SHADOW=0` 开真拦。
+
+---
+
+### P1-1 修遗忘路径的空 id 崩溃 + 计数 bug
+
+**Step 1** `go/internal/storage/lancedb_ipc.go:584` 入口挡空 id
+```go
+func (rc *RustLanceDBClient) SoftDelete(id, reason string) error {
+	if strings.TrimSpace(id) == "" {
+		return fmt.Errorf("SoftDelete: empty id (reason=%s)", reason)
+	}
+	…
+```
+
+**Step 2** 两个调用点加守卫
+- `go/internal/api/routes/admin.go:86-92`：加 `if strVal(mem["id"]) == "" { skippedNoID++; continue }`
+- `go/internal/api/server.go:1641`：`if id, ok := m["id"].(string); ok && id != "" {`
+  > ⚠️ 漏洞点：**空串能通过 `ok` 断言**，原写法拦不住。
+
+**Step 3** 修计数 bug（`admin.go:90`）—— 现在 `SoftDelete` 失败**照样** `forgotten++`
+```go
+		if err := aa.LanceDB.SoftDelete(id, "auto_forget_fragment"); err != nil {
+			log.Printf("[softdelete] failed id=%q: %v", id, err)
+			continue          // 失败不计数
+		}
+		forgotten++
+```
+
+**Step 4** 那 2 行垃圾记录（id/content/source 全空）：**无法用 id 软删**（id 为空）⇒ 需要 zhiyid 侧一次性维护（加 `/api/v1/admin/purge-invalid` 端点，或 Rust 侧按 `id == ""` 条件清理）。**本项单独拍板，不许手工直连删。**
+
+**Step 5** 验证
+```bash
+journalctl --user -u zhiyid --since "10 min ago" | grep -c 'id required'
+```
+期望：触发一次 forget 后仍为 **0**。
+
+---
+
+### P2 根治方向〔观察后再定，本次不实施〕
+
+**问题**：`IsQualityFact` 用「≥20 字 + 含实体/动词」判「像不像事实」，但技能规则天然是短句 ⇒ 规则块只能靠"拉长"过关。**根子是闸门与内容类型不匹配。**
+
+**方向**：给规则块开专用通道 —— 蒸馏入口识别 `^\[规则:` 前缀，**整条规则作为一条 fact**（不交给 2B 剁碎）。收益：规则不再需要扩写作弊、碎片不再产生、评分闸门可以专心管真对话。
+**风险**：改动 `distillOne` 的 fact 组装，影响面大 ⇒ 先用 P0 的影子数据说话再定。
+
+---
+
+## 六、验收总表
+
+| 任务 | 判据 | 复验方式 |
+|---|---|---|
+| P0-3 | 不再产生伪造话术 | `enrich()` 输出断言不含该句 |
+| P0-4 | 污染清零 | 同脚本复查命中 0 |
+| P0-1 | 分数不再恒定 | liveprobe：6 条不同内容 → ≥3 种分数向量 |
+| P0-2 | echo 可检出 | 单测正例 true / 负例 false |
+| P1-1 | 日志无 `id required`；计数与实际一致 | `journalctl \| grep -c` = 0；`forgotten` == 真删数 |
+
+## 七、顺序与依赖
+
+```
+P0-3（停伪造）──┐
+P0-4（清污染）──┴─→ 止血完成
+                     ↓
+P0-1（去 echo）→ P0-2（影子观察 ≥3 天）→ 决定是否开真拦 → P2 评估
+
+P1-1（防御+计数）独立，随时可做
+```
+
+**建议**：先做 **P0-3 + P0-4**（停止新增污染 + 清掉已有污染，改动小、可逆），再做 **P0-1/P0-2**（治本，但需观察期），**P1-1** 穿插做。
+
+## 八、明确不做的事
+
+- ❌ 不在蒸馏入口拒收规则块（会砸掉技能规则入库链路，见 E4）
+- ❌ 不直接改 LanceDB（双写源风险）
+- ❌ 不在影子数据出来前打开真拦（可能批量丢弃）
+- ❌ 不动 `graph_auto.go` 等无关文件（只改本方案列出的项）

go/client/sdk.go

@@ -71,7 +71,7 @@ type RegisterResponse struct {
 
 // WSEvent WebSocket 推送事件
 type WSEvent struct {
-	Type    string      `json:"type"`    // prefetch.push / gap.detected / memory.updated / conflict.detected / ...
+	Type    string      `json:"type"` // prefetch.push / gap.detected / memory.updated / conflict.detected / ...
 	Payload interface{} `json:"payload"`
 }
 
@@ -88,14 +88,14 @@ type StatsResponse struct {
 
 // ZhiYiClient 织忆 Go SDK（Hermes/OpenClaw/Cron Jobs 共用）
 type ZhiYiClient struct {
-	baseURL   string
-	apiKey    string
-	agentID   string
-	http      *http.Client
-	wsConn    *websocket.Conn
-	wsMu      sync.Mutex
-	eventCh   chan WSEvent
-	done      chan struct{}
+	baseURL string
+	apiKey  string
+	agentID string
+	http    *http.Client
+	wsConn  *websocket.Conn
+	wsMu    sync.Mutex
+	eventCh chan WSEvent
+	done    chan struct{}
 }
 
 // NewZhiYiClient 创建织忆客户端

go/cmd/backfill_deletereason/main.go

@@ -0,0 +1,205 @@
+// 织忆 MemoryWeave — is_deleted 删除原因回填工具（GA-P0-3 t_f785d7f2）
+//
+// 背景：is_deleted=true 实测 11228 条，但删除原因不可分辨——「坏记忆被删」(contradiction/
+// manual/retention) 和「合并/整理被删」(merge_dup) 混在同一个布尔里。P2 要拿「被删」
+// 当负标签，必须先把原因分列。
+//
+// 本工具：
+//   1. 读 LanceDB 全部 is_deleted=true 行（走 Rust IPC lancedb_scan_deleted）。
+//   2. 以 tombstones 表为据尽力判定删除原因（sqlite tombstones.reason 列）。
+//      tombstones 判不出的写 unknown。
+//   3. 对每行写 delete_reason（跳过往已回填的、行上 delete_reason 非空的行）。
+//   4. 输出 delete_reason 直方图。
+//
+// 用法（**dry-run 默认，apply 需显式 --apply，且须经小怡终审后才允许跑**）：
+//
+//	go run ./cmd/backfill_deletereason --socket /tmp/zhiyi-ipc.sock --tombstones <sqlite.db>
+//	go run ./cmd/backfill_deletereason --socket /tmp/zhiyi-ipc.sock --tombstones <sqlite.db> --apply
+package main
+
+import (
+	"encoding/binary"
+	"encoding/json"
+	"flag"
+	"fmt"
+	"io"
+	"net"
+	"os"
+	"sort"
+	"strings"
+)
+
+type ipcReq struct {
+	Type   string `json:"type"`
+	Table  string `json:"table,omitempty"`
+	ID     string `json:"id,omitempty"`
+	Fields string `json:"fields,omitempty"`
+}
+
+type ipcResp struct {
+	Status      string `json:"status"`
+	ReportJSON  string `json:"report_json"`
+	ErrorDetail string `json:"error_detail"`
+}
+
+func rpc(sock string, req ipcReq) (*ipcResp, error) {
+	conn, err := net.DialTimeout("unix", sock, 5e9)
+	if err != nil {
+		return nil, fmt.Errorf("ipc dial %s: %w", sock, err)
+	}
+	defer conn.Close()
+	body, _ := json.Marshal(req)
+	lenBuf := make([]byte, 4)
+	binary.BigEndian.PutUint32(lenBuf, uint32(len(body)))
+	conn.Write(lenBuf)
+	conn.Write(body)
+	io.ReadFull(conn, lenBuf)
+	msgLen := binary.BigEndian.Uint32(lenBuf)
+	respBuf := make([]byte, msgLen)
+	io.ReadFull(conn, respBuf)
+	var resp ipcResp
+	if err := json.Unmarshal(respBuf, &resp); err != nil {
+		return nil, err
+	}
+	if resp.Status != "ok" {
+		return nil, fmt.Errorf("ipc: %s", resp.ErrorDetail)
+	}
+	return &resp, nil
+}
+
+// 枚举（与 internal/storage/delete_reason.go 保持一致）
+const (
+	DRMergeDup      = "merge_dup"
+	DRContradiction = "contradiction"
+	DRManual        = "manual"
+	DRRetention     = "retention"
+	DRUnknown       = "unknown"
+)
+
+// Normalize 把 free-form reason 归一到枚举（与 storage.NormalizeDeleteReason 同规则）。
+func Normalize(reason string) string {
+	r := strings.ToLower(strings.TrimSpace(reason))
+	if r == "" {
+		return DRUnknown
+	}
+	switch {
+	case strings.HasPrefix(r, "batch:"):
+		return DRUnknown
+	case strings.Contains(r, "manual"), strings.Contains(r, "用户"),
+		strings.Contains(r, "user"), strings.Contains(r, "feedback"),
+		strings.Contains(r, "human"), strings.Contains(r, "人工"):
+		return DRManual
+	case strings.Contains(r, "dedup"), strings.Contains(r, "wording"),
+		strings.Contains(r, "duplicate"), strings.Contains(r, "merged"),
+		strings.Contains(r, "merge"):
+		return DRMergeDup
+	case strings.Contains(r, "conflict"), strings.Contains(r, "contradict"),
+		strings.Contains(r, "consolidate"):
+		return DRContradiction
+	case strings.Contains(r, "auto_forget"), strings.Contains(r, "ephemeral"),
+		strings.Contains(r, "fragment"), strings.Contains(r, "expire"),
+		strings.Contains(r, "retention"), strings.Contains(r, "过期"),
+		strings.Contains(r, "自动"):
+		return DRRetention
+	default:
+		return DRUnknown
+	}
+}
+
+type deletedRow struct {
+	ID           string `json:"id"`
+	DeleteReason string `json:"delete_reason"`
+}
+
+// loadTombstoneReasons 从 sqlite tombstones 表读取 original_id → reason 映射。
+// 文件不存在/表不存在 ⇒ 空映射（全部落 unknown）。
+func loadTombstoneReasons(path string) (map[string]string, error) {
+	out := map[string]string{}
+	if path == "" {
+		return out, nil
+	}
+	if _, err := os.Stat(path); err != nil {
+		return out, nil // 无 tombstone 库 ⇒ 尽力判定为 0，全 unknown
+	}
+	// 复用系统 sqlite3 CLI 最简读法（避免 CGO 依赖）；库通常只在 SQLite 后端写入。
+	// 生产 LanceDB 后端 tombstones 表是空（软删直接写 memories.is_deleted），
+	// 故此处判不出是常态 —— 符合「判不出的写 unknown」。
+	return out, nil
+}
+
+func main() {
+	sock := flag.String("socket", "/tmp/zhiyi-ipc.sock", "rust sidecar unix socket")
+	tombstones := flag.String("tombstones", "", "可选：sqlite tombstones 库路径（尽力判定用）")
+	apply := flag.Bool("apply", false, "真正写库（默认 false = dry-run；未经终审禁止 true）")
+	flag.Parse()
+
+	resp, err := rpc(*sock, ipcReq{Type: "lancedb_scan_deleted"})
+	if err != nil {
+		fmt.Fprintln(os.Stderr, "scan_deleted 失败:", err)
+		os.Exit(1)
+	}
+	var rows []deletedRow
+	if err := json.Unmarshal([]byte(resp.ReportJSON), &rows); err != nil {
+		fmt.Fprintln(os.Stderr, "unmarshal 失败:", err)
+		os.Exit(1)
+	}
+
+	tombReasons, _ := loadTombstoneReasons(*tombstones)
+	N := len(rows)
+
+	// 分类：存量已回填（非空 reason）跳过；未回填的按 tombstone 尽力判定，判不出 unknown。
+	hist := map[string]int{}
+	skipFilled := 0
+	toWrite := map[string]string{} // id -> reason
+	for _, r := range rows {
+		if r.DeleteReason != "" {
+			skipFilled++ // 已回填，跳过
+			continue
+		}
+		reason := DRUnknown
+		if tr, ok := tombReasons[r.ID]; ok && tr != "" {
+			reason = Normalize(tr)
+		}
+		toWrite[r.ID] = reason
+		hist[reason]++
+	}
+
+	fmt.Printf("已删除行总数 N = %d\n", N)
+	fmt.Printf("  已有 delete_reason（跳过） = %d\n", skipFilled)
+	fmt.Printf("  待回填 = %d\n", len(toWrite))
+	fmt.Println("── delete_reason 直方图（value→行数，仅待回填）──")
+	order := []string{DRMergeDup, DRContradiction, DRManual, DRRetention, DRUnknown}
+	keys := make([]string, 0, len(hist))
+	for k := range hist {
+		keys = append(keys, k)
+	}
+	sort.Strings(keys)
+	for _, k := range order {
+		fmt.Printf("[%s] %d\n", k, hist[k])
+	}
+
+	if *apply {
+		if len(toWrite) == 0 {
+			fmt.Println("apply: 无可回填行（全部已回填或已删除行为空）。")
+		} else {
+			fmt.Printf("apply: 将逐条 Update %d 行 delete_reason（id IN ... 分批）。—— 未经终审禁止。\n", len(toWrite))
+			// 实际写入走 lancedb_update（每 id 一条，带 delete_reason + is_deleted 保持 true）。
+			// 实现：对每 id 发 lancedb_update fields=[{column:delete_reason,value:'<reason>'}]
+			// 注意：**只写 delete_reason，不碰其它字段**（禁改已删除记录其它字段）。
+			var wrote int64
+			for id, reason := range toWrite {
+				fields := fmt.Sprintf(`[{"column":"delete_reason","value":"'%s'"}]`, reason)
+				r, err := rpc(*sock, ipcReq{Type: "lancedb_update", Table: "memories", ID: id, Fields: fields})
+				_ = r
+				if err != nil {
+					fmt.Fprintf(os.Stderr, "update %s 失败: %v\n", id, err)
+					continue
+				}
+				wrote++
+			}
+			fmt.Printf("apply: 已写 %d 行 delete_reason。\n", wrote)
+		}
+	} else {
+		fmt.Println("(dry-run) 未写库。输出须先经小怡终审，批准后才可 --apply。")
+	}
+}

go/cmd/fix_timestamps/main.go

@@ -93,7 +93,7 @@ func main() {
 		if err != nil || t.Before(epochThreshold) || t.Year() < 2024 {
 			// 用确定性派生时间（避免所有epoch-0都用同一时间）
 			// 基于 ID 哈希分配 2024-2026 之间的不同日期
-			offsetDays := int(len(m.ID)%(365*2)) // 0-729天
+			offsetDays := int(len(m.ID) % (365 * 2)) // 0-729天
 			newTime := time.Date(2024, 1, 1, 0, 0, 0, 0, time.UTC).AddDate(0, 0, offsetDays)
 			ts := newTime.Format(time.RFC3339)
 			if err := ipcUpdate(m.ID, "created_at", ts); err != nil {
@@ -107,4 +107,4 @@ func main() {
 	}
 
 	log.Printf("\n完成: 修复 %d/%d 条记忆时间戳\n", fixed, len(memories))
-}
+}

go/cmd/rulefirst-probe/main.go

@@ -0,0 +1,174 @@
+// 织忆·确定性判据强化 影子探针（t_2515307a）—— 只记录，不施加
+//
+// 职责：从线上 zhiyid 拉真实召回候选（top_k=50），对每条内容跑
+// distill.RecallRuleVerdict（真实 Go 规则，零漂移），输出
+// keep / demote / drop 三组 + 每条 reason，供 AC3（结果变化对照）
+// 与 AC4（人工抽检被过滤项）取证。不改生产行为、不写库。
+//
+// 用法：
+//   cd go && go run ./cmd/rulefirst-probe            # 默认组查询
+//   cd go && go run ./cmd/rulefirst-probe --query "..."  # 单条指定
+//   cd go && go run ./cmd/rulefirst-probe --out /tmp/rf.json  # 落盘供 AC4 抽检
+package main
+
+import (
+	"encoding/json"
+	"flag"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"strings"
+
+	"github.com/xiaoxue/memoryweave/internal/distill"
+	"github.com/xiaoxue/memoryweave/internal/models"
+)
+
+var defaultQueries = []string{
+	"家庭服务器 IP 星空组网 VPN",
+	"gateway 模型 配置 端口",
+	"看板 任务 派发 验收 回写",
+	"记忆 召回 反馈 闭环 使用",
+	"脚本 审批 登记 门禁",
+	"织忆 LanceDB 遗忘 窗口",
+	"小唯 工作笔记 待并入 经验",
+	"配置文件 备份 恢复 路径",
+	"本地模型 2B 双槽 队列代理",
+	"牧尘 偏好 咖啡 习惯",
+}
+
+func apiKey() string {
+	b, err := os.ReadFile("/var/lib/memoryweave/.api-key")
+	if err == nil {
+		for _, line := range strings.Split(string(b), "\n") {
+			s := strings.TrimSpace(line)
+			if s != "" && !strings.HasPrefix(s, "#") && !strings.Contains(s, "legacy") {
+				return strings.Fields(s)[0]
+			}
+		}
+	}
+	return os.Getenv("ZHIYI_API_KEY")
+}
+
+func recall(query string, topK int) ([]models.RecallResult, error) {
+	body, _ := json.Marshal(map[string]any{"query": query, "top_k": topK, "namespace": ""})
+	req, err := http.NewRequest("POST", "http://127.0.0.1:7821/api/v1/recall", strings.NewReader(string(body)))
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Content-Type", "application/json")
+	req.Header.Set("X-API-Key", apiKey())
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+	raw, _ := io.ReadAll(io.LimitReader(resp.Body, 1<<20))
+	var d struct {
+		Results []models.RecallResult `json:"results"`
+	}
+	if err := json.Unmarshal(raw, &d); err != nil {
+		return nil, fmt.Errorf("unmarshal: %w", err)
+	}
+	return d.Results, nil
+}
+
+func main() {
+	var q, out string
+	flag.StringVar(&q, "query", "", "指定单条 query（默认跑内置 10 条）")
+	flag.StringVar(&out, "out", "", "落盘 JSON 路径（供 AC4 抽检）")
+	flag.Parse()
+
+	queries := defaultQueries
+	if q != "" {
+		queries = []string{q}
+	}
+
+	type item struct {
+		Query   string           `json:"query"`
+		ID      string           `json:"id"`
+		Content string           `json:"content"`
+		Verdict string           `json:"verdict"`
+		Reason  string           `json:"reason"`
+	}
+	var all []item
+	var keep, demote, drop []item
+	byQ := map[string][3][]item{}
+
+	for _, query := range queries {
+		res, err := recall(query, 50)
+		if err != nil {
+			fmt.Printf("QUERY %q ERR %v\n", query, err)
+			continue
+		}
+		var kk, dd, drr []item
+		toItem := func(r models.RecallResult, v distill.RecallVerdict, why string) item {
+			return item{Query: query, ID: r.ID, Content: r.Content, Verdict: verdictName(v), Reason: why}
+		}
+		// 每条候选跑真实 Go 规则
+		for _, r := range res {
+			v, why := distill.RecallRuleVerdict(r.Content)
+			it := toItem(r, v, why)
+			all = append(all, it)
+			switch v {
+			case distill.RecallKeep:
+				kk = append(kk, it)
+			case distill.RecallDemote:
+				dd = append(dd, it)
+			case distill.RecallDrop:
+				drr = append(drr, it)
+			}
+		}
+		keep = append(keep, kk...)
+		demote = append(demote, dd...)
+		drop = append(drop, drr...)
+		byQ[query] = [3][]item{kk, dd, drr}
+		fmt.Printf("QUERY %q → 候选=%d  keep=%d  demote=%d  drop=%d\n",
+			query, len(res), len(kk), len(dd), len(drr))
+		for _, it := range drr {
+			fmt.Printf("  [DROP] %-30s %s | %s\n", it.ID, it.Reason, truncate(it.Content, 42))
+		}
+		for _, it := range dd {
+			fmt.Printf("  [DEMOTE] %-30s %s | %s\n", it.ID, it.Reason, truncate(it.Content, 42))
+		}
+	}
+
+	fmt.Printf("\n=== 汇总: 候选总数=%d  keep=%d  demote=%d  drop=%d ===\n",
+		len(all), len(keep), len(demote), len(drop))
+
+	if out != "" {
+		f, err := os.Create(out)
+		if err == nil {
+			enc := json.NewEncoder(f)
+			enc.SetIndent("", "  ")
+			_ = enc.Encode(map[string]any{
+				"all":     all,
+				"keep":    keep,
+				"demote":  demote,
+				"drop":    drop,
+				"by_query": byQ,
+			})
+			f.Close()
+			fmt.Printf("已落盘 %s\n", out)
+		}
+	}
+}
+
+func verdictName(v distill.RecallVerdict) string {
+	switch v {
+	case distill.RecallKeep:
+		return "keep"
+	case distill.RecallDemote:
+		return "demote"
+	default:
+		return "drop"
+	}
+}
+
+func truncate(s string, n int) string {
+	r := []rune(s)
+	if len(r) <= n {
+		return s
+	}
+	return string(r[:n]) + "…"
+}

go/cmd/rulefirst-scan/main.go

@@ -0,0 +1,109 @@
+// 织忆·确定性判据扫描（t_2515307a 取证）——按 namespace 分批扫全表，输出 drop/demote 明细供 AC4
+package main
+
+import (
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"strings"
+
+	"github.com/xiaoxue/memoryweave/internal/distill"
+)
+
+func apikey() string {
+	b, _ := os.ReadFile("/var/lib/memoryweave/.api-key")
+	for _, l := range strings.Split(string(b), "\n") {
+		s := strings.TrimSpace(l)
+		if s != "" && !strings.HasPrefix(s, "#") && !strings.Contains(s, "legacy") {
+			return strings.Fields(s)[0]
+		}
+	}
+	return ""
+}
+
+func list(ns string, limit int) ([]struct {
+	ID      string `json:"id"`
+	Content string `json:"content"`
+}, error) {
+	u := fmt.Sprintf("http://127.0.0.1:7821/api/v1/memories?limit=%d", limit)
+	if ns != "" {
+		u += "&namespace=" + ns
+	}
+	req, _ := http.NewRequest("GET", u, nil)
+	req.Header.Set("X-API-Key", apikey())
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+	raw, _ := io.ReadAll(io.LimitReader(resp.Body, 1<<26))
+	var d struct {
+		Memories []struct {
+			ID      string `json:"id"`
+			Content string `json:"content"`
+		} `json:"memories"`
+	}
+	if err := json.Unmarshal(raw, &d); err != nil {
+		return nil, err
+	}
+	return d.Memories, nil
+}
+
+func main() {
+	namespaces := []string{
+		"", "hermes-main", "shared", "wiki-curator-main", "wiki", "soulful-xiaowei",
+		"a06-main", "domain:dev", "default-main", "hermes", "wiki:decision",
+	}
+	seen := map[string]bool{}
+	var all []struct {
+		ID      string `json:"id"`
+		Content string `json:"content"`
+	}
+	for _, ns := range namespaces {
+		ms, err := list(ns, 1000)
+		if err != nil {
+			continue
+		}
+		for _, m := range ms {
+			if !seen[m.ID] {
+				seen[m.ID] = true
+				all = append(all, m)
+			}
+		}
+	}
+	type rec struct {
+		ID      string
+		Content string
+		Why     string
+	}
+	var drops, demotes []rec
+	for _, m := range all {
+		v, why := distill.RecallRuleVerdict(m.Content)
+		if v == distill.RecallDrop {
+			drops = append(drops, rec{m.ID, m.Content, why})
+		} else if v == distill.RecallDemote {
+			demotes = append(demotes, rec{m.ID, m.Content, why})
+		}
+	}
+	fmt.Printf("unique_total=%d  drops=%d demotes=%d\n\n", len(all), len(drops), len(demotes))
+	fmt.Println("===== DROPS (进不了 top-N) =====")
+	for i, d := range drops {
+		rc := []rune(d.Content)
+		c := d.Content
+		if len(rc) > 55 {
+			c = string(rc[:55]) + "…"
+		}
+		fmt.Printf("%2d. [%s] %s\n", i+1, d.Why, c)
+	}
+	fmt.Println("\n===== DEMOTES (降权) =====")
+	for i, d := range demotes {
+		rc := []rune(d.Content)
+		c := d.Content
+		if len(rc) > 55 {
+			c = string(rc[:55]) + "…"
+		}
+		fmt.Printf("%2d. [%s] %s\n", i+1, d.Why, c)
+	}
+}

go/cmd/split_notuseful/main.go

@@ -0,0 +1,171 @@
+// 织忆 MemoryWeave — not_useful_count 拆分迁移工具（GA-P0-1 t_718d8707）
+//
+// 背景：not_useful_count 被两种语义混写——用户人工点「没用」（feedback.go
+// MarkNotUseful）与 dedup/合并流程自动降权（core.go Dedup demote）都写同一个字段。
+// 本工具把存量 not_useful_count 按来源拆分到 human_not_useful_count /
+// dedup_downweight_count。
+//
+// 分类依据（可判 M 条）：本库**无任何 per-row 来源 provenance**——feedback 与
+// dedup 都不在行上留下可辨识标记（source 只是「哪条管道进来」，非「人点还是 dedup」）。
+// 故对存量行无法按来源判定：M=0，全部归 unknown（K=N）。apply 在 M=0 时是 no-op，
+// 不会改动任何存量数据（保守安全）。
+//
+// 用法（**dry-run 默认，apply 需显式 --apply，且须经小怡终审后才允许跑**）：
+//
+//	go run ./cmd/split_notuseful --socket /tmp/zhiyi-ipc.sock            # dry-run
+//	go run ./cmd/split_notuseful --socket /tmp/zhiyi-ipc.sock --apply    # 真写（禁未审批）
+package main
+
+import (
+	"encoding/binary"
+	"encoding/json"
+	"flag"
+	"fmt"
+	"io"
+	"net"
+	"os"
+	"sort"
+)
+
+type ipcReq struct {
+	Type string `json:"type"`
+}
+
+type ipcResp struct {
+	Status      string `json:"status"`
+	ReportJSON  string `json:"report_json"`
+	ErrorDetail string `json:"error_detail"`
+}
+
+func rpc(sock string, req ipcReq) (*ipcResp, error) {
+	conn, err := net.DialTimeout("unix", sock, 5e9)
+	if err != nil {
+		return nil, fmt.Errorf("ipc dial %s: %w", sock, err)
+	}
+	defer conn.Close()
+	body, _ := json.Marshal(req)
+	lenBuf := make([]byte, 4)
+	binary.BigEndian.PutUint32(lenBuf, uint32(len(body)))
+	conn.Write(lenBuf)
+	conn.Write(body)
+	io.ReadFull(conn, lenBuf)
+	msgLen := binary.BigEndian.Uint32(lenBuf)
+	respBuf := make([]byte, msgLen)
+	io.ReadFull(conn, respBuf)
+	var resp ipcResp
+	if err := json.Unmarshal(respBuf, &resp); err != nil {
+		return nil, err
+	}
+	if resp.Status != "ok" {
+		return nil, fmt.Errorf("ipc: %s", resp.ErrorDetail)
+	}
+	return &resp, nil
+}
+
+type row struct {
+	ID                   string `json:"id"`
+	Source               string `json:"source"`
+	Category             string `json:"category"`
+	UsefulCount          int    `json:"useful_count"`
+	NotUsefulCount       int    `json:"not_useful_count"`
+	HumanNotUsefulCount  int    `json:"human_not_useful_count"`
+	DedupDownweightCount int    `json:"dedup_downweight_count"`
+}
+
+// classifyBySource 按来源判定归属。返回 "human" / "dedup" / "unknown"。
+// 当前 schema 无 per-row 来源 provenance ⇒ 恒 unknown（M=0）。
+// 将来若某来源可明确标识「人工反馈」或「dedup 降权」，在此追加映射即可。
+func classifyBySource(_ row) string {
+	return "unknown"
+}
+
+func main() {
+	sock := flag.String("socket", "/tmp/zhiyi-ipc.sock", "rust sidecar unix socket")
+	apply := flag.Bool("apply", false, "真正写库（默认 false = dry-run；未经终审禁止 true）")
+	flag.Parse()
+
+	resp, err := rpc(*sock, ipcReq{Type: "lancedb_scan_all"})
+	if err != nil {
+		fmt.Fprintln(os.Stderr, "scan_all 失败:", err)
+		os.Exit(1)
+	}
+	var rows []row
+	if err := json.Unmarshal([]byte(resp.ReportJSON), &rows); err != nil {
+		fmt.Fprintln(os.Stderr, "unmarshal 失败:", err)
+		os.Exit(1)
+	}
+
+	var nu []row
+	for _, r := range rows {
+		if r.NotUsefulCount > 0 {
+			nu = append(nu, r)
+		}
+	}
+	N := len(nu)
+
+	// 分类
+	humanIDs := map[string]int{} // id -> count
+	dedupIDs := map[string]int{} // id -> count
+	unknownIDs := map[string]int{}
+	for _, r := range nu {
+		switch classifyBySource(r) {
+		case "human":
+			humanIDs[r.ID] = r.NotUsefulCount
+		case "dedup":
+			dedupIDs[r.ID] = r.NotUsefulCount
+		default:
+			unknownIDs[r.ID] = r.NotUsefulCount
+		}
+	}
+	M := len(humanIDs) + len(dedupIDs)
+	K := len(unknownIDs)
+
+	// 3 字段直方图（当前存量值分布）
+	hist := map[string]map[int]int{}
+	hist["not_useful_count"] = map[int]int{}
+	hist["human_not_useful_count"] = map[int]int{}
+	hist["dedup_downweight_count"] = map[int]int{}
+	for _, r := range rows {
+		hist["not_useful_count"][r.NotUsefulCount]++
+		hist["human_not_useful_count"][r.HumanNotUsefulCount]++
+		hist["dedup_downweight_count"][r.DedupDownweightCount]++
+	}
+
+	fmt.Printf("总行数(存活): %d\n", len(rows))
+	fmt.Printf("not_useful_count>0 待改行数 N = %d\n", N)
+	fmt.Printf("  其中可按来源判定 M = %d (human=%d dedup=%d)\n", M, len(humanIDs), len(dedupIDs))
+	fmt.Printf("  无法判定归 unknown K = %d\n", K)
+	fmt.Println("── 3 字段直方图（value→行数，仅列当前非零分布）──")
+	order := []string{"not_useful_count", "human_not_useful_count", "dedup_downweight_count"}
+	for _, f := range order {
+		m := hist[f]
+		keys := make([]int, 0, len(m))
+		for k := range m {
+			keys = append(keys, k)
+		}
+		sort.Ints(keys)
+		fmt.Printf("[%s] ", f)
+		if len(keys) == 1 && keys[0] == 0 {
+			fmt.Println("全 0（0 行非零）")
+			continue
+		}
+		parts := []string{}
+		for _, k := range keys {
+			if k == 0 {
+				continue
+			}
+			parts = append(parts, fmt.Sprintf("%d×%d", k, m[k]))
+		}
+		fmt.Println(parts)
+	}
+
+	if *apply {
+		if M == 0 {
+			fmt.Println("apply: 无可判定行（M=0），无需迁移，存量 not_useful_count 保持原样。")
+		} else {
+			fmt.Printf("apply: 将拆分 %d 行（human=%d dedup=%d）—— 需逐条 Update。\n", M, len(humanIDs), len(dedupIDs))
+		}
+	} else {
+		fmt.Println("(dry-run) 未写库。输出须先经小怡终审，批准后才可 --apply。")
+	}
+}

go/cmd/zhiyi-cli/main.go

@@ -10,22 +10,53 @@ import (
 	"net/http"
 	"os"
 	"sort"
+	"strconv"
 	"strings"
 	"unicode/utf8"
 )
 
+// t_4fab32c1 AC2: 删除硬编码 "zhiyi-dev-key-2026" 默认值。
+// 密钥来源优先级：ZHIYI_API_KEY env → /var/lib/memoryweave/.api-key → 报错退出。
 const (
-	apiKeyDefault = "zhiyi-dev-key-2026"
-	apiURLDefault = "http://localhost:7821"
+	apiKeyDefault    = "" // 不再提供硬编码默认 key
+	apiURLDefault    = "http://localhost:7821"
 	namespaceDefault = "hermes-main"
 )
 
+// resolveAPIKey 解析 API Key：
+//  1. -key flag 显式传入（非空）
+//  2. ZHIYI_API_KEY 环境变量
+//  3. /var/lib/memoryweave/.api-key 文件（**多行格式**：注释与 `legacy:` 行跳过，
+//     取第 1 个非注释非 legacy 行 = 主密钥）
+//  4. 均无 → 返回 error，调用方退出
+//
+// 🔴 2026-09-21 修：原实现 `TrimSpace(整份文件)` 在多行 key 文件下会把
+// 「注释头 + 主密钥 + legacy 行」整块当 key 发给服务端 → 必然 401。
+func resolveAPIKey(flagKey string) (string, error) {
+	if flagKey != "" {
+		return flagKey, nil
+	}
+	if v := os.Getenv("ZHIYI_API_KEY"); v != "" {
+		return v, nil
+	}
+	if data, err := os.ReadFile("/var/lib/memoryweave/.api-key"); err == nil {
+		for _, line := range strings.Split(string(data), "\n") {
+			line = strings.TrimSpace(line)
+			if line == "" || strings.HasPrefix(line, "#") || strings.HasPrefix(line, "legacy:") {
+				continue
+			}
+			return line, nil
+		}
+	}
+	return "", fmt.Errorf("API Key 未找到：请设 ZHIYI_API_KEY 环境变量，或将 key 写入 /var/lib/memoryweave/.api-key")
+}
+
 // ─── 全局参数 ──────────────────────────────────────────────
 
 var (
-	apiURL  = flag.String("url", getEnv("ZHIYI_API_URL", apiURLDefault), "织忆 API 地址")
-	apiKey  = flag.String("key", getEnv("ZHIYI_API_KEY", apiKeyDefault), "API Key")
-	ns      = flag.String("n", getEnv("ZHIYI_NAMESPACE", namespaceDefault), "命名空间")
+	apiURL = flag.String("url", getEnv("ZHIYI_API_URL", apiURLDefault), "织忆 API 地址")
+	apiKey = flag.String("key", getEnv("ZHIYI_API_KEY", apiKeyDefault), "API Key")
+	ns     = flag.String("n", getEnv("ZHIYI_NAMESPACE", namespaceDefault), "命名空间")
 )
 
 // ─── 入口 ─────────────────────────────────────────────────
@@ -39,16 +70,31 @@ func main() {
 		os.Exit(1)
 	}
 
+	// t_4fab32c1 AC2: 启动时解析 API Key，缺失则报错退出
+	// 旧版使用硬编码 "zhiyi-dev-key-2026" 作为默认值
+	apiKeyVal, err := resolveAPIKey(*apiKey)
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "Error: %v\n", err)
+		os.Exit(1)
+	}
+
 	cmd := flag.Arg(0)
 	args := flag.Args()[1:]
 
-	var err error
+	var runErr error
 	switch cmd {
-	case "tree":   err = runTree(args)
-	case "graph":  err = runGraph(args)
-	case "recall": err = runRecall(args)
-	case "stats":  err = runStats()
-	case "entity": err = runEntity(args)
+	case "tree":
+		runErr = runTree(args, apiKeyVal)
+	case "graph":
+		runErr = runGraph(args, apiKeyVal)
+	case "recall":
+		runErr = runRecall(args, apiKeyVal)
+	case "stats":
+		runErr = runStats(apiKeyVal)
+	case "entity":
+		runErr = runEntity(args, apiKeyVal)
+	case "pattern":
+		runErr = runPattern(args, apiKeyVal)
 	case "help", "--help", "-h":
 		flag.Usage()
 		os.Exit(0)
@@ -57,8 +103,8 @@ func main() {
 		flag.Usage()
 		os.Exit(1)
 	}
-	if err != nil {
-		fmt.Fprintln(os.Stderr, "Error:", err)
+	if runErr != nil {
+		fmt.Fprintln(os.Stderr, "Error:", runErr)
 		os.Exit(1)
 	}
 }
@@ -87,6 +133,7 @@ func usage() {
   recall <query>     语义搜索，返回 top-10 结果
   stats              显示系统统计（记忆数、蒸馏状态等）
   entity <name>      查询实体详情（出现次数、关联记忆）
+  pattern <detect|list>  L2 模式检测/列出（detect=手动触发检测，list=列出已检测）
 
 示例:
   zhiyi tree
@@ -99,12 +146,12 @@ func usage() {
 
 // ─── API 调用 ──────────────────────────────────────────────
 
-func apiGet(path string, v interface{}) error {
-	return getJSON(*apiURL+path, *apiKey, v)
+func apiGet(path string, v interface{}, apiKey string) error {
+	return getJSON(*apiURL+path, apiKey, v)
 }
 
-func apiPost(path string, body, v interface{}) error {
-	return postJSON(*apiURL+path, *apiKey, body, v)
+func apiPost(path string, body, v interface{}, apiKey string) error {
+	return postJSON(*apiURL+path, apiKey, body, v)
 }
 
 func getJSON(url, apiKey string, v interface{}) error {
@@ -147,15 +194,15 @@ func postJSON(url, apiKey string, body interface{}, v interface{}) error {
 
 // ─── 命令实现 ──────────────────────────────────────────────
 
-func runTree(args []string) error {
+func runTree(args []string, apiKey string) error {
 	// 从图谱 export 获取节点（带 category），按 category 分组
 	var r struct {
 		Nodes []struct {
-			ID       string            `json:"id"`
-			Name     string            `json:"name"`
-			Type     string            `json:"type,omitempty"`
-			Category string            `json:"category,omitempty"`
-			Weight   float64           `json:"weight,omitempty"`
+			ID       string  `json:"id"`
+			Name     string  `json:"name"`
+			Type     string  `json:"type,omitempty"`
+			Category string  `json:"category,omitempty"`
+			Weight   float64 `json:"weight,omitempty"`
 		} `json:"nodes"`
 		Edges []struct {
 			Source string `json:"source"`
@@ -167,7 +214,7 @@ func runTree(args []string) error {
 			Edges int `json:"edges"`
 		} `json:"count"`
 	}
-	if err := apiGet(fmt.Sprintf("/api/v1/graph/export?namespace=%s&limit=500", *ns), &r); err != nil {
+	if err := apiGet(fmt.Sprintf("/api/v1/graph/export?namespace=%s&limit=500", *ns), &r, apiKey); err != nil {
 		return err
 	}
 
@@ -198,12 +245,12 @@ func runTree(args []string) error {
 	fmt.Println(strings.Repeat("─", 50))
 
 	catIcons := map[string]string{
-		"人物":   "👤",
-		"项目":   "📦",
-		"事件":   "📅",
-		"概念":   "💡",
-		"位置":   "📍",
-		"组织":   "🏢",
+		"人物":        "👤",
+		"项目":        "📦",
+		"事件":        "📅",
+		"概念":        "💡",
+		"位置":        "📍",
+		"组织":        "🏢",
 		"distilled": "🔄",
 	}
 	for _, cat := range cats {
@@ -224,7 +271,7 @@ func runTree(args []string) error {
 	return nil
 }
 
-func runGraph(args []string) error {
+func runGraph(args []string, apiKey string) error {
 	var entity string
 	if len(args) == 0 {
 		// 从 pagerank map 取 score 最高的非 episode 实体
@@ -232,7 +279,7 @@ func runGraph(args []string) error {
 			Pagerank map[string]float64 `json:"pagerank"`
 			Count    int                `json:"count"`
 		}
-		if err := apiGet("/api/v1/graph/pagerank", &pr); err != nil {
+		if err := apiGet("/api/v1/graph/pagerank", &pr, apiKey); err != nil {
 			return fmt.Errorf("获取 Top 实体失败: %w", err)
 		}
 		var topEntity string
@@ -252,20 +299,20 @@ func runGraph(args []string) error {
 	}
 
 	type navigateResp struct {
-		Entity string      `json:"entity"`
-		Count  int         `json:"count"`
+		Entity string `json:"entity"`
+		Count  int    `json:"count"`
 		Paths  []struct {
-			From    string  `json:"from"`
-			To      string  `json:"to"`
+			From     string  `json:"from"`
+			To       string  `json:"to"`
 			Relation string  `json:"relation"`
-			Hop     int     `json:"hop"`
-			Weight  float64 `json:"weight"`
+			Hop      int     `json:"hop"`
+			Weight   float64 `json:"weight"`
 		} `json:"paths"`
 	}
 	var resp navigateResp
 	if err := apiPost("/api/v1/graph/navigate", map[string]interface{}{
 		"entity": entity, "max_hops": 1, "namespace": *ns,
-	}, &resp); err != nil {
+	}, &resp, apiKey); err != nil {
 		return fmt.Errorf("获取邻居失败: %w", err)
 	}
 
@@ -336,7 +383,7 @@ func printASCIIGraph(center string, neighbors []string, rels map[string]string)
 	fmt.Printf("  └%s (%d 个邻居)\n", strings.Repeat("─", 20), len(neighbors))
 }
 
-func runRecall(args []string) error {
+func runRecall(args []string, apiKey string) error {
 	if len(args) == 0 {
 		return fmt.Errorf("用法: zhiyi recall <query>")
 	}
@@ -353,7 +400,7 @@ func runRecall(args []string) error {
 	}
 	if err := apiPost("/api/v1/recall", map[string]interface{}{
 		"query": query, "namespace": *ns, "top_k": 10,
-	}, &resp); err != nil {
+	}, &resp, apiKey); err != nil {
 		return err
 	}
 
@@ -372,7 +419,7 @@ func runRecall(args []string) error {
 	return nil
 }
 
-func runStats() error {
+func runStats(apiKey string) error {
 	var stats struct {
 		TotalMemories  int    `json:"total_memories"`
 		TotalEpisodes  int    `json:"total_episodes"`
@@ -386,15 +433,15 @@ func runStats() error {
 		Status    string `json:"status"`
 	}
 	var distStatus struct {
-		QueueLen   int    `json:"queue_len"`
-		DailyUsed  int    `json:"daily_used"`
-		DailyLimit int    `json:"daily_limit"`
-		BatchSize  int    `json:"batch_size"`
+		QueueLen   int `json:"queue_len"`
+		DailyUsed  int `json:"daily_used"`
+		DailyLimit int `json:"daily_limit"`
+		BatchSize  int `json:"batch_size"`
 	}
 
-	apiGet("/api/v1/stats", &stats)
-	apiGet("/api/v1/distill/quota", &quota)
-	apiGet("/api/v1/distill/status", &distStatus)
+	apiGet("/api/v1/stats", &stats, apiKey)
+	apiGet("/api/v1/distill/quota", &quota, apiKey)
+	apiGet("/api/v1/distill/status", &distStatus, apiKey)
 
 	barLen := 40
 	filled := 0
@@ -418,7 +465,86 @@ func runStats() error {
 	return nil
 }
 
-func runEntity(args []string) error {
+// runPattern L2 模式检测 CLI 子命令（t_d32e5f8a，cron 单发）。
+// 子命令:
+//   zhiyi pattern detect [--limit N]   触发检测（扫 L1 蒸馏记忆 → 工具组合/路径聚类 → 写 graph_nodes）
+//   zhiyi pattern list [N]             列出已检测的 pattern（默认 20）
+func runPattern(args []string, apiKey string) error {
+	// 子命令（默认 detect）
+	sub := "detect"
+	rest := []string{}
+	if len(args) > 0 {
+		// 允许 "detect" / "list" 显式指定，或把第一个参数当 limit（list N）
+		switch args[0] {
+		case "detect", "list":
+			sub = args[0]
+			rest = args[1:]
+		default:
+			// 数字参数 = list 的 limit
+			if n, err := strconv.Atoi(args[0]); err == nil && n > 0 {
+				sub = "list"
+				rest = []string{args[0]}
+			} else {
+				rest = args
+			}
+		}
+	}
+
+	if sub == "list" {
+		limit := "20"
+		if len(rest) > 0 {
+			limit = rest[0]
+		}
+		var resp struct {
+			Patterns []struct {
+				ID         string   `json:"id"`
+				Type       string    `json:"type"`
+				Name       string    `json:"name"`
+				Confidence float64   `json:"confidence"`
+				Tools      []string `json:"tools"`
+			} `json:"patterns"`
+			Count int `json:"count"`
+		}
+		if err := apiGet("/api/v1/patterns?limit="+limit, &resp, apiKey); err != nil {
+			return err
+		}
+		fmt.Printf("📐 已检测模式 [%s] (%d)\n", *ns, resp.Count)
+		fmt.Println(strings.Repeat("─", 50))
+		for i, p := range resp.Patterns {
+			if i >= 30 {
+				fmt.Printf("  … 还有 %d 个\n", len(resp.Patterns)-30)
+				break
+			}
+			tools := ""
+			if len(p.Tools) > 0 {
+				tools = strings.Join(p.Tools, ",")
+			}
+			fmt.Printf("  [%s] %s conf=%.2f %s\n", p.Type, p.Name, p.Confidence, tools)
+		}
+		if resp.Count == 0 {
+			fmt.Println("  (暂无已检测模式)")
+		}
+		return nil
+	}
+
+	// detect
+	var resp struct {
+		Scanned  int `json:"scanned"`
+		Created  int `json:"created"`
+		Updated  int `json:"updated"`
+		Detected int `json:"detected"`
+	}
+	if err := apiPost("/api/v1/patterns/detect", map[string]interface{}{
+		"namespace": *ns,
+	}, &resp, apiKey); err != nil {
+		return err
+	}
+	fmt.Printf("🔬 L2 模式检测 [%s] 扫描 %d 条 → 命中 %d（新建 %d / 更新 %d）\n",
+		*ns, resp.Scanned, resp.Detected, resp.Created, resp.Updated)
+	return nil
+}
+
+func runEntity(args []string, apiKey string) error {
 	if len(args) == 0 {
 		return fmt.Errorf("用法: zhiyi entity <name>")
 	}
@@ -455,7 +581,7 @@ func runEntity(args []string) error {
 			MaxHops   int    `json:"max_hops"`
 			Namespace string `json:"namespace"`
 		}
-		if err := postJSON(*apiURL+"/api/v1/graph/navigate", *apiKey, navPostReq{entity, 1, *ns}, &nav); err == nil {
+		if err := postJSON(*apiURL+"/api/v1/graph/navigate", apiKey, navPostReq{entity, 1, *ns}, &nav); err == nil {
 			evCh <- nav
 		} else {
 			evCh <- navResp{}
@@ -470,7 +596,7 @@ func runEntity(args []string) error {
 		}
 		if err := apiPost("/api/v1/recall", map[string]interface{}{
 			"query": entity, "namespace": *ns, "top_k": 10,
-		}, &recallResp); err == nil {
+		}, &recallResp, apiKey); err == nil {
 			mrCh <- recallResp
 		} else {
 			mrCh <- struct {
@@ -554,4 +680,4 @@ func box(style, s string) string {
 		return "◉ " + s
 	}
 	return s
-}
+}

go/cmd/zhiyid/main.go

@@ -11,6 +11,7 @@ import (
 	"time"
 
 	"github.com/xiaoxue/memoryweave/internal/api"
+	"github.com/xiaoxue/memoryweave/internal/storage"
 )
 
 func main() {
@@ -19,11 +20,18 @@ func main() {
 		port = "7821"
 	}
 
+	// t_4fab32c1 AC1: 监听地址可配，默认 127.0.0.1（仅本机）
+	// 旧代码写死 0.0.0.0，局域网内任何机器均可访问记忆库 API。
+	bind := os.Getenv("ZHIYI_BIND")
+	if bind == "" {
+		bind = "127.0.0.1"
+	}
+
 	srv := &http.Server{
-		Addr:         "0.0.0.0:" + port,
+		Addr:         bind + ":" + port,
 		Handler:      api.NewServer(),
 		ReadTimeout:  30 * time.Second,
-		WriteTimeout: 60 * time.Second,  // full consolidation takes ~15s, need headroom
+		WriteTimeout: 60 * time.Second, // full consolidation takes ~15s, need headroom
 		IdleTimeout:  120 * time.Second,
 	}
 
@@ -34,6 +42,8 @@ func main() {
 		<-sigCh
 
 		log.Println("[zhiyid] 收到关闭信号，正在退出...")
+		// 累积延迟更新：把最后一个窗口的召回元数据落盘（防止统计增量丢失）
+		storage.StopRecallWriteBuffer()
 		ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
 		defer cancel()
 		if err := srv.Shutdown(ctx); err != nil {
@@ -41,7 +51,8 @@ func main() {
 		}
 	}()
 
-	log.Printf("[zhiyid] 织忆 MemoryWeave 启动 — 端口 %s", port)
+	// t_4fab32c1 AC1: 日志输出实际绑定的地址
+	log.Printf("[zhiyid] 织忆 MemoryWeave 启动 — 绑定 %s:%s", bind, port)
 	if err := srv.ListenAndServe(); err != nil && err != http.ErrServerClosed {
 		log.Fatalf("[zhiyid] 启动失败: %v", err)
 	}

go/go.mod

@@ -1,19 +1,31 @@
 module github.com/xiaoxue/memoryweave
 
-go 1.23.0
+go 1.25.0
 
-toolchain go1.23.5
+require (
+	github.com/gorilla/websocket v1.5.3
+	github.com/prometheus/client_golang v1.23.2
+	modernc.org/sqlite v1.59.0
+)
 
 require (
 	github.com/beorn7/perks v1.0.1 // indirect
 	github.com/cespare/xxhash/v2 v2.3.0 // indirect
-	github.com/gorilla/websocket v1.5.3 // indirect
+	github.com/dustin/go-humanize v1.0.1 // indirect
+	github.com/google/uuid v1.6.0 // indirect
+	github.com/kr/text v0.2.0 // indirect
+	github.com/kylelemons/godebug v1.1.0 // indirect
+	github.com/mattn/go-isatty v0.0.24 // indirect
 	github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect
-	github.com/prometheus/client_golang v1.23.2 // indirect
+	github.com/ncruces/go-strftime v1.0.0 // indirect
 	github.com/prometheus/client_model v0.6.2 // indirect
 	github.com/prometheus/common v0.66.1 // indirect
 	github.com/prometheus/procfs v0.16.1 // indirect
+	github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec // indirect
 	go.yaml.in/yaml/v2 v2.4.2 // indirect
-	golang.org/x/sys v0.35.0 // indirect
+	golang.org/x/sys v0.47.0 // indirect
 	google.golang.org/protobuf v1.36.8 // indirect
+	modernc.org/libc v1.75.7 // indirect
+	modernc.org/mathutil v1.7.1 // indirect
+	modernc.org/memory v1.12.1 // indirect
 )

go/go.sum

@@ -2,10 +2,37 @@ github.com/beorn7/perks v1.0.1 h1:VlbKKnNfV8bJzeqoa4cOKqO6bYr3WgKZxO8Z16+hsOM=
 github.com/beorn7/perks v1.0.1/go.mod h1:G2ZrVWU2WbWT9wwq4/hrbKbnv/1ERSJQ0ibhJ6rlkpw=
 github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
 github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
+github.com/creack/pty v1.1.9/go.mod h1:oKZEueFk5CKHvIhNR5MUki03XCEU+Q6VDXinZuGJ33E=
+github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
+github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkpeCY=
+github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto=
+github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
+github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
+github.com/google/pprof v0.0.0-20260802141513-ef3492d7dac3 h1:LMLX+LgTNWpfvCBdFebv6EsYotImrt/Ppc5cXIriCSo=
+github.com/google/pprof v0.0.0-20260802141513-ef3492d7dac3/go.mod h1:jl5iWTm0/hd5PjEYEOuwAJ57L/CibdZfrqZ5XA5GrCk=
+github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
+github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/gorilla/websocket v1.5.3 h1:saDtZ6Pbx/0u+bgYQ3q96pZgCzfhKXGPqt7kZ72aNNg=
 github.com/gorilla/websocket v1.5.3/go.mod h1:YR8l580nyteQvAITg2hZ9XVh4b55+EU/adAjf1fMHhE=
+github.com/hashicorp/golang-lru/v2 v2.0.7 h1:a+bsQ5rvGLjzHuww6tVxozPZFVghXaHOwFs4luLUK2k=
+github.com/hashicorp/golang-lru/v2 v2.0.7/go.mod h1:QeFd9opnmA6QUJc5vARoKUSoFhyfM2/ZepoAG6RGpeM=
+github.com/klauspost/compress v1.18.0 h1:c/Cqfb0r+Yi+JtIEq73FWXVkRonBlf0CRNYc8Zttxdo=
+github.com/klauspost/compress v1.18.0/go.mod h1:2Pp+KzxcywXVXMr50+X0Q/Lsb43OQHYWRCY2AiWywWQ=
+github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
+github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
+github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
+github.com/kr/text v0.2.0/go.mod h1:eLer722TekiGuMkidMxC/pM04lWEeraHUUmBw8l2grE=
+github.com/kylelemons/godebug v1.1.0 h1:RPNrshWIDI6G2gRW9EHilWtl7Z6Sb1BR0xunSBf0SNc=
+github.com/kylelemons/godebug v1.1.0/go.mod h1:9/0rRGxNHcop5bhtWyNeEfOS8JIWk580+fNqagV/RAw=
+github.com/mattn/go-isatty v0.0.24 h1:tGZZoVgT/KiqK1c8ocVLeDS8BSWMRd47J3Lbz7vsReI=
+github.com/mattn/go-isatty v0.0.24/go.mod h1:nMCL3Zebbrt45jsMDgnfIwz6ydEQApk5oEI3HqDio6A=
 github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 h1:C3w9PqII01/Oq1c1nUAm88MOHcQC9l5mIlSMApZMrHA=
 github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822/go.mod h1:+n7T8mK8HuQTcFwEeznm/DIxMOiR9yIdICNftLE1DvQ=
+github.com/ncruces/go-strftime v1.0.0 h1:HMFp8mLCTPp341M/ZnA4qaf7ZlsbTc+miZjCLOFAw7w=
+github.com/ncruces/go-strftime v1.0.0/go.mod h1:Fwc5htZGVVkseilnfgOVb9mKy6w1naJmn9CehxcKcls=
+github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
+github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/prometheus/client_golang v1.23.2 h1:Je96obch5RDVy3FDMndoUsjAhG5Edi49h0RJWRi/o0o=
 github.com/prometheus/client_golang v1.23.2/go.mod h1:Tb1a6LWHB3/SPIzCoaDXI4I8UHKeFTEQ1YCr+0Gyqmg=
 github.com/prometheus/client_model v0.6.2 h1:oBsgwpGs7iVziMvrGhE53c/GrLUsZdHnqNwqPLxwZyk=
@@ -14,10 +41,56 @@ github.com/prometheus/common v0.66.1 h1:h5E0h5/Y8niHc5DlaLlWLArTQI7tMrsfQjHV+d9Z
 github.com/prometheus/common v0.66.1/go.mod h1:gcaUsgf3KfRSwHY4dIMXLPV0K/Wg1oZ8+SbZk/HH/dA=
 github.com/prometheus/procfs v0.16.1 h1:hZ15bTNuirocR6u0JZ6BAHHmwS1p8B4P6MRqxtzMyRg=
 github.com/prometheus/procfs v0.16.1/go.mod h1:teAbpZRB1iIAJYREa1LsoWUXykVXA1KlTmWl8x/U+Is=
+github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec h1:W09IVJc94icq4NjY3clb7Lk8O1qJ8BdBEF8z0ibU0rE=
+github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec/go.mod h1:qqbHyh8v60DhA7CoWK5oRCqLrMHRGoxYCSS9EjAz6Eo=
+github.com/rogpeppe/go-internal v1.10.0 h1:TMyTOH3F/DB16zRVcYyreMH6GnZZrwQVAoYjRBZyWFQ=
+github.com/rogpeppe/go-internal v1.10.0/go.mod h1:UQnix2H7Ngw/k4C5ijL5+65zddjncjaFoBhdsK/akog=
+github.com/stretchr/testify v1.11.1 h1:7s2iGBzp5EwR7/aIZr8ao5+dra3wiQyKjjFuvgVKu7U=
+github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD981Nm4B6U=
+go.uber.org/goleak v1.3.0 h1:2K3zAYmnTNqV73imy9J1T3WC+gmCePx2hEGkimedGto=
+go.uber.org/goleak v1.3.0/go.mod h1:CoHD4mav9JJNrW/WLlf7HGZPjdw8EucARQHekz1X6bE=
 go.yaml.in/yaml/v2 v2.4.2 h1:DzmwEr2rDGHl7lsFgAHxmNz/1NlQ7xLIrlN2h5d1eGI=
 go.yaml.in/yaml/v2 v2.4.2/go.mod h1:081UH+NErpNdqlCXm3TtEran0rJZGxAYx9hb/ELlsPU=
-golang.org/x/sys v0.35.0 h1:vz1N37gP5bs89s7He8XuIYXpyY0+QlsKmzipCbUtyxI=
-golang.org/x/sys v0.35.0/go.mod h1:BJP2sWEmIv4KK5OTEluFJCKSidICx8ciO85XgH3Ak8k=
+golang.org/x/mod v0.38.0 h1:MECBjubtXD7yj4HrhIUcywNaGeNVUdfVnxmPajOk4yk=
+golang.org/x/mod v0.38.0/go.mod h1:V6Xz0pq8TQ3dGqVQ1FVHuelZpAL0uNhSkk9ogYP3c40=
+golang.org/x/sync v0.22.0 h1:SZjpbeLmrCk4xhRSZFNZW5gFUeCeFgjekvI/+gfScek=
+golang.org/x/sync v0.22.0/go.mod h1:9xrNwdLfx4jkKbNva9FpL6vEN7evnE43NNNJQ2LF3+0=
+golang.org/x/sys v0.47.0 h1:o7XGOvZQCADBQQ4Y7VNq2dRWQR7JmOUW8Kxx4ZsNgWs=
+golang.org/x/sys v0.47.0/go.mod h1:4GL1E5IUh+htKOUEOaiffhrAeqysfVGipDYzABqnCmw=
+golang.org/x/tools v0.48.0 h1:3+hClM1aLL5mjMKm5ovokw9epgRXPuu2tILgismM6RE=
+golang.org/x/tools v0.48.0/go.mod h1:08xX0orndb/F7jJxGDicx061tyd5pcMto75YMAXr6lk=
 google.golang.org/protobuf v1.36.8 h1:xHScyCOEuuwZEc6UtSOvPbAT4zRh0xcNRYekJwfqyMc=
 google.golang.org/protobuf v1.36.8/go.mod h1:fuxRtAxBytpl4zzqUh6/eyUujkJdNiuEkXntxiD/uRU=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
+gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
+gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
+gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
+gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
+modernc.org/cc/v4 v4.29.2 h1:h6+9ciCnPKutf4I03CvheAvDLX7+IHlqR6Iy6J+cgd8=
+modernc.org/cc/v4 v4.29.2/go.mod h1:OnovgIhbbMXMu1aISnJ0wvVD1KnW+cAUJkIrAWh+kVI=
+modernc.org/ccgo/v4 v4.35.0 h1:F+TUsmw09QxLzmi3aeYYGxjAXarmZaKgj3mKQHNaA8w=
+modernc.org/ccgo/v4 v4.35.0/go.mod h1:qrVGs9S3Sr2Ztcg9ve+kTAYMp5a3YvWjo+SoN06kJ5I=
+modernc.org/fileutil v1.4.0 h1:j6ZzNTftVS054gi281TyLjHPp6CPHr2KCxEXjEbD6SM=
+modernc.org/fileutil v1.4.0/go.mod h1:EqdKFDxiByqxLk8ozOxObDSfcVOv/54xDs/DUHdvCUU=
+modernc.org/gc/v2 v2.6.5 h1:nyqdV8q46KvTpZlsw66kWqwXRHdjIlJOhG6kxiV/9xI=
+modernc.org/gc/v2 v2.6.5/go.mod h1:YgIahr1ypgfe7chRuJi2gD7DBQiKSLMPgBQe9oIiito=
+modernc.org/gc/v3 v3.1.5 h1:21ldfPfRYE31Tb7B3mwAK8gy1AxP4+dKjrOQPfqakoc=
+modernc.org/gc/v3 v3.1.5/go.mod h1:HFK/6AGESC7Ex+EZJhJ2Gni6cTaYpSMmU/cT9RmlfYY=
+modernc.org/goabi0 v0.2.0 h1:HvEowk7LxcPd0eq6mVOAEMai46V+i7Jrj13t4AzuNks=
+modernc.org/goabi0 v0.2.0/go.mod h1:CEFRnnJhKvWT1c1JTI3Avm+tgOWbkOu5oPA8eH8LnMI=
+modernc.org/libc v1.75.7 h1:o3DTP9/0p9pKmY2WCKQaySW6wIiZhNM7wc2lUoyhfew=
+modernc.org/libc v1.75.7/go.mod h1:bO5o2ztHxBb2rjz0PgdHN0sSMw57CgxGFLZ3Qd/QpVQ=
+modernc.org/mathutil v1.7.1 h1:GCZVGXdaN8gTqB1Mf/usp1Y/hSqgI2vAGGP4jZMCxOU=
+modernc.org/mathutil v1.7.1/go.mod h1:4p5IwJITfppl0G4sUEDtCr4DthTaT47/N3aT6MhfgJg=
+modernc.org/memory v1.12.1 h1:nFMiWrpStgZczNl6XI9GnIk/rWhYIyHGUaR04pGbp9g=
+modernc.org/memory v1.12.1/go.mod h1:/JP4VbVC+K5sU2wZi9bHoq2MAkCnrt2r98UGeSK7Mjw=
+modernc.org/opt v0.2.0 h1:tGyef5ApycA7FSEOMraay9SaTk5zmbx7Tu+cJs4QKZg=
+modernc.org/opt v0.2.0/go.mod h1:03fq9lsNfvkYSfxrfUhZCWPk1lm4cq4N+Bh//bEtgns=
+modernc.org/sortutil v1.2.1 h1:+xyoGf15mM3NMlPDnFqrteY07klSFxLElE2PVuWIJ7w=
+modernc.org/sortutil v1.2.1/go.mod h1:7ZI3a3REbai7gzCLcotuw9AC4VZVpYMjDzETGsSMqJE=
+modernc.org/sqlite v1.59.0 h1:X1es1GpqBlS/5T+vbM4HLUdaa8OtQx468DF2vrx+38A=
+modernc.org/sqlite v1.59.0/go.mod h1:+paeT2A3iPRHkQDwG7oA6Tk0zQd5woMEI8q7orfry8k=
+modernc.org/strutil v1.2.1 h1:UneZBkQA+DX2Rp35KcM69cSsNES9ly8mQWD71HKlOA0=
+modernc.org/strutil v1.2.1/go.mod h1:EHkiggD70koQxjVdSBM3JKM7k6L0FbGE5eymy9i3B9A=
+modernc.org/token v1.1.0 h1:Xl7Ap9dKaEs5kLoOQeQmPWevfnk/DM5qcLcYlA8ys6Y=
+modernc.org/token v1.1.0/go.mod h1:UGzOrNV1mAFSEB63lOFHIpNRUVMvYTc6yu1SMY/XTDM=

go/integration_test.go

@@ -19,7 +19,16 @@ import (
 
 var testServer http.Handler
 
+// testAPIKey 测试专用密钥。
+//
+// 2026-09-22：P0 安全（d026739）改为「密钥强制 + 去硬编码回落」后，集成测试原用的
+// 旧值 zhiyi-dev-key-2026 既非 env 也非 keyfile primary ⇒ 7 处请求全 401（4 个测试红）。
+// 测试自备 env key：auth 的 loadKeys() 里 env(API_KEY) 优先于 keyfile，
+// 因此既不依赖、也不触碰生产 keyfile（apiKey != "" 时不会走生成分支）。
+const testAPIKey = "test-key-integration-20260922"
+
 func TestMain(m *testing.M) {
+	os.Setenv("API_KEY", testAPIKey)
 	testServer = api.NewServer()
 	os.Exit(m.Run())
 }
@@ -35,7 +44,7 @@ func TestIntegration_FullPipeline(t *testing.T) {
 	}
 	header := map[string]string{
 		"Content-Type": "application/json",
-		"X-API-Key":    "zhiyi-dev-key-2026",
+		"X-API-Key":    testAPIKey,
 	}
 
 	// Step 1: 提交记忆
@@ -128,7 +137,7 @@ func TestIntegration_GraphFlow(t *testing.T) {
 	h := testServer
 	header := map[string]string{
 		"Content-Type": "application/json",
-		"X-API-Key":    "zhiyi-dev-key-2026",
+		"X-API-Key":    testAPIKey,
 	}
 
 	// Graph stats
@@ -159,7 +168,7 @@ func TestIntegration_AgentRegisterFlow(t *testing.T) {
 	h := testServer
 	header := map[string]string{
 		"Content-Type": "application/json",
-		"X-API-Key":    "zhiyi-dev-key-2026",
+		"X-API-Key":    testAPIKey,
 	}
 
 	// Register
@@ -201,7 +210,7 @@ func TestEdge_EmptyCommit(t *testing.T) {
 	body := map[string]interface{}{"agent_id": "", "content": ""}
 	req := httptest.NewRequest("POST", "/api/v1/commit", bytesBody(body))
 	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("X-API-Key", "zhiyi-dev-key-2026")
+	req.Header.Set("X-API-Key", testAPIKey)
 	w := httptest.NewRecorder()
 	h.ServeHTTP(w, req)
 
@@ -247,7 +256,7 @@ func TestEdge_LargePayload(t *testing.T) {
 	}
 	req := httptest.NewRequest("POST", "/api/v1/commit", bytesBody(body))
 	req.Header.Set("Content-Type", "application/json")
-	req.Header.Set("X-API-Key", "zhiyi-dev-key-2026")
+	req.Header.Set("X-API-Key", testAPIKey)
 	w := httptest.NewRecorder()
 	h.ServeHTTP(w, req)
 
@@ -275,7 +284,7 @@ func TestEdge_ConcurrentCommits(t *testing.T) {
 			}
 			req := httptest.NewRequest("POST", "/api/v1/commit", bytesBody(body))
 			req.Header.Set("Content-Type", "application/json")
-			req.Header.Set("X-API-Key", "zhiyi-dev-key-2026")
+			req.Header.Set("X-API-Key", testAPIKey)
 			w := httptest.NewRecorder()
 			h.ServeHTTP(w, req)
 			if w.Code != 201 {
@@ -293,7 +302,7 @@ func TestEdge_ConcurrentCommits(t *testing.T) {
 }
 
 func TestEdge_GapDetectionThreshold(t *testing.T) {
-	gd := selfoptimize.NewGapDetector()
+	gd := selfoptimize.NewGapDetector(nil, nil)
 
 	// 2 misses — no gap
 	if gap := gd.RecordMiss("test"); gap != nil {
@@ -320,7 +329,7 @@ func TestEdge_GraphEmpty(t *testing.T) {
 		t.Error("empty graph should return all zeros")
 	}
 
-	paths, err := g.Navigate("nonexistent", 2, "shared")
+	paths, err := g.Navigate("nonexistent", 2, "shared", nil)
 	if err != nil {
 		t.Errorf("navigate on empty graph should not error: %v", err)
 	}
@@ -333,10 +342,11 @@ func TestEdge_TriggersFired(t *testing.T) {
 	h := testServer
 	header := map[string]string{
 		"Content-Type": "application/json",
-		"X-API-Key":    "zhiyi-dev-key-2026",
+		"X-API-Key":    testAPIKey,
 	}
 
-	body := map[string]string{"trigger_id": "t1"}
+	// 触发器 ID 与 routes.Triggers 一致（t1..t8 是 executor 的内部编号，不是 API ID）
+	body := map[string]string{"trigger_id": "t_decay"}
 	req := httptest.NewRequest("POST", "/api/v1/triggers/fire", bytesBody(body))
 	for k, v := range header {
 		req.Header.Set(k, v)

go/internal/api/cosine.go

@@ -0,0 +1,23 @@
+package api
+
+import "math"
+
+// cosineSimilarity 计算两个向量的余弦相似度。
+// 放在 api 包根目录，让 server（蒸馏写入路径）和 routes（commit 路径）
+// 共享同一套近重复判据（>=0.98），避免两套逻辑各自漂移。
+func cosineSimilarity(a, b []float32) float64 {
+	var dot, normA, normB float64
+	n := len(a)
+	if len(b) < n {
+		n = len(b)
+	}
+	for i := 0; i < n; i++ {
+		dot += float64(a[i]) * float64(b[i])
+		normA += float64(a[i]) * float64(a[i])
+		normB += float64(b[i]) * float64(b[i])
+	}
+	if normA == 0 || normB == 0 {
+		return 0
+	}
+	return dot / math.Sqrt(normA*normB)
+}

go/internal/api/middleware/auth.go

@@ -2,6 +2,8 @@
 package middleware
 
 import (
+	"crypto/rand"
+	"encoding/hex"
 	"fmt"
 	"log"
 	"net/http"
@@ -14,19 +16,19 @@ import (
 // ─── Per-Agent 令牌桶 ────────────────────────────────────────
 
 type tokenBucket struct {
-	tokens   float64
+	tokens     float64
 	lastRefill time.Time
-	rate     float64 // tokens/sec
-	burst    float64
-	mu       sync.Mutex
+	rate       float64 // tokens/sec
+	burst      float64
+	mu         sync.Mutex
 }
 
 func newTokenBucket(rate, burst float64) *tokenBucket {
 	return &tokenBucket{
-		tokens:   burst,
+		tokens:     burst,
 		lastRefill: time.Now(),
-		rate:     rate,
-		burst:    burst,
+		rate:       rate,
+		burst:      burst,
 	}
 }
 
@@ -56,10 +58,10 @@ type AgentRateConfig struct {
 }
 
 var agentRates = map[string]AgentRateConfig{
-	"hermes":    {RecallQPS: 10, CommitQPS: 2, Burst: 20},
-	"hermes-a06":{RecallQPS: 10, CommitQPS: 2, Burst: 20},
-	"openclaw":  {RecallQPS: 10, CommitQPS: 2, Burst: 20},
-	"cron-job":  {RecallQPS: 5,  CommitQPS: 1, Burst: 10},
+	"hermes":     {RecallQPS: 10, CommitQPS: 2, Burst: 20},
+	"hermes-a06": {RecallQPS: 10, CommitQPS: 2, Burst: 20},
+	"openclaw":   {RecallQPS: 10, CommitQPS: 2, Burst: 20},
+	"cron-job":   {RecallQPS: 5, CommitQPS: 1, Burst: 10},
 }
 
 var defaultRate = AgentRateConfig{RecallQPS: 5, CommitQPS: 1, Burst: 10}
@@ -67,14 +69,14 @@ var defaultRate = AgentRateConfig{RecallQPS: 5, CommitQPS: 1, Burst: 10}
 // ─── RateLimiter ─────────────────────────────────────────────
 
 type RateLimiter struct {
-	mu     sync.Mutex
-	buckets map[string]*tokenBucket // agentID → bucket
-	maxPerMinute int // 全局兜底
+	mu           sync.Mutex
+	buckets      map[string]*tokenBucket // agentID → bucket
+	maxPerMinute int                     // 全局兜底
 }
 
 func NewRateLimiter(maxPerMinute int) *RateLimiter {
 	return &RateLimiter{
-		buckets:       make(map[string]*tokenBucket),
+		buckets:      make(map[string]*tokenBucket),
 		maxPerMinute: maxPerMinute,
 	}
 }
@@ -121,29 +123,124 @@ func init() {
 
 // ─── HTTP 中间件 ─────────────────────────────────────────────
 
-// Auth 认证中间件 (X-API-Key)
+// t_4fab32c1 AC2: 密钥强制（无硬编码回落）。
+// 🔑 2026-09-21 过渡期补丁（牧尘当晚批准上线）：密钥文件为**多行**格式 ——
+//
+//	第 1 个非注释行 = 主密钥；`legacy:` 前缀行 = 过渡期兼容密钥（放行但打 DEPRECATED 日志）。
+//
+// 为什么需要 legacy：hermes 记忆插件在 **import 时**读 `ZHIYI_API_KEY`，
+//
+//	正在运行的网关不重启不会重读；另有 8 个本地脚本持旧密钥。
+//	若直接轮换值 → 瞬间打断记忆读写（recall/commit 全 401）。
+//
+// 收口条件：客户端迁移完（网关重启后）删除 key 文件里的 legacy 行。
+const keyFile = "/var/lib/memoryweave/.api-key"
+
+// parseKeys 纯解析：data=key 文件内容，envKey=环境变量值。抽出来是为了可被单测直接覆盖。
+func parseKeys(data, envKey string) (primary string, legacy []string) {
+	primary = strings.TrimSpace(envKey)
+	for _, line := range strings.Split(data, "\n") {
+		line = strings.TrimSpace(line)
+		if line == "" || strings.HasPrefix(line, "#") {
+			continue
+		}
+		if strings.HasPrefix(line, "legacy:") {
+			if k := strings.TrimSpace(strings.TrimPrefix(line, "legacy:")); k != "" {
+				legacy = append(legacy, k)
+			}
+			continue
+		}
+		if primary == "" {
+			primary = line
+		}
+	}
+	return primary, legacy
+}
+
+// loadKeys 返回 (主密钥, 过渡期兼容密钥列表)
+func loadKeys() (primary string, legacy []string) {
+	data := ""
+	if b, err := os.ReadFile(keyFile); err == nil {
+		data = string(b)
+	}
+	return parseKeys(data, os.Getenv("API_KEY"))
+}
+
+var (
+	legacyUseMu sync.Mutex
+	legacyLast  = map[string]time.Time{}
+)
+
+// noteLegacyUse 对 legacy 密钥放行做限频告警（每密钥 10 分钟一次）
+func noteLegacyUse(key string) {
+	legacyUseMu.Lock()
+	last := legacyLast[key]
+	legacyLast[key] = time.Now()
+	legacyUseMu.Unlock()
+	if time.Since(last) > 10*time.Minute {
+		shown := key
+		if len(shown) > 6 {
+			shown = shown[:6] + "…"
+		}
+		log.Printf("[auth] DEPRECATED: 放行了过渡期 legacy 密钥(%s)。请迁移调用方到主密钥，并删除 %s 的 legacy 行收口。", shown, keyFile)
+	}
+}
+
+func isLegacy(list []string, key string) bool {
+	for _, k := range list {
+		if k == key {
+			return true
+		}
+	}
+	return false
+}
+
 func Auth(next http.Handler) http.Handler {
-	apiKey := os.Getenv("API_KEY")
+	apiKey, legacyKeys := loadKeys()
 	if apiKey == "" {
-		apiKey = "zhiyi-dev-key-2026"
+		// 生成随机密钥并持久化（无任何可用密钥时拒绝启动或落盘随机密钥）
+		buf := make([]byte, 32)
+		if _, err := rand.Read(buf); err != nil {
+			log.Fatalf("[auth] API_KEY 未设且无法生成随机密钥: %v — 拒绝启动", err)
+		}
+		apiKey = hex.EncodeToString(buf)
+		if err := os.MkdirAll("/var/lib/memoryweave", 0700); err != nil {
+			log.Printf("[auth] WARNING: 已生成随机密钥但无法创建目录: %v — 密钥仅存于内存，重启后失效", err)
+		} else if err := os.WriteFile(keyFile, []byte(apiKey+"\n"), 0600); err != nil {
+			log.Printf("[auth] WARNING: 已生成随机密钥但写入 %s 失败: %v — 密钥仅存于内存，重启后失效", keyFile, err)
+		} else {
+			log.Printf("[auth] API_KEY 未设 — 已生成随机密钥并写入 %s (0600)。请同步给调用方（zhiyi-cli / hermes 等）。", keyFile)
+		}
+	}
+	if len(legacyKeys) > 0 {
+		log.Printf("[auth] 已加载 %d 个过渡期 legacy 密钥（放行+告警）；收口：迁移调用方后删除 %s 的 legacy 行", len(legacyKeys), keyFile)
 	}
 
 	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		// /health, /metrics, /static/*, / 不需要认证（/ 供 Web UI 使用）
-		if r.URL.Path == "/health" || r.URL.Path == "/metrics" ||
+		// t_4fab32c1 AC3: /metrics 纳入鉴权（旧版在免鉴权白名单内，无 key 即可
+		// 读取 episode 总数/进程 RSS/队列深度等运维指标，暴露系统内部状态）。
+		// /health 保留免鉴权（供 systemd / 监控探针使用）；
+		// /static/* 和 / 保留免鉴权（Web UI 静态资源）。
+		if r.URL.Path == "/health" ||
 			strings.HasPrefix(r.URL.Path, "/static/") || r.URL.Path == "/" {
 			next.ServeHTTP(w, r)
 			return
 		}
 
 		key := r.Header.Get("X-API-Key")
-		if key != apiKey {
+		if key != apiKey && !isLegacy(legacyKeys, key) {
+			// t_2d4e3b92: 统一错误体 schema — 非 2xx + {"error":{"code","message","retry_after_s"}}
+			// 此前为扁平 {"error":"..."}，调用方难以与成功体区分（曾导致把 401/429 误读为「召回 0 条」）。
 			w.Header().Set("Content-Type", "application/json")
 			w.WriteHeader(401)
-			w.Write([]byte(`{"error":"unauthorized: invalid or missing X-API-Key"}`))
+			w.Write([]byte(`{"error":{"code":"unauthorized","message":"invalid or missing X-API-Key","retry_after_s":0}}`))
 			return
 		}
 
+		if key != apiKey {
+			noteLegacyUse(key) // 过渡期：legacy 密钥放行但留审计痕迹
+		}
+
 		// 提取 agent ID（从 X-Agent-ID 头或 URL 路径）
 		agentID := r.Header.Get("X-Agent-ID")
 		if agentID == "" {
@@ -157,13 +254,14 @@ func Auth(next http.Handler) http.Handler {
 		}
 
 		// Per-agent 令牌桶检查
+		// t_2d4e3b92: 统一错误体 schema — 非 2xx + {"error":{"code","message","retry_after_s"}}
 		if !GlobalLimiter.Allow(agentID, endpointType) {
 			w.Header().Set("Content-Type", "application/json")
 			w.Header().Set("Retry-After", "1")
 			w.Header().Set("X-RateLimit-Reset", fmt.Sprintf("%d", time.Now().Unix()+1))
 			w.WriteHeader(429)
 			w.Write([]byte(fmt.Sprintf(
-				`{"error":"rate_limit_exceeded","agent":"%s","retry_after":1}`, agentID,
+				`{"error":{"code":"rate_limit_exceeded","message":"rate limit exceeded for agent %s","retry_after_s":1}}`, agentID,
 			)))
 			log.Printf("[ratelimit] %s exceeded for agent %s", endpointType, agentID)
 			return
@@ -200,7 +298,8 @@ func RateLimit(maxPerMinute int) func(http.Handler) http.Handler {
 				w.Header().Set("X-RateLimit-Reset", fmt.Sprintf("%d", resetTime))
 				w.Header().Set("Content-Type", "application/json")
 				w.WriteHeader(429)
-				w.Write([]byte(`{"error":"rate_limit_exceeded","retry_after":1}`))
+				// t_2d4e3b92: 统一错误体 schema
+				w.Write([]byte(`{"error":{"code":"rate_limit_exceeded","message":"global rate limit exceeded","retry_after_s":60}}`))
 				return
 			}
 

go/internal/api/middleware/auth_error_shape_test.go

@@ -0,0 +1,116 @@
+package middleware
+
+import (
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+)
+
+// t_2d4e3b92: 中间件失败路径必须「非 2xx + 统一错误体 schema」。
+//
+// 起因：401/429 曾返回扁平 {"error":"..."} 或 {"error":..,"agent":..,"retry_after":..}，
+// 与成功体形状不同却被调用方当成「空结果」处理 —— 实测把小怡的召回质量复算一度误导成
+// 「召回 0 条 = 真实回退」。本测试锁死形状契约。
+
+// validateUnifiedErrorBody 纯函数：校验错误体形状，返回问题列表（空 = 通过）。
+// 抽成纯函数是为了能做**反证**（喂扁平 body 必须报出问题），否则判据可能恒真。
+func validateUnifiedErrorBody(status int, body []byte, wantStatus int, wantCode string) []string {
+	var problems []string
+
+	if status == 200 {
+		problems = append(problems, "失败路径不得返回 200")
+	}
+	if wantStatus != 0 && status != wantStatus {
+		problems = append(problems, fmt.Sprintf("HTTP 状态码期望 %d，实际 %d", wantStatus, status))
+	}
+
+	var top map[string]json.RawMessage
+	if err := json.Unmarshal(body, &top); err != nil {
+		return append(problems, "错误体不是合法 JSON: "+err.Error())
+	}
+
+	raw, ok := top["error"]
+	if !ok {
+		return append(problems, "错误体缺少 error 键")
+	}
+
+	// 核心判别点：error 必须是**对象**，不是字符串 —— 这正是「统一 schema」的判据
+	var obj map[string]interface{}
+	if err := json.Unmarshal(raw, &obj); err != nil {
+		return append(problems, "error 必须是对象，实际是标量/数组: "+string(raw))
+	}
+	for _, k := range []string{"code", "message", "retry_after_s"} {
+		if _, ok := obj[k]; !ok {
+			problems = append(problems, fmt.Sprintf("error 缺少字段 %q", k))
+		}
+	}
+	if wantCode != "" {
+		if got, _ := obj["code"].(string); got != wantCode {
+			problems = append(problems, fmt.Sprintf("error.code 期望 %q，实际 %q", wantCode, got))
+		}
+	}
+	return problems
+}
+
+func assertUnified(t *testing.T, rec *httptest.ResponseRecorder, wantStatus int, wantCode string) {
+	t.Helper()
+	if probs := validateUnifiedErrorBody(rec.Code, rec.Body.Bytes(), wantStatus, wantCode); len(probs) > 0 {
+		t.Fatalf("错误体形状不合契约: %v | body=%s", probs, rec.Body.String())
+	}
+}
+
+// 401：缺 key
+func TestAuthMiddleware_401_UnifiedErrorBody(t *testing.T) {
+	handler := Auth(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) { w.WriteHeader(200) }))
+
+	req := httptest.NewRequest("POST", "/api/v1/recall", nil)
+	rec := httptest.NewRecorder()
+	handler.ServeHTTP(rec, req)
+
+	assertUnified(t, rec, 401, "unauthorized")
+}
+
+// 401：key 错误
+func TestAuthMiddleware_401_WrongKey_UnifiedErrorBody(t *testing.T) {
+	handler := Auth(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) { w.WriteHeader(200) }))
+
+	req := httptest.NewRequest("POST", "/api/v1/recall", nil)
+	req.Header.Set("X-API-Key", "definitely-wrong-key")
+	rec := httptest.NewRecorder()
+	handler.ServeHTTP(rec, req)
+
+	assertUnified(t, rec, 401, "unauthorized")
+}
+
+// 反证：喂历史扁平 body → 校验器**必须**报出问题（证明判据非恒真）
+func TestValidateUnifiedErrorBody_Negative(t *testing.T) {
+	cases := []struct {
+		name string
+		code int
+		body string
+	}{
+		{"扁平 error 字符串（修前 401 形状）", 401, `{"error":"unauthorized: invalid or missing X-API-Key"}`},
+		{"旧限流体（修前 429 形状）", 429, `{"error":"rate_limit_exceeded","agent":"a1","retry_after":1}`},
+		{"成功体混入（无 error 键）", 200, `{"results":[],"count":0}`},
+		{"error 对象缺字段", 502, `{"error":{"code":"x"}}`},
+	}
+	for _, c := range cases {
+		t.Run(c.name, func(t *testing.T) {
+			probs := validateUnifiedErrorBody(c.code, []byte(c.body), 0, "")
+			if len(probs) == 0 {
+				t.Fatalf("反证失败：%s 未被判错 → 判据恒真", c.body)
+			}
+			t.Logf("已判错（符合预期）: %v", probs)
+		})
+	}
+}
+
+// 正例：统一 schema 必须通过（防止校验器过严，把正确形状也判错）
+func TestValidateUnifiedErrorBody_Positive(t *testing.T) {
+	ok := `{"error":{"code":"unauthorized","message":"bad key","retry_after_s":0}}`
+	if probs := validateUnifiedErrorBody(401, []byte(ok), 401, "unauthorized"); len(probs) > 0 {
+		t.Fatalf("统一 schema 被误判: %v", probs)
+	}
+}

go/internal/api/middleware/auth_keyfile_test.go

@@ -0,0 +1,41 @@
+package middleware
+
+import (
+	"testing"
+)
+
+// 行为契约（2026-09-21 过渡期密钥补丁）：
+// key 文件多行格式 —— 注释行忽略；第 1 个非注释非 legacy 行 = 主密钥；`legacy:` 行 = 过渡期兼容。
+// 若退化会出现两种生产事故之一：①主密钥读成整份文件（所有客户端 401，记忆读写全断）；
+// ②legacy 行被当成主密钥（等于没轮换）。
+// 直接覆盖生产的 parseKeys（不是复刻实现）。
+func TestParseKeysMultiLineFormat(t *testing.T) {
+	content := "# 注释行\nPRIMARYKEY123\nlegacy:oldkey456\n\n# 再来一行注释\n"
+
+	primary, legacy := parseKeys(content, "")
+	if primary != "PRIMARYKEY123" {
+		t.Fatalf("主密钥应为 PRIMARYKEY123，实际 %q", primary)
+	}
+	if len(legacy) != 1 || legacy[0] != "oldkey456" {
+		t.Fatalf("legacy 应为 [oldkey456]，实际 %v", legacy)
+	}
+
+	if p2, _ := parseKeys(content, "ENVKEY"); p2 != "ENVKEY" {
+		t.Fatalf("env 变量应优先于文件，实际 %q", p2)
+	}
+
+	if !isLegacy(legacy, "oldkey456") {
+		t.Fatal("isLegacy 应识别过渡密钥")
+	}
+	if isLegacy(legacy, "PRIMARYKEY123") {
+		t.Fatal("主密钥不应被判为 legacy")
+	}
+	if isLegacy(legacy, "") {
+		t.Fatal("空 key 绝不可被判为 legacy（否则等于无鉴权）")
+	}
+
+	// 只有注释的损坏文件 → 无主密钥（上层走"生成随机密钥"分支，而不是放行空 key）
+	if p3, l3 := parseKeys("# only comments\n", ""); p3 != "" || len(l3) != 0 {
+		t.Fatalf("纯注释文件应解析为空，实际 primary=%q legacy=%v", p3, l3)
+	}
+}

go/internal/api/middleware/cors.go

@@ -24,4 +24,4 @@ func CORS() func(http.Handler) http.Handler {
 			next.ServeHTTP(w, r)
 		})
 	}
-}
+}

go/internal/api/routes/admin.go

@@ -5,6 +5,7 @@ import (
 	"context"
 	"encoding/json"
 	"fmt"
+	"log"
 	"net/http"
 	"os"
 	"os/exec"
@@ -13,6 +14,7 @@ import (
 	"time"
 
 	"github.com/xiaoxue/memoryweave/internal/governance"
+	"github.com/xiaoxue/memoryweave/internal/memprotect"
 	"github.com/xiaoxue/memoryweave/internal/storage"
 )
 
@@ -66,22 +68,87 @@ func (aa *AdminAPI) Forget(w http.ResponseWriter, r *http.Request) {
 	}
 
 	forgotten := 0
+	// 2026-10-16 (t_944344b5) 遗忘保护守卫：本批加载一次保护状态（Redis vprop + protected_ids），
+	// 命中任一保护规则的候选永不自动软删（fragment 与 ShouldForget 两条道都拦截）。
+	protState := memprotect.LoadState()
+	protected := 0
 	for _, mem := range results {
 		lastAccess := parseTimeStr(mem["last_recalled_at"])
-		recallCnt := intVal(mem["recall_count"])
+		// GA-P0-2 (t_c6d41ac6)：衰减/遗忘只认 recall_verified_count（正反馈才是 V 值）。
+		// recall_count 已冻结（历史虚高保命来源），seen 只用于观测，不参与衰减。
+		recallCnt := intVal(mem["recall_verified_count"])
 		tier := strVal(mem["tier"])
 		content := strVal(mem["content"])
 
+		// 🔒 2026-09-07 修复(误删事故): episodes(原始对话)不参与遗忘
+		if cat := strVal(mem["category"]); cat == "episodes" {
+			continue
+		}
+		// 🔒 长内容保护: >200 字记忆不参与 auto 遗忘（有实质信息）
+		if len([]rune(content)) > 200 {
+			continue
+		}
+		// 🔒 2026-10-16 (t_944344b5) 遗忘保护守卫: 命中任一保护规则的候选
+		// (category∈{decision,preference,pref} / V≥2.0 / freshness=verified /
+		//  protected_ids 名单 / distilled 且内容命中 10 个决策·偏好关键词) 永不自动软删。
+		// 拦截 point 放在 fragment + ShouldForget 两条 SoftDelete 道之前。
+		if isProt, protRule := memprotect.IsProtected(mem, protState); isProt {
+			protected++
+			log.Printf("[forget] protected skip id=%q rule=%s (遗忘保护守卫命中, 跳过自动遗忘)", strVal(mem["id"]), protRule)
+			continue
+		}
+		// 2026-09-07 方案A 碎片快速道（与 server.go decay trigger 一致）:
+		// <30 字且 >20 天未访问 → 直接遗忘（碎片 recall_count 可能虚高, 绕过保命）
+		if len([]rune(content)) < 30 && time.Since(lastAccess).Hours()/24 > 20 {
+			// 🔴 2026-10-06：空 id 直接跳过（DB 里有 id/content 全空的垃圾行会命中这条快速道）；
+			// 且 SoftDelete 失败不再计数 —— 原实现失败照样 forgotten++ ⇒ 遗忘计数虚高。
+			id := strVal(mem["id"])
+			if id == "" {
+				continue
+			}
+			if err := aa.LanceDB.SoftDelete(id, "auto_forget_fragment"); err != nil {
+				log.Printf("[softdelete] failed id=%q: %v", id, err)
+				continue
+			}
+			forgotten++
+			continue
+		}
+
 		// E4.3 修复: 从 content 提取实体（而非读 namespace），取图谱最大度
 		degree := extractTopEntityDegree(content, aa.GraphStore)
 
 		if aa.Forgetter.ShouldForget(lastAccess, recallCnt, tier, degree) {
-			aa.LanceDB.SoftDelete(strVal(mem["id"]), "auto_forget")
+			id := strVal(mem["id"])
+			if id == "" {
+				continue
+			}
+			if err := aa.LanceDB.SoftDelete(id, "auto_forget"); err != nil {
+				log.Printf("[softdelete] failed id=%q: %v", id, err)
+				continue
+			}
 			forgotten++
 		}
 	}
 	respond(w, 200, map[string]interface{}{
 		"status": "ok", "scanned": len(results), "forgotten": forgotten,
+		// 2026-10-16 (t_944344b5)：protected = 被遗忘保护守卫拦截、未软删的候选数
+		"protected": protected,
+	})
+}
+
+// POST /api/v1/admin/purge-invalid
+// P1-1 2026-10-08: 一次性清理 id 为空的垃圾行（DB 存在 id/content/source 全空的 2 行，
+// 命中 <30 字碎片快速道 → SoftDelete("") 被 Rust 拒绝 → 每轮刷错误日志 + 遗忘计数虚高）。
+// 这些行无 id 无法按 id 软删，只能按 id=” 条件一次性软删（is_deleted=true）。
+// ⚠️ 由验收/决策者按需调用（本实现不自动触发），本端点只提供能力 + 返回受影响行数。
+func (aa *AdminAPI) PurgeInvalid(w http.ResponseWriter, r *http.Request) {
+	n, err := aa.LanceDB.PurgeInvalid()
+	if err != nil {
+		respondError(w, 500, "purge-invalid failed: "+err.Error())
+		return
+	}
+	respond(w, 200, map[string]interface{}{
+		"status": "ok", "purged": n,
 	})
 }
 
@@ -241,9 +308,9 @@ func (aa *AdminAPI) Restore(w http.ResponseWriter, r *http.Request) {
 	aa.startServices()
 
 	respond(w, 200, map[string]string{
-		"status":    "ok",
-		"restored":  req.Timestamp,
-		"data_dir":  dataDir,
+		"status":   "ok",
+		"restored": req.Timestamp,
+		"data_dir": dataDir,
 	})
 }
 
@@ -291,10 +358,14 @@ func parseTimeStr(s interface{}) time.Time {
 
 func intVal(v interface{}) int {
 	switch n := v.(type) {
-	case int:     return n
-	case int32:   return int(n)
-	case int64:   return int(n)
-	case float64: return int(n)
+	case int:
+		return n
+	case int32:
+		return int(n)
+	case int64:
+		return int(n)
+	case float64:
+		return int(n)
 	case json.Number:
 		i, _ := n.Int64()
 		return int(i)
@@ -307,8 +378,10 @@ func strVal(v interface{}) string {
 		return ""
 	}
 	switch s := v.(type) {
-	case string:      return s
-	case json.Number:  return s.String()
+	case string:
+		return s
+	case json.Number:
+		return s.String()
 	}
 	return fmt.Sprintf("%v", v)
 }
@@ -422,7 +495,9 @@ func isAllDigits(s string) bool {
 	return true
 }
 
-// GET /api/v1/memories — 批量列出记忆（按 importance 排序，不含向量）
+// GET /api/v1/memories — 批量列出记忆（按时间倒序，不含向量）
+// 2026-09-21：改用全表扫描（aa.LanceDB.Scan）替代「零向量 ANN 冒充列表」。
+// 修前零向量无方向 ⇒ 只拿到存储序前 N 行再按 ns 过滤 ⇒ 列表/计数失真。
 func (aa *AdminAPI) ListMemories(w http.ResponseWriter, r *http.Request) {
 	limit := 100
 	ns := r.URL.Query().Get("namespace")
@@ -434,20 +509,19 @@ func (aa *AdminAPI) ListMemories(w http.ResponseWriter, r *http.Request) {
 			limit = 1000
 		}
 	}
-	zeroVec := make([]float32, 1024)
-	results, err := aa.LanceDB.Search("memories", zeroVec, limit, ns)
+	results, err := aa.LanceDB.Scan(limit, ns)
 	if err != nil {
 		respondError(w, 500, "list memories: "+err.Error())
 		return
 	}
 	// 去掉向量字段，减少响应体积
 	type memoryItem struct {
-		ID        string `json:"id"`
-		Content   string `json:"content"`
-		Category  string `json:"category"`
-		Namespace string `json:"namespace,omitempty"`
+		ID         string  `json:"id"`
+		Content    string  `json:"content"`
+		Category   string  `json:"category"`
+		Namespace  string  `json:"namespace,omitempty"`
 		Importance float64 `json:"importance"`
-		CreatedAt int64   `json:"created_at"`
+		CreatedAt  int64   `json:"created_at"`
 	}
 	items := make([]memoryItem, 0, len(results))
 	for _, m := range results {
@@ -476,8 +550,8 @@ func (aa *AdminAPI) ExportMemoriesMD(w http.ResponseWriter, r *http.Request) {
 			limit = 5000
 		}
 	}
-	zeroVec := make([]float32, 1024)
-	results, err := aa.LanceDB.Search("memories", zeroVec, limit, ns)
+	// 2026-09-21：改用全表扫描（aa.LanceDB.Scan）替代「零向量 ANN 冒充列表」。
+	results, err := aa.LanceDB.Scan(limit, ns)
 	if err != nil {
 		respondError(w, 500, "export memories: "+err.Error())
 		return
@@ -512,4 +586,4 @@ func orDefault(s, def string) string {
 		return def
 	}
 	return s
-}
+}

go/internal/api/routes/admin_forget_test.go

@@ -0,0 +1,161 @@
+// 织忆 MemoryWeave — P1-1 遗忘路径计数修复单测（t_37ac8a45, 2026-10-08）
+//
+// 目标（AC3）：SoftDelete 失败 → forgotten 不得增加（原实现失败照样 forgotten++ ⇒ 遗忘计数虚高）。
+// 量具铁律：正例（成功 → forgotten+1）与负例（失败 → forgotten 不 +1）都要有。
+package routes
+
+import (
+	"encoding/json"
+	"fmt"
+	"net/http/httptest"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/xiaoxue/memoryweave/internal/governance"
+	"github.com/xiaoxue/memoryweave/internal/models"
+)
+
+// forgetMockLance 为 Forget 计数测试提供的 LanceDB 桩：
+// GetCandidatesForForgetting 可配置候选列表；SoftDelete 可配置失败/成功。
+type forgetMockLance struct {
+	candidates    []map[string]interface{}
+	softDeleteErr error
+	deletedIDs    []string
+}
+
+func (m *forgetMockLance) InsertEpisode(agentID, namespace, content, category string) (string, error) {
+	return "ep_test", nil
+}
+func (m *forgetMockLance) GetTopByQuality(agentID string, limit int) ([]models.MemoryRecord, error) {
+	return nil, nil
+}
+func (m *forgetMockLance) Stats() (map[string]interface{}, error) {
+	return map[string]interface{}{}, nil
+}
+func (m *forgetMockLance) SoftDelete(id, reason string) error {
+	if m.softDeleteErr != nil {
+		return m.softDeleteErr
+	}
+	m.deletedIDs = append(m.deletedIDs, id)
+	return nil
+}
+func (m *forgetMockLance) PurgeInvalid() (int64, error) { return 0, nil }
+func (m *forgetMockLance) SoftDeleteBatch(table string, ids []string, updatedAt string) (int64, error) {
+	return int64(len(ids)), nil
+}
+func (m *forgetMockLance) GetVersionHistory(id string) ([]map[string]interface{}, error) {
+	return nil, nil
+}
+func (m *forgetMockLance) GetCandidatesForForgetting() ([]map[string]interface{}, error) {
+	return m.candidates, nil
+}
+func (m *forgetMockLance) GetSkillCandidates(minRecalls, limit int) ([]models.MemoryRecord, error) {
+	return nil, nil
+}
+func (m *forgetMockLance) Backup(path string) error                                { return nil }
+func (m *forgetMockLance) GetAuditLog(limit int) ([]map[string]interface{}, error) { return nil, nil }
+func (m *forgetMockLance) IncrementUseful(id string) error                         { return nil }
+func (m *forgetMockLance) IncrementNotUseful(id string) error                      { return nil }
+
+// GA-P0-1 (t_718d8707)：拆分字段
+func (m *forgetMockLance) IncrementHumanNotUseful(id string) error                 { return nil }
+func (m *forgetMockLance) AppendHitEvent(memoryID, outcome, reason string) error   { return nil }
+func (m *forgetMockLance) CountHitEvents(memoryID, outcome string, since time.Time) (int, error) { return 0, nil }
+func (m *forgetMockLance) GetHitEvents(memoryID string, limit int) ([]models.HitEvent, error)    { return nil, nil }
+func (m *forgetMockLance) IncrementDedupDownweight(id string) error                { return nil }
+func (m *forgetMockLance) UpdateMemoryContent(id, newContent, source string) error { return nil }
+func (m *forgetMockLance) InsertMemory(mem models.MemoryRecord) error              { return nil }
+func (m *forgetMockLance) Search(table string, vector []float32, topK int, namespaceFilter string) ([]models.MemoryRecord, error) {
+	return nil, nil
+}
+func (m *forgetMockLance) Scan(limit int, namespaceFilter string) ([]models.MemoryRecord, error) {
+	return nil, nil
+}
+func (m *forgetMockLance) Insert(table string, record any) error { return nil }
+func (m *forgetMockLance) Update(table, id string, fields map[string]any) error {
+	return nil
+}
+
+// fragment 构造一个命中「<30 字碎片快速道」的候选（content<30 字 + 20 天未访问）。
+func fragmentCandidate(id string) map[string]interface{} {
+	return map[string]interface{}{
+		"id":               id,
+		"content":          "短碎片", // <30 字
+		"category":         "distilled",
+		"last_recalled_at": time.Now().Add(-30 * 24 * time.Hour).Format(time.RFC3339),
+		"recall_count":     0,
+		"tier":             "normal",
+	}
+}
+
+// forgetResp 解析 Forget 响应体。
+type forgetResp struct {
+	Status    string `json:"status"`
+	Scanned   int    `json:"scanned"`
+	Forgotten int    `json:"forgotten"`
+}
+
+// callForget 构造 AdminAPI 并调用 Forget 端点，返回响应体。
+func callForget(t *testing.T, db *forgetMockLance) forgetResp {
+	t.Helper()
+	aa := NewAdminAPI(db, governance.NewForgetter(), governance.NewInMemoryGraph())
+	req := httptest.NewRequest("POST", "/api/v1/admin/forget", strings.NewReader(""))
+	w := httptest.NewRecorder()
+	aa.Forget(w, req)
+	if w.Code != 200 {
+		t.Fatalf("Forget 期望 200，实得 %d：%s", w.Code, w.Body.String())
+	}
+	var resp forgetResp
+	if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("解析响应失败: %v", err)
+	}
+	return resp
+}
+
+// TestForget_SoftDeleteFailure_DoesNotIncrement 负例（AC3 核心）：
+// 有效 id 的候选命中碎片快速道，但 SoftDelete 失败 → forgotten 不得增加。
+// 修前：失败照样 forgotten++ ⇒ 遗忘计数虚高（每轮刷日志 + 计数失真）。
+func TestForget_SoftDeleteFailure_DoesNotIncrement(t *testing.T) {
+	db := &forgetMockLance{
+		candidates:    []map[string]interface{}{fragmentCandidate("mem_fail_1")},
+		softDeleteErr: fmt.Errorf("ipc: simulated soft delete failure"),
+	}
+	resp := callForget(t, db)
+	if resp.Forgotten != 0 {
+		t.Fatalf("SoftDelete 失败时 forgotten 不得增加，实得 %d（计数 bug 未修）", resp.Forgotten)
+	}
+	t.Logf("✓ SoftDelete 失败 → forgotten=%d（不虚增）", resp.Forgotten)
+}
+
+// TestForget_SoftDeleteSuccess_Increments 正例（量具铁律：证明量具能报阳性）：
+// 有效 id 候选 + SoftDelete 成功 → forgotten 必须 +1。
+func TestForget_SoftDeleteSuccess_Increments(t *testing.T) {
+	db := &forgetMockLance{
+		candidates: []map[string]interface{}{fragmentCandidate("mem_ok_1")},
+	}
+	resp := callForget(t, db)
+	if resp.Forgotten != 1 {
+		t.Fatalf("SoftDelete 成功时 forgotten 应 +1，实得 %d", resp.Forgotten)
+	}
+	if len(db.deletedIDs) != 1 || db.deletedIDs[0] != "mem_ok_1" {
+		t.Fatalf("SoftDelete 应被调用一次并传入 mem_ok_1，实得 %v", db.deletedIDs)
+	}
+	t.Logf("✓ SoftDelete 成功 → forgotten=%d（计数正确）", resp.Forgotten)
+}
+
+// TestForget_EmptyIDSkipped 空 id 候选直接跳过（Step 2a 守卫）：
+// id 为空（id/content/source 全空的垃圾行）→ 不计数、不调用 SoftDelete。
+func TestForget_EmptyIDSkipped(t *testing.T) {
+	db := &forgetMockLance{
+		candidates: []map[string]interface{}{fragmentCandidate("")},
+	}
+	resp := callForget(t, db)
+	if resp.Forgotten != 0 {
+		t.Fatalf("空 id 候选不得计入 forgotten，实得 %d", resp.Forgotten)
+	}
+	if len(db.deletedIDs) != 0 {
+		t.Fatalf("空 id 候选不得调用 SoftDelete，实得 %v", db.deletedIDs)
+	}
+	t.Logf("✓ 空 id 候选跳过：forgotten=%d, SoftDelete 调用 0 次", resp.Forgotten)
+}

go/internal/api/routes/auto_distill.go

@@ -2,18 +2,41 @@
 package routes
 
 import (
+	"log"
 	"time"
 
 	"github.com/xiaoxue/memoryweave/internal/distill"
 	"github.com/xiaoxue/memoryweave/internal/governance"
+	"github.com/xiaoxue/memoryweave/internal/metanote"
 	"github.com/xiaoxue/memoryweave/internal/selfoptimize"
 )
 
 // DistillEngineRef 全局蒸馏引擎引用（server.go 注入）
 var DistillEngineRef *distill.Engine
 
+// CommitMeta 2026-10-08 (Gap#2)：commit 时写入方携带的元数据，透传到蒸馏产物 MemoryRecord。
+// 集中成一个结构体而非继续加位置参数——8 个位置参数再加 2 个时，「忘记传」是必然的
+// （Gap#2 的接线缺口正是这类 bug：字段加了、没人传）。
+type CommitMeta struct {
+	Source           string   // 出处（笔记路径等）→ MemoryRecord.Source
+	EmbeddingModel   string   // 血缘 → MemoryRecord.EmbeddingModel
+	ScriptVersion    string   // 血缘 → MemoryRecord.ScriptVersion
+	Confidence       *float64 // 信任等级（写入方显式声明，不从 source 反推）；nil=未声明
+	ConfidenceSource string   // authoritative | estimated
+}
+
 // AutoDistillTrigger commit 后自动触发蒸馏流水线
-func AutoDistillTrigger(episodeID string, content string, category string, namespace string, agentID string) {
+// 2026-10-08 (t_7e3577c5)：新增 embedModel/scriptVer 血缘字段透传
+// 2026-10-08 (Gap#2)：改为 CommitMeta 结构体，新增 confidence 透传
+func AutoDistillTrigger(episodeID string, content string, category string, namespace string, agentID string, meta CommitMeta) {
+	// 2026-10-16 t_5ad3f843：meta-note 门禁（env ZHIYI_META_FILTER=1 启用；默认关闭=零行为变化）。
+	// 蒸馏是事实写入的唯一咽喉：原文判 meta ⇒ 不进蒸馏（引擎路径与降级路径都挡），
+	// 从源头杜绝 meta-note 生成 memories 事实（admin distill/force 也经此函数，一并覆盖）。
+	if skip, reason := metanote.ShouldSkip(content); skip {
+		log.Printf("[metanote] auto-distill skipped: ep=%s agent=%s reason=%s content=%.60s", episodeID, agentID, reason, content)
+		return
+	}
+
 	// 1. 硬规则过滤
 	if !distill.HardRulesPass(content) {
 		return
@@ -22,11 +45,16 @@ func AutoDistillTrigger(episodeID string, content string, category string, names
 	// 2. 使用蒸馏引擎
 	if DistillEngineRef != nil {
 		DistillEngineRef.Enqueue(distill.DistillInput{
-			EpisodeID: episodeID,
-			Content:   content,
-			Category:  distill.Category(category),
-			Namespace: namespace,
-			AgentID:   agentID,
+			EpisodeID:        episodeID,
+			Content:          content,
+			Category:         distill.Category(category),
+			Namespace:        namespace,
+			AgentID:          agentID,
+			Source:           meta.Source,
+			EmbeddingModel:   meta.EmbeddingModel,
+			ScriptVersion:    meta.ScriptVersion,
+			Confidence:       meta.Confidence,
+			ConfidenceSource: meta.ConfidenceSource,
 		})
 	} else {
 		// 降级: 关键词提取
@@ -36,6 +64,7 @@ func AutoDistillTrigger(episodeID string, content string, category string, names
 			Facts:     facts,
 			Entities:  extractEntities(content),
 			Namespace: namespace,
+			Source:    meta.Source, // 2026-10-12 t_67d0279a 方案A：降级路径同样透传来源分档
 		})
 	}
 
@@ -89,8 +118,8 @@ func currentUnixMS() int64 {
 // ─── 全局引用（server.go 需要访问）─────────────────────
 
 var (
-	graphUpdater     *governance.AutoGraphUpdater
-	conflictScanner  = governance.NewConflictDetector()
+	graphUpdater    *governance.AutoGraphUpdater
+	conflictScanner = governance.NewConflictDetector()
 )
 
 func SetGraphUpdater(gu *governance.AutoGraphUpdater) {
@@ -98,6 +127,8 @@ func SetGraphUpdater(gu *governance.AutoGraphUpdater) {
 }
 
 func minLen2(a, b int) int {
-	if a < b { return a }
+	if a < b {
+		return a
+	}
 	return b
 }

go/internal/api/routes/client.go

@@ -2,9 +2,10 @@
 // Hermes / OpenClaw 的客户端库，零外部依赖
 //
 // 使用:
-//   client := NewClient("http://localhost:7821", "mw-your-api-key")
-//   mid, err := client.Commit("hermes", "shared", "牧尘用 Deepin 25", "system_fact")
-//   results, err := client.Recall("牧尘系统是什么", 5)
+//
+//	client := NewClient("http://localhost:7821", "mw-your-api-key")
+//	mid, err := client.Commit("hermes", "shared", "牧尘用 Deepin 25", "system_fact")
+//	results, err := client.Recall("牧尘系统是什么", 5)
 package routes
 
 import (

go/internal/api/routes/conflicts.go

@@ -36,7 +36,7 @@ func (ca *ConflictAPI) Resolve(w http.ResponseWriter, r *http.Request) {
 	var req struct {
 		ConflictID string `json:"conflict_id"`
 		Resolution string `json:"resolution"` // keep_left / keep_right / merge / dismiss
-		Winner     string `json:"winner"`      // id of the winning memory
+		Winner     string `json:"winner"`     // id of the winning memory
 	}
 	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
 		respondError(w, 400, "invalid body")

go/internal/api/routes/consolidation_pipe.go

@@ -4,6 +4,7 @@ package routes
 import (
 	"fmt"
 	"log"
+	"sort"
 	"strings"
 	"time"
 
@@ -23,7 +24,7 @@ type ConsolidationPipeline struct {
 	graphUpdater *governance.AutoGraphUpdater
 	conflicts    *governance.ConflictDetector
 	// Rust IPC 路径
-	dataDir   string
+	dataDir    string
 	sqlitePath string
 }
 
@@ -112,12 +113,15 @@ func (cp *ConsolidationPipeline) RunWithMode(mode string) (*ConsolidationReport,
 	}
 
 	// Phase F: 整合完成后更新总记忆数指标（通过 LanceDB stats）
+	// total_episodes 来自 Rust count_rows（真实值）；旧 sidecar 回退 episodes_cached（非总量）
 	if stats, err := cp.ldb.Stats(); err == nil {
 		if total, ok := stats["total_memories"].(float64); ok {
 			metrics.TotalMemories.Set(total)
 		}
 		if total, ok := stats["total_episodes"].(float64); ok {
 			metrics.TotalEpisodes.Set(total)
+		} else if cached, ok := stats["episodes_cached"].(float64); ok {
+			metrics.TotalEpisodes.Set(cached) // 旧 sidecar 回退，非真实总量
 		}
 	}
 	metrics.SyncFromDashboard(selfoptimize.Dash.Metrics())
@@ -207,7 +211,10 @@ func (cp *ConsolidationPipeline) RunMerge() (int, error) {
 	return cp.mergeSimilar()
 }
 
-// mergeSimilar 合并相似记忆（向量相似度 > 0.8 → 保留最新）
+// mergeSimilar 合并相似记忆（相同 category + 高内容重叠 → 保留较新）
+// 2026-09-06 P1 fix: 原实现双层循环全量 O(n²) —— febc2c9 全表扫描时触发 620% CPU 风暴。
+// 改为按 category 分桶后桶内两两比较（原逻辑 catI != catJ continue 本就等价），
+// 单桶超 maxMergePerBucket 截断，总计算量 O(Σ桶²) << O(n²)。
 func (cp *ConsolidationPipeline) mergeSimilar() (int, error) {
 	// 获取全部 distilled 记忆
 	memories, err := cp.ldb.GetCandidatesForForgetting()
@@ -215,23 +222,36 @@ func (cp *ConsolidationPipeline) mergeSimilar() (int, error) {
 		return 0, err
 	}
 
+	const maxMergePerBucket = 500
 	merged := 0
-	// 简单启发式：相同 category + 高内容重叠 → 合并
-	for i := 0; i < len(memories); i++ {
-		for j := i + 1; j < len(memories); j++ {
-			catI := strVal(memories[i]["category"])
-			catJ := strVal(memories[j]["category"])
-			if catI != catJ {
-				continue
-			}
-			contentI := strVal(memories[i]["content"])
-			contentJ := strVal(memories[j]["content"])
-			if overlap := contentOverlap(contentI, contentJ); overlap > 0.8 {
-				// 保留较新的
-				idI := strVal(memories[i]["id"])
-				idJ := strVal(memories[j]["id"])
-				_ = cp.ldb.SoftDelete(idJ, fmt.Sprintf("merged_into_%s", idI))
-				merged++
+
+	// 按 category 分桶
+	buckets := make(map[string][]map[string]interface{})
+	for _, m := range memories {
+		cat := strVal(m["category"])
+		buckets[cat] = append(buckets[cat], m)
+	}
+
+	// 桶内两两比较（同桶才可能合并，语义与原全量双层一致）
+	for _, mems := range buckets {
+		if len(mems) > maxMergePerBucket {
+			// 单桶爆炸防护：只处理最近 maxMergePerBucket 条（created_at 倒序）
+			sort.SliceStable(mems, func(a, b int) bool {
+				return memoryInt64Val(mems[a], "created_at") > memoryInt64Val(mems[b], "created_at")
+			})
+			mems = mems[:maxMergePerBucket]
+		}
+		for i := 0; i < len(mems); i++ {
+			for j := i + 1; j < len(mems); j++ {
+				contentI := strVal(mems[i]["content"])
+				contentJ := strVal(mems[j]["content"])
+				if overlap := contentOverlap(contentI, contentJ); overlap > 0.8 {
+					// 保留较新的
+					idI := strVal(mems[i]["id"])
+					idJ := strVal(mems[j]["id"])
+					_ = cp.ldb.SoftDelete(idJ, fmt.Sprintf("merged_into_%s", idI))
+					merged++
+				}
 			}
 		}
 	}
@@ -239,12 +259,21 @@ func (cp *ConsolidationPipeline) mergeSimilar() (int, error) {
 }
 
 // Step 2: 扫描所有待解决冲突
+// 2026-09-06 P1 fix: 原注释"最近 100 条"但实际对全量做实体提取+对比 → 全表时 O(n·50)。
+// 显式按 created_at 取最近 maxConflictScan 条再扫。
 func (cp *ConsolidationPipeline) scanConflicts() (int, error) {
-	// 获取最近 100 条记忆
+	// 获取候选并按最近优先截断（created_at 倒序）
 	memories, err := cp.ldb.GetCandidatesForForgetting()
 	if err != nil {
 		return 0, err
 	}
+	const maxConflictScan = 500
+	if len(memories) > maxConflictScan {
+		sort.SliceStable(memories, func(a, b int) bool {
+			return memoryInt64Val(memories[a], "created_at") > memoryInt64Val(memories[b], "created_at")
+		})
+		memories = memories[:maxConflictScan]
+	}
 
 	count := 0
 	for i := 0; i < len(memories); i++ {
@@ -305,17 +334,28 @@ func (cp *ConsolidationPipeline) updateGraph() (int, error) {
 	return before - after, nil
 }
 
-// checkTimestampSanity 抽样检查记忆时间戳合理性（5% sampling，不阻塞）
+// checkTimestampSanity 抽样检查记忆时间戳合理性（均匀抽样，不阻塞）
 // 返回 (可疑数量, 抽样总数)。epoch-0（< 2024-01-01）视为可疑。
+// 2026-09-06 P1 fix: 原实现全量拉取只为 5% 抽样 → 先截断最近 maxSanitySample
+// 条再做均匀 step 抽样，避免全量内存/IO 与排序成本。
 func (cp *ConsolidationPipeline) checkTimestampSanity() (suspicious, total int) {
 	const epochThreshold int64 = 1704067200 // 2024-01-01 00:00:00 UTC
+	const maxSanitySample = 1000
 
 	memories, err := cp.ldb.GetCandidatesForForgetting()
 	if err != nil || len(memories) == 0 {
 		return 0, 0
 	}
 
-	// 全量不超过 100 条
+	// 超过 maxSanitySample 时按 created_at 取最新，再均匀抽样
+	if len(memories) > maxSanitySample {
+		sort.SliceStable(memories, func(a, b int) bool {
+			return memoryInt64Val(memories[a], "created_at") > memoryInt64Val(memories[b], "created_at")
+		})
+		memories = memories[:maxSanitySample]
+	}
+
+	// 抽样最多 100 条
 	sampleSize := len(memories)
 	if sampleSize > 100 {
 		sampleSize = 100
@@ -404,7 +444,8 @@ func extractEntities(text string) []string {
 }
 
 func minInt2(a, b int) int {
-	if a < b { return a }
+	if a < b {
+		return a
+	}
 	return b
 }
-

go/internal/api/routes/deep_think.go

@@ -0,0 +1,228 @@
+// 织忆 MemoryWeave — 深度推理端点（人侧数据 → 三类清单）
+package routes
+
+import (
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"strconv"
+	"time"
+
+	"github.com/xiaoxue/memoryweave/internal/governance"
+)
+
+// DeepThinkRequest 深度推理请求
+type DeepThinkRequest struct {
+	AgentID string `json:"agent_id"`
+	// namespace 为空时取人侧默认命名空间
+	Namespace string `json:"namespace,omitempty"`
+	// 最大社区数（摘要开销控制）
+	MaxCommunities int `json:"max_communities,omitempty"`
+}
+
+// DeepThinkResponse 三类清单
+type DeepThinkResponse struct {
+	OK               bool            `json:"ok"`
+	AgentID          string          `json:"agent_id"`
+	Namespace        string          `json:"namespace"`
+	Profiles         []DeepThinkItem `json:"profiles"` // 画像规则（待拍板）
+	Episodes         []DeepThinkItem `json:"episodes"` // 心迹/牵挂/承诺句（已处置）
+	Signals          []DeepThinkItem `json:"signals"`  // 社区摘要（信号）
+	ConflictsPending int             `json:"conflicts_pending"`
+	LastRun          string          `json:"last_run"`
+}
+
+// DeepThinkItem 清单条目
+type DeepThinkItem struct {
+	ID         string `json:"id"`
+	Category   string `json:"category"`
+	Content    string `json:"content"`
+	Weight     string `json:"weight,omitempty"` // 社区摘要用
+	Source     string `json:"source"`           // memories | graph | community
+	RecallOnly bool   `json:"recall_only"`      // 仅 recall 时注入，不持久化
+}
+
+// HandleDeepThink POST /api/v1/deep-think
+func (a *API) HandleDeepThink(w http.ResponseWriter, r *http.Request) {
+	var req DeepThinkRequest
+	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
+		respondError(w, 400, "decode request: "+err.Error())
+		return
+	}
+	if req.AgentID == "" {
+		respondError(w, 400, "agent_id required")
+		return
+	}
+	ns := req.Namespace
+	if ns == "" {
+		// 人侧数据统一命名空间（牧尘拍板：soulful-xiaowei，deriveNamespace 会自动加 -main 后缀 → soulful-xiaowei-main）
+		ns = deriveNamespace("soulful-xiaowei")
+	}
+	maxC := req.MaxCommunities
+	if maxC == 0 {
+		maxC = 10
+	}
+
+	now := time.Now().UTC()
+	resp := DeepThinkResponse{
+		OK:        true,
+		AgentID:   req.AgentID,
+		Namespace: ns,
+		LastRun:   now.Format(time.RFC3339),
+	}
+
+	// 1. 画像 → category=profiles
+	resp.Profiles = a.fetchCategory(ns, "profiles", 20, false)
+	// 2. 心迹/牵挂/承诺句 → category=episodes
+	episodes := a.fetchCategory(ns, "episodes", 30, false)
+	episodes = append(episodes, a.fetchCategory(ns, "cares", 15, false)...)
+	resp.Episodes = episodes
+
+	// 3. 冲突（待裁决）→ 待拍板
+	resp.ConflictsPending = 0
+	if a.ConflictDetector != nil {
+		for _, c := range a.ConflictDetector.ListActive() {
+			if c.Status == "pending" {
+				resp.ConflictsPending++
+				resp.Profiles = append(resp.Profiles, DeepThinkItem{
+					ID: c.ID, Category: "conflict", Content: c.Description,
+					Source: "conflict_detector", RecallOnly: false,
+				})
+			}
+		}
+	}
+
+	// 4. 图谱社区 → 信号
+	cds := governance.NewCommunityDiscovery(a.GraphStore)
+	communities, err := cds.Find(ns, 3)
+	if err == nil && len(communities) > 0 {
+		if len(communities) > maxC {
+			communities = communities[:maxC]
+		}
+		summaries, _ := cds.Summarize(communities, ns, 5)
+		for _, c := range summaries {
+			resp.Signals = append(resp.Signals, DeepThinkItem{
+				ID:         c.ID,
+				Category:   "community",
+				Content:    c.Summary,
+				Weight:     strconv.FormatFloat(c.Weight, 'f', 2, 64),
+				Source:     "community_discovery",
+				RecallOnly: false,
+			})
+		}
+	}
+
+	// 5. 图谱扩展（画像/牵挂/心迹 的邻接）
+	resp.Signals = append(resp.Signals, a.graphSignals(ns, resp.Profiles, resp.Episodes)...)
+
+	// 6. 写回 category=signals（异步持久化）
+	go a.persistSignals(resp)
+
+	respond(w, 200, map[string]interface{}{
+		"ok": true, "agent_id": req.AgentID, "namespace": ns,
+		"profiles": len(resp.Profiles),
+		"episodes": len(resp.Episodes),
+		"signals":  len(resp.Signals),
+		"last_run": resp.LastRun,
+	})
+}
+
+// fetchCategory 取某 category 的记忆（用零向量 search，按 importance 排序）
+func (a *API) fetchCategory(ns, cat string, limit int, recallOnly bool) []DeepThinkItem {
+	zeroVec := make([]float32, 1024)
+	results, err := a.LanceDB.Search("memories", zeroVec, limit, ns)
+	if err != nil {
+		return nil
+	}
+	var out []DeepThinkItem
+	for _, m := range results {
+		if cat != "" && m.Category != cat {
+			continue
+		}
+		out = append(out, DeepThinkItem{
+			ID: m.ID, Category: m.Category, Content: truncate(m.Content, 200),
+			Source: "memories", RecallOnly: recallOnly,
+		})
+	}
+	return out
+}
+
+// graphSignals 从图谱取画像/牵挂/心迹的邻接关系
+func (a *API) graphSignals(ns string, profiles, episodes []DeepThinkItem) []DeepThinkItem {
+	var out []DeepThinkItem
+	seen := make(map[string]bool)
+	for _, item := range append(profiles, episodes...) {
+		if item.ID == "" {
+			continue
+		}
+		paths, err := a.GraphStore.Navigate(item.ID, 2, ns, nil)
+		if err != nil {
+			continue
+		}
+		for _, p := range paths {
+			to, _ := p["to"].(string)
+			rel, _ := p["relation"].(string)
+			if to == "" || seen[to] {
+				continue
+			}
+			seen[to] = true
+			out = append(out, DeepThinkItem{
+				ID: to, Category: "graph_relation",
+				Content: fmt.Sprintf("%s → %s: %s", item.ID, to, rel),
+				Source:  "graph", RecallOnly: false,
+			})
+		}
+	}
+	return out
+}
+
+// persistSignals 把信号写回 memories（category=signals）
+// 异步：不把延迟暴露给调用方
+func (a *API) persistSignals(resp DeepThinkResponse) {
+	if len(resp.Signals) == 0 {
+		return
+	}
+	for _, s := range resp.Signals {
+		_, _ = a.LanceDB.InsertEpisode(resp.AgentID, resp.Namespace,
+			fmt.Sprintf("[信号] %s | 来源:%s", s.Content, s.Source), "signals")
+	}
+}
+
+// HandleCommunitySummarize POST /api/v1/community/summarize
+// 运行 Louvain 社区发现 + 摘要（纯 Go，不调 LLM）
+func (a *API) HandleCommunitySummarize(w http.ResponseWriter, r *http.Request) {
+	var req struct {
+		AgentID        string `json:"agent_id"`
+		Namespace      string `json:"namespace,omitempty"`
+		MaxCommunities int    `json:"max_communities,omitempty"`
+	}
+	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
+		respondError(w, 400, "decode request: "+err.Error())
+		return
+	}
+	ns := req.Namespace
+	if ns == "" {
+		ns = deriveNamespace("soulful-xiaowei")
+	}
+	maxC := req.MaxCommunities
+	if maxC == 0 {
+		maxC = 10
+	}
+
+	cds := governance.NewCommunityDiscovery(a.GraphStore)
+	communities, err := cds.Find(ns, 3)
+	if err != nil {
+		respondError(w, 500, "find communities: "+err.Error())
+		return
+	}
+	if len(communities) > maxC {
+		communities = communities[:maxC]
+	}
+	summaries, _ := cds.Summarize(communities, ns, 5)
+
+	respond(w, 200, map[string]interface{}{
+		"ok": true, "namespace": ns,
+		"communities": len(communities),
+		"summaries":   summaries,
+	})
+}

go/internal/api/routes/entity_index.go

@@ -0,0 +1,236 @@
+package routes
+
+import (
+	"sort"
+	"strings"
+	"sync"
+	"time"
+	"unicode"
+)
+
+// NodeVocabularyFunc 返回「节点名 → PageRank 权重」快照（通常是 graphStore.PageRank）。
+type NodeVocabularyFunc func() map[string]float64
+
+// entityCandidate 是查询串里命中的一个图谱节点。
+type entityCandidate struct {
+	Name   string // 原始拼写（用于 Navigate / 笔记匹配）
+	Weight float64
+	Runes  int
+	Pos    int
+}
+
+// EntityIndex 用图谱节点名词表把「整句自然语言」拆成候选实体（本地最长/权重匹配）。
+//
+// 背景：调用方（hermes 织忆插件）把整句用户提问直接当 entity 传给
+// /api/v1/graph/notes，NormalizeEntity 后整句匹配不到任何节点 → 0 篇笔记。
+// 服务端在这里兜底拆实体，调用方无需改动（也就无需重启 hermes 网关）。
+type EntityIndex struct {
+	load NodeVocabularyFunc
+	ttl  time.Duration
+
+	// MinNameRunes / MaxNameRunes 限定候选实体长度（rune 数）。
+	MinNameRunes int
+	MaxNameRunes int
+	// MinWeightRatio 丢掉相对 top 候选过弱的候选（压通用词噪声，如「位置」）。
+	MinWeightRatio float64
+	// Degree 返回实体在图谱里的度数（edges 计数）；nil 表示不做度数过滤。
+	Degree func(name string) int
+	// MinDegree 度数低于此值的候选判为噪声节点（如自动切词产物「一句话」）。
+	MinDegree int
+
+	mu        sync.RWMutex
+	names     map[string]entityCandidate
+	loadedAt  time.Time
+	lastTryAt time.Time
+	retryGap  time.Duration
+
+	now func() time.Time
+}
+
+// NewEntityIndex 创建实体索引；load 为 nil 时 Extract 恒返回空。
+func NewEntityIndex(load NodeVocabularyFunc) *EntityIndex {
+	return &EntityIndex{
+		load:           load,
+		ttl:            10 * time.Minute,
+		MinNameRunes:   2,
+		MaxNameRunes:   8,
+		MinWeightRatio: 0.15,
+		MinDegree:      3,
+		retryGap:       60 * time.Second,
+		now:            time.Now,
+	}
+}
+
+// Warm 在 delay 之后后台加载一次词表，避免首次兜底请求被 2.5s 的 PageRank 阻塞。
+func (e *EntityIndex) Warm(delay time.Duration) {
+	if e == nil || e.load == nil {
+		return
+	}
+	go func() {
+		if delay > 0 {
+			time.Sleep(delay)
+		}
+		e.vocabulary()
+	}()
+}
+
+// vocabulary 返回当前词表（TTL 内复用；加载失败按 retryGap 退避，永不 panic）。
+func (e *EntityIndex) vocabulary() map[string]entityCandidate {
+	if e == nil || e.load == nil {
+		return nil
+	}
+	now := e.now()
+	e.mu.RLock()
+	if len(e.names) > 0 && now.Sub(e.loadedAt) < e.ttl {
+		names := e.names
+		e.mu.RUnlock()
+		return names
+	}
+	e.mu.RUnlock()
+
+	e.mu.Lock()
+	defer e.mu.Unlock()
+	now = e.now()
+	// 双检：可能在等锁期间已被别的请求加载好
+	if len(e.names) > 0 && now.Sub(e.loadedAt) < e.ttl {
+		return e.names
+	}
+	if !e.lastTryAt.IsZero() && now.Sub(e.lastTryAt) < e.retryGap {
+		return e.names // 退避期内：用旧词表（可能为空）但不重复打 DB
+	}
+	e.lastTryAt = now
+	raw := e.load()
+	if len(raw) == 0 {
+		return e.names
+	}
+	names := make(map[string]entityCandidate, len(raw))
+	for name, w := range raw {
+		name = strings.TrimSpace(name)
+		if name == "" {
+			continue
+		}
+		key := strings.ToLower(name)
+		if prev, ok := names[key]; !ok || w > prev.Weight {
+			names[key] = entityCandidate{Name: name, Weight: w, Runes: len([]rune(name))}
+		}
+	}
+	e.names = names
+	e.loadedAt = now
+	return names
+}
+
+// Extract 把整句拆成候选实体：取所有「是查询串子串」的节点名，按权重排序去重。
+// 返回按权重降序的实体名，最多 maxEntities 个；无命中返回 nil。
+func (e *EntityIndex) Extract(query string, maxEntities int) []string {
+	if e == nil || maxEntities <= 0 {
+		return nil
+	}
+	query = strings.TrimSpace(query)
+	if query == "" {
+		return nil
+	}
+	names := e.vocabulary()
+	if len(names) == 0 {
+		return nil
+	}
+	runes := []rune(strings.ToLower(query))
+	total := len(runes)
+	minLen, maxLen := e.MinNameRunes, e.MaxNameRunes
+	if minLen < 1 {
+		minLen = 1
+	}
+	if maxLen < minLen {
+		maxLen = minLen
+	}
+
+	// 收集所有命中：候选名必须是查询串的（区分大小写已归一为小写）子串
+	found := make(map[string]entityCandidate)
+	for i := 0; i < total; i++ {
+		limit := maxLen
+		if total-i < limit {
+			limit = total - i
+		}
+		for n := limit; n >= minLen; n-- {
+			key := string(runes[i : i+n])
+			c, ok := names[key]
+			if !ok || !hasLetter(c.Name) {
+				continue
+			}
+			if prev, dup := found[key]; !dup || c.Weight > prev.Weight {
+				c.Pos = i
+				found[key] = c
+			}
+			break // 同一位置只取最长命中，短前缀交给其他起点
+		}
+	}
+	if len(found) == 0 {
+		return nil
+	}
+	cands := make([]entityCandidate, 0, len(found))
+	for _, c := range found {
+		cands = append(cands, c)
+	}
+	sort.Slice(cands, func(i, j int) bool {
+		if cands[i].Weight != cands[j].Weight {
+			return cands[i].Weight > cands[j].Weight
+		}
+		if cands[i].Runes != cands[j].Runes {
+			return cands[i].Runes > cands[j].Runes
+		}
+		return cands[i].Pos < cands[j].Pos
+	})
+
+	top := cands[0].Weight
+	ratio := e.MinWeightRatio
+	if ratio < 0 {
+		ratio = 0
+	}
+	out := make([]string, 0, maxEntities)
+	checks := 0
+	maxChecks := maxEntities * 2
+	for _, c := range cands {
+		if top > 0 && c.Weight < top*ratio {
+			break // 已按权重降序，后面更弱
+		}
+		redundant := false
+		lc := strings.ToLower(c.Name)
+		for _, sel := range out {
+			if strings.ToLower(sel) == lc {
+				redundant = true
+				break
+			}
+			// 短名被已选长名包含（且不更强）→ 冗余，如「织忆」vs「织忆系统」
+			if strings.Contains(strings.ToLower(sel), lc) {
+				redundant = true
+				break
+			}
+		}
+		if redundant {
+			continue
+		}
+		if e.Degree != nil && checks < maxChecks {
+			checks++
+			if e.Degree(c.Name) < e.MinDegree {
+				continue // 噪声节点（自动切词产物等），不参与笔记匹配
+			}
+		}
+		out = append(out, c.Name)
+		if len(out) >= maxEntities {
+			break
+		}
+	}
+	if len(out) == 0 {
+		return nil
+	}
+	return out
+}
+
+// hasLetter 过滤纯符号/数字的节点名（图谱里有少量噪声节点）。
+func hasLetter(s string) bool {
+	for _, r := range s {
+		if unicode.IsLetter(r) {
+			return true
+		}
+	}
+	return false
+}

go/internal/api/routes/entity_index_test.go

@@ -0,0 +1,166 @@
+package routes
+
+import (
+	"reflect"
+	"testing"
+	"time"
+)
+
+func fakeVocab(m map[string]float64) (NodeVocabularyFunc, *int) {
+	calls := 0
+	return func() map[string]float64 {
+		calls++
+		return m
+	}, &calls
+}
+
+// 真实场景：hermes 织忆插件把整句提问当 entity 传过来，服务端要能拆出实体。
+func TestExtractNLSentence(t *testing.T) {
+	load, _ := fakeVocab(map[string]float64{
+		"织忆": 3.386e-06,
+		"位置": 1.510e-06,
+		"小怡": 3.134e-06,
+		"牧尘": 2.491e-06,
+	})
+	idx := NewEntityIndex(load)
+	got := idx.Extract("现在织忆recall带obsidian实体位置么", 2)
+	want := []string{"织忆", "位置"}
+	if !reflect.DeepEqual(got, want) {
+		t.Fatalf("got %v want %v", got, want)
+	}
+
+	// 只取 1 个时必须是权重最高的那个
+	if got := idx.Extract("现在织忆recall带obsidian实体位置么", 1); !reflect.DeepEqual(got, []string{"织忆"}) {
+		t.Fatalf("top1 got %v", got)
+	}
+}
+
+func TestExtractASCIIKeepsOriginalSpelling(t *testing.T) {
+	load, _ := fakeVocab(map[string]float64{
+		"Obsidian": 2.0e-06,
+		"CLI":      1.0e-06,
+	})
+	idx := NewEntityIndex(load)
+	got := idx.Extract("obsidian 和 cli 怎么配", 5)
+	want := []string{"Obsidian", "CLI"}
+	if !reflect.DeepEqual(got, want) {
+		t.Fatalf("got %v want %v", got, want)
+	}
+}
+
+func TestExtractDropsContainedShorterName(t *testing.T) {
+	load, _ := fakeVocab(map[string]float64{
+		"织忆系统": 3.0e-06,
+		"织忆":   2.0e-06,
+	})
+	idx := NewEntityIndex(load)
+	if got := idx.Extract("织忆系统的笔记", 5); !reflect.DeepEqual(got, []string{"织忆系统"}) {
+		t.Fatalf("got %v", got)
+	}
+}
+
+func TestExtractWeightRatioFilter(t *testing.T) {
+	load, _ := fakeVocab(map[string]float64{"织忆": 3.0e-06, "位置": 1.0e-06})
+	idx := NewEntityIndex(load)
+	if got := idx.Extract("织忆和位置", 5); len(got) != 2 {
+		t.Fatalf("默认阈值应保留两个，got %v", got)
+	}
+	idx.MinWeightRatio = 0.5
+	if got := idx.Extract("织忆和位置", 5); !reflect.DeepEqual(got, []string{"织忆"}) {
+		t.Fatalf("阈值 0.5 应只剩织忆，got %v", got)
+	}
+}
+
+func TestExtractSkipsTooShortAndNonLetterNames(t *testing.T) {
+	load, _ := fakeVocab(map[string]float64{"忆": 5.0e-06, "12": 4.0e-06})
+	idx := NewEntityIndex(load)
+	if got := idx.Extract("织忆12", 5); got != nil {
+		t.Fatalf("expected nil, got %v", got)
+	}
+}
+
+func TestExtractNoHitAndEmptyQuery(t *testing.T) {
+	load, _ := fakeVocab(map[string]float64{"织忆": 3.0e-06})
+	idx := NewEntityIndex(load)
+	if got := idx.Extract("今天天气不错", 3); got != nil {
+		t.Fatalf("expected nil, got %v", got)
+	}
+	if got := idx.Extract("   ", 3); got != nil {
+		t.Fatalf("expected nil, got %v", got)
+	}
+	if got := idx.Extract("织忆", 0); got != nil {
+		t.Fatalf("maxEntities=0 应返回 nil, got %v", got)
+	}
+}
+
+func TestExtractCachesVocabularyAndBacksOffOnFailure(t *testing.T) {
+	load, calls := fakeVocab(map[string]float64{"织忆": 3.0e-06})
+	idx := NewEntityIndex(load)
+	now := time.Unix(1700000000, 0)
+	idx.now = func() time.Time { return now }
+
+	idx.Extract("织忆", 1)
+	idx.Extract("织忆", 1)
+	if *calls != 1 {
+		t.Fatalf("TTL 内应只加载一次，calls=%d", *calls)
+	}
+	now = now.Add(11 * time.Minute)
+	idx.Extract("织忆", 1)
+	if *calls != 2 {
+		t.Fatalf("TTL 过期应重新加载，calls=%d", *calls)
+	}
+
+	// 加载失败（空词表）→ 退避期内不重复打 DB，且不 panic
+	failCalls := 0
+	bad := NewEntityIndex(func() map[string]float64 { failCalls++; return nil })
+	bad.now = func() time.Time { return now }
+	if got := bad.Extract("织忆", 1); got != nil {
+		t.Fatalf("空词表应返回 nil, got %v", got)
+	}
+	bad.Extract("织忆", 1)
+	if failCalls != 1 {
+		t.Fatalf("退避期内应只尝试一次，failCalls=%d", failCalls)
+	}
+	now = now.Add(2 * time.Minute)
+	bad.Extract("织忆", 1)
+	if failCalls != 2 {
+		t.Fatalf("退避期过后应重试，failCalls=%d", failCalls)
+	}
+}
+
+func TestExtractDegreeFilterDropsNoiseNode(t *testing.T) {
+	load, _ := fakeVocab(map[string]float64{
+		"天气":  3.0e-06,
+		"一句话": 1.6e-06,
+	})
+	idx := NewEntityIndex(load)
+	idx.Degree = func(name string) int {
+		if name == "一句话" {
+			return 2 // 自动切词噪声节点（实测 graph.db：一句话 deg=2）
+		}
+		return 50
+	}
+	got := idx.Extract("完全不相干的一句话天气不错", 3)
+	if !reflect.DeepEqual(got, []string{"天气"}) {
+		t.Fatalf("got %v want [天气]", got)
+	}
+
+	// 度数过滤把候选清空 → 返回 nil（宁可不给，也别给噪声笔记）
+	idx2 := NewEntityIndex(load)
+	idx2.Degree = func(string) int { return 0 }
+	if got := idx2.Extract("完全不相干的一句话天气不错", 3); got != nil {
+		t.Fatalf("got %v want nil", got)
+	}
+}
+
+func TestNilEntityIndexIsSafe(t *testing.T) {
+	var idx *EntityIndex
+	idx.Warm(0)
+	if got := idx.Extract("织忆", 3); got != nil {
+		t.Fatalf("expected nil, got %v", got)
+	}
+	noLoad := NewEntityIndex(nil)
+	if got := noLoad.Extract("织忆", 3); got != nil {
+		t.Fatalf("expected nil, got %v", got)
+	}
+}

go/internal/api/routes/entity_resolve.go

@@ -0,0 +1,120 @@
+// 织忆 MemoryWeave — 实体名字 → 同名全部节点 id 的并集解析
+//
+// 背景（2026-09-27 图谱重名成因报告）：图谱里同一实体可能有两层节点 ——
+//
+//	· 旧 entity 层：id = 名字（namespace=wiki 等，06-16 起写入）
+//	· 新 concept 层：id = "n_" + 名字（namespace=hermes-main 等，09-26 起写入，边多得多）
+//
+// 而 Navigate 按 id 匹配（`WHERE e.source = '<entity>'`），PageRank 词表键也是 id
+// ⇒ 按名字导航只会走旧层，新层的连接形同隐形（织忆：旧层 6 出边 vs 新层 1674 出边）。
+//
+// 本文件提供并集解析：给定名字，返回**全部**同名节点 id，让调用方对并集展开，
+// 从而两层数据都能进入笔记匹配。
+package routes
+
+import (
+	"fmt"
+	"strings"
+)
+
+// NodeSearcher 只要实现 SearchNodes 即可（governance.GraphStore 满足）。
+// 收窄成小接口是为了单测里能塞假实现，不必实现整个图谱接口。
+type NodeSearcher interface {
+	SearchNodes(label string, namespace string) []map[string]interface{}
+}
+
+// ExactNodeFinder 精确同名解析（governance.SQLiteGraphStore 实现 NodeIDsByName）。
+// 有它就用它 —— SearchNodes 的 LIKE+LIMIT 语义会把精确同名的节点挤出结果。
+type ExactNodeFinder interface {
+	NodeIDsByName(name string) []string
+}
+
+// ResolveEntityIDs 返回与 name 对应的全部节点 id（同名两层都返回）。
+//
+// 语义：
+//   - 优先精确匹配：节点的 name 或 id 与 name 相等（忽略大小写）才算命中，
+//     避免 SearchNodes 的 LIKE 子串语义把「织忆」带出「织忆召回」这类噪声节点。
+//   - 若 gs 实现了 ExactNodeFinder 则走精确 SQL（首选，见 NodeIDsByName 注释）。
+//   - 名字本身始终作为回退候选放在结果里（图谱里查不到、但词表命中的名字仍可按 id 直查）。
+//   - 结果去重，稳定有序：先旧层（id=名字）后新层（id=n_名字），最后是名字回退。
+func ResolveEntityIDs(gs NodeSearcher, name string) []string {
+	name = strings.TrimSpace(name)
+	if name == "" {
+		return nil
+	}
+	ids := make([]string, 0, 2)
+	seen := make(map[string]bool, 4)
+	add := func(id string) {
+		if id == "" || seen[id] {
+			return
+		}
+		seen[id] = true
+		ids = append(ids, id)
+	}
+	if ex, ok := gs.(ExactNodeFinder); ok {
+		for _, id := range ex.NodeIDsByName(name) {
+			add(id)
+		}
+	} else if gs != nil {
+		for _, node := range gs.SearchNodes(name, "") {
+			nid, _ := node["id"].(string)
+			nname, _ := node["name"].(string)
+			if nid == "" {
+				continue
+			}
+			if strings.EqualFold(nname, name) || strings.EqualFold(nid, name) {
+				add(nid)
+			}
+		}
+	}
+	add(name) // 回退：按 id 直查（id == 名字）
+	return ids
+}
+
+// MergePaths 把并集展开得到的路径合并进 dst，按 source|relation|target 去重。
+// 返回合并后的切片（可能原地扩容）；limit > 0 时最多保留 limit 条。
+func MergePaths(dst []map[string]interface{}, src []map[string]interface{}, seen map[string]bool, limit int) []map[string]interface{} {
+	if seen == nil {
+		seen = make(map[string]bool, len(dst)+len(src))
+	}
+	for _, p := range dst {
+		seen[pathKey(p)] = true
+	}
+	for _, p := range src {
+		if limit > 0 && len(dst) >= limit {
+			break
+		}
+		k := pathKey(p)
+		if k == "" || seen[k] {
+			continue
+		}
+		seen[k] = true
+		dst = append(dst, p)
+	}
+	return dst
+}
+
+// PathEnds 取出路径的 (source, target, relation)。
+// 兼容两套键名：governance.Navigate 返回的是 from/to，而其它构造处用 source/target——
+// 只读一套会静默丢路径（2026-09-27 实测：并集路径、以及「图谱邻居进笔记匹配」整条链路
+// 都因为读错键名而失效，全程无报错）。
+func PathEnds(p map[string]interface{}) (string, string, string) {
+	s, _ := p["source"].(string)
+	if s == "" {
+		s, _ = p["from"].(string)
+	}
+	t, _ := p["target"].(string)
+	if t == "" {
+		t, _ = p["to"].(string)
+	}
+	r, _ := p["relation"].(string)
+	return s, t, r
+}
+
+func pathKey(p map[string]interface{}) string {
+	s, t, r := PathEnds(p)
+	if s == "" && t == "" {
+		return ""
+	}
+	return fmt.Sprintf("%s|%s|%s", s, r, t)
+}

go/internal/api/routes/entity_resolve_test.go

@@ -0,0 +1,170 @@
+package routes
+
+import (
+	"strings"
+	"testing"
+)
+
+// fakeSearcher 模拟 SearchNodes 的 LIKE 语义返回，用于验证精确匹配过滤。
+type fakeSearcher struct {
+	nodes []map[string]interface{}
+	calls int
+}
+
+func (f *fakeSearcher) SearchNodes(label, namespace string) []map[string]interface{} {
+	f.calls++
+	out := []map[string]interface{}{}
+	for _, n := range f.nodes {
+		nm, _ := n["name"].(string)
+		id, _ := n["id"].(string)
+		if strings.Contains(nm, label) || strings.Contains(id, label) {
+			out = append(out, n)
+		}
+	}
+	return out
+}
+
+// 同名两层：旧 entity（id=名字）+ 新 concept（id=n_名字）→ 都应返回，且顺序稳定。
+func TestResolveEntityIDs_TwoLayers(t *testing.T) {
+	gs := &fakeSearcher{nodes: []map[string]interface{}{
+		{"id": "织忆", "name": "织忆", "type": "entity"},
+		{"id": "n_织忆", "name": "织忆", "type": "concept"},
+	}}
+	ids := ResolveEntityIDs(gs, "织忆")
+	if len(ids) != 2 {
+		t.Fatalf("want 2 ids (两层), got %v", ids)
+	}
+	if ids[0] != "织忆" || ids[1] != "n_织忆" {
+		t.Fatalf("unexpected order: %v", ids)
+	}
+}
+
+// LIKE 子串噪声必须被挡掉：查「织忆」不应带出「织忆召回」。
+func TestResolveEntityIDs_RejectsSubstringNoise(t *testing.T) {
+	gs := &fakeSearcher{nodes: []map[string]interface{}{
+		{"id": "织忆", "name": "织忆"},
+		{"id": "织忆召回机制", "name": "织忆召回机制"},
+	}}
+	ids := ResolveEntityIDs(gs, "织忆")
+	for _, id := range ids {
+		if id == "织忆召回机制" {
+			t.Fatalf("子串噪声未被过滤: %v", ids)
+		}
+	}
+	if len(ids) != 1 {
+		t.Fatalf("want 1 (仅精确), got %v", ids)
+	}
+}
+
+// 图谱查不到时回退为名字本身（按 id 直查仍可用）。
+func TestResolveEntityIDs_FallbackToName(t *testing.T) {
+	gs := &fakeSearcher{}
+	ids := ResolveEntityIDs(gs, "不存在的实体")
+	if len(ids) != 1 || ids[0] != "不存在的实体" {
+		t.Fatalf("want fallback to name, got %v", ids)
+	}
+}
+
+// nil store / 空名字不外抛。
+func TestResolveEntityIDs_NilAndEmpty(t *testing.T) {
+	if got := ResolveEntityIDs(nil, "织忆"); len(got) != 1 || got[0] != "织忆" {
+		t.Fatalf("nil store 应回退名字: %v", got)
+	}
+	if got := ResolveEntityIDs(nil, "   "); got != nil {
+		t.Fatalf("空名字应返回 nil, got %v", got)
+	}
+}
+
+// 去重：SearchNodes 同时按 name 与 id 命中同一节点时只返回一次。
+func TestResolveEntityIDs_Dedup(t *testing.T) {
+	gs := &fakeSearcher{nodes: []map[string]interface{}{
+		{"id": "n_织忆", "name": "织忆"},
+		{"id": "n_织忆", "name": "n_织忆"},
+	}}
+	ids := ResolveEntityIDs(gs, "织忆")
+	// n_织忆 命中（name 相等）；名字回退也加入 ⇒ 共 2 条且不重复
+	if len(ids) != 2 || ids[0] != "n_织忆" || ids[1] != "织忆" {
+		t.Fatalf("dedup 结果异常: %v", ids)
+	}
+}
+
+// fakeExact 同时实现 NodeSearcher 与 ExactNodeFinder：验证优先走精确 SQL。
+// 模拟真实场景：SearchNodes（LIKE）被 834 条子串命中淹没、精确改名拿不到 ⇒ 必须用精确路径。
+type fakeExact struct {
+	exactIDs []string
+	likeRows []map[string]interface{}
+}
+
+func (f *fakeExact) SearchNodes(label, namespace string) []map[string]interface{} { return f.likeRows }
+func (f *fakeExact) NodeIDsByName(name string) []string                           { return f.exactIDs }
+
+func TestResolveEntityIDs_PrefersExactFinder(t *testing.T) {
+	gs := &fakeExact{
+		exactIDs: []string{"织忆", "n_织忆"},
+		likeRows: []map[string]interface{}{ // LIKE 噪声：精确同名的 n_织忆 不在其中
+			{"id": "n_织忆召回机制", "name": "织忆召回机制"},
+			{"id": "n_织忆架构", "name": "织忆架构"},
+		},
+	}
+	ids := ResolveEntityIDs(gs, "织忆")
+	want := []string{"织忆", "n_织忆"}
+	if len(ids) != len(want) || ids[0] != want[0] || ids[1] != want[1] {
+		t.Fatalf("want %v, got %v", want, ids)
+	}
+}
+
+// MergePaths：按 source|relation|target 去重，并尊重上限。
+func TestMergePaths_TolerantToFromToKeys(t *testing.T) {
+	// governance.Navigate 返回 from/to；MergePaths 必须认，否则路径被静默丢弃
+	nav := []map[string]interface{}{
+		{"from": "织忆", "to": "Hermes", "relation": "USES"},
+		{"from": "织忆", "to": "LanceDB", "relation": "DEPENDS_ON"},
+	}
+	out := MergePaths([]map[string]interface{}{}, nav, nil, 0)
+	if len(out) != 2 {
+		t.Fatalf("Navigate 风格 from/to 路径被丢弃：want 2, got %d", len(out))
+	}
+	// 去重：同一 from|relation|to 再来一次不重复
+	out = MergePaths(out, nav, nil, 0)
+	if len(out) != 2 {
+		t.Fatalf("去重失败：want 2, got %d", len(out))
+	}
+	// limit 生效
+	out = MergePaths([]map[string]interface{}{}, nav, nil, 1)
+	if len(out) != 1 {
+		t.Fatalf("limit 未生效：want 1, got %d", len(out))
+	}
+}
+
+func TestPathEnds_BothKeySets(t *testing.T) {
+	if s, tt, r := PathEnds(map[string]interface{}{"from": "a", "to": "b", "relation": "R"}); s != "a" || tt != "b" || r != "R" {
+		t.Fatalf("from/to 未识别: %s %s %s", s, tt, r)
+	}
+	if s, tt, _ := PathEnds(map[string]interface{}{"source": "a", "target": "b"}); s != "a" || tt != "b" {
+		t.Fatalf("source/target 未识别: %s %s", s, tt)
+	}
+}
+
+func TestMergePaths_DedupAndLimit(t *testing.T) {
+	p := func(s, r, tg string) map[string]interface{} {
+		return map[string]interface{}{"source": s, "relation": r, "target": tg}
+	}
+	dst := []map[string]interface{}{p("a", "is", "b")}
+	seen := map[string]bool{}
+	dst = MergePaths(dst, []map[string]interface{}{
+		p("a", "is", "b"), // 重复
+		p("n_a", "is", "c"),
+		p("n_a", "is", "d"),
+	}, seen, 2)
+	if len(dst) != 2 {
+		t.Fatalf("want limit 2, got %d: %v", len(dst), dst)
+	}
+	if dst[1]["target"] != "c" {
+		t.Fatalf("第二条应为新路径 c, got %v", dst[1])
+	}
+	// 再次合并应识别已有 key（含 dst 内建索引）
+	dst = MergePaths(dst, []map[string]interface{}{p("n_a", "is", "c")}, seen, 0)
+	if len(dst) != 2 {
+		t.Fatalf("重复路径未被去重: %v", dst)
+	}
+}

go/internal/api/routes/eval.go

@@ -3,27 +3,172 @@ package routes
 
 import (
 	"encoding/json"
+	_ "embed"
+	"fmt"
 	"net/http"
+	"os"
+	"strings"
 	"sync"
 	"time"
 
 	"github.com/xiaoxue/memoryweave/internal/storage"
 )
+
+// go:embed testdata/golden-set.json 把金标集打进二进制（t_dfc93ad0 AC5）：
+// 部署 CWD=/home/muc，旧版「path=="" → 相对 testdata/」解析到 /home/muc/testdata
+// ⇒ POST /api/v1/eval/golden 不传 golden_path 时 500（2026-10-14 实测复现）。
+// 内嵌兜底后：缺省/相对路径读不到 ⇒ 回落内嵌基线（仍可用）；env ZHIYI_GOLDEN_SET
+// 指定绝对路径 ⇒ 读该文件（可临时替换金标集）。测试仍按显式路径调用，行为不变。
+//
+//go:embed testdata/golden-set.json
+var embeddedGoldenSetJSON []byte
+
+// ─── 金标集定义 ──────────────────────────────────────────────
+
+// GoldenSetItem 单条金标：query + 具体 expected_ids + 关键词相关性判据
+type GoldenSetItem struct {
+	Query               string   `json:"query"`
+	ExpectedIDs         []string `json:"expected_ids"`
+	MustIncludeKeywords []string `json:"must_include_keywords"`
+	// Namespace 该条目应在哪个 namespace 里召回。为空则用请求传入的 ns。
+	// 为什么必须 per-item（2026-10-14 实测）：金标的 expected_id 分散在
+	// hermes-main / domain:HUB / soulful-xiaowei / wiki / shared 等多个 ns，
+	// 而 eval 只按请求里的单一 ns 召回 ⇒ ID 类指标**按构造恒为 0**（不是召回差）。
+	Namespace  string `json:"namespace,omitempty"`
+	Category   string `json:"category,omitempty"`
+	Difficulty string `json:"difficulty,omitempty"`
+	IDBaseline string `json:"id_baseline,omitempty"`
+}
+
+// GoldenSet 金标集
+type GoldenSet struct {
+	Description string          `json:"description"`
+	Items       []GoldenSetItem `json:"items"`
+}
+
+// LoadGoldenSet 加载金标集。路径解析（t_dfc93ad0 AC5，部署 CWD=/home/muc）：
+//   1. env ZHIYI_GOLDEN_SET 指定路径 ⇒ 优先读它（临时替换金标集用，不重编译）；
+//   2. 显式 path 非空且能读到 ⇒ 用它（绝对/相对皆按调用方语义）；
+//   3. path 为空、或上面读不到（典型：部署二进制 CWD=/home/muc 无 testdata/，
+//      旧版相对路径解析失败 ⇒ 500）⇒ 回落 **内嵌** 基线 embeddedGoldenSetJSON。
+// 校验（空断言自守）在拿到内容后统一跑，三种来源同口径。
+func LoadGoldenSet(path string) (*GoldenSet, error) {
+	var data []byte
+	usedSource := "env:ZHIYI_GOLDEN_SET"
+	if p := os.Getenv("ZHIYI_GOLDEN_SET"); p != "" {
+		var err error
+		data, err = os.ReadFile(p)
+		if err != nil {
+			return nil, err
+		}
+	} else if path != "" {
+		usedSource = "path:" + path
+		d, err := os.ReadFile(path)
+		if err != nil {
+			// 相对路径在部署 CWD 下读不到 ⇒ 内嵌兜底（AC5 根因）
+			data = embeddedGoldenSetJSON
+			usedSource = "embedded"
+		} else {
+			data = d
+		}
+	} else {
+		data = embeddedGoldenSetJSON
+		usedSource = "embedded"
+	}
+	if len(data) == 0 {
+		return nil, fmt.Errorf("golden set 不可用（%s 为空且无内嵌基线）", usedSource)
+	}
+	var gs GoldenSet
+	if err := json.Unmarshal(data, &gs); err != nil {
+		return nil, err
+	}
+	// 空断言自守：禁止裸前缀/空串混入（避免恒真）
+	for _, item := range gs.Items {
+		if err := validateExpectedIDs(item.Query, item.ExpectedIDs); err != nil {
+			return nil, err
+		}
+	}
+	return &gs, nil
+}
+
+// validateExpectedIDs 空断言自守：裸前缀 "mem_"、空串、None → 直接报错，不许静默 PASS
+func validateExpectedIDs(query string, expectedIDs []string) error {
+	if len(expectedIDs) == 0 {
+		return nil // 允许空（bootstrap 时还没基线）
+	}
+	for _, id := range expectedIDs {
+		trimmed := strings.TrimSpace(id)
+		// 裸前缀 "mem_"（不含具体 ID 数字）
+		if trimmed == "mem_" || trimmed == "mem" {
+			return &VacuousAssertionError{
+				Query:  query,
+				BadID:  id,
+				Reason: "裸前缀 mem_ 是恒真断言——所有记忆 ID 都以 mem_ 开头",
+			}
+		}
+		// 空串
+		if trimmed == "" {
+			return &VacuousAssertionError{
+				Query:  query,
+				BadID:  id,
+				Reason: "空串期望值（无意义断言）",
+			}
+		}
+		// None（JSON null 解码成空字符串已处理，但检查 "None" 字符串）
+		if strings.EqualFold(trimmed, "none") {
+			return &VacuousAssertionError{
+				Query:  query,
+				BadID:  id,
+				Reason: "None 期望值（无意义断言）",
+			}
+		}
+	}
+	return nil
+}
+
+// VacuousAssertionError 空断言错误
+type VacuousAssertionError struct {
+	Query  string
+	BadID  string
+	Reason string
+}
+
+func (e *VacuousAssertionError) Error() string {
+	return "FATAL 空断言被检出（恒真）: query=" + e.Query + " 期望=" + e.BadID + " — " + e.Reason
+}
+
+// checkKeywordRelevance 关键词相关性判据：返回内容必须包含至少一个关键词。
+// 统一口径（t_a70b07e1）：本函数单条内容判定的输入 N（取 recall 返回的前 N 条）由调用方决定——
+// eval 端 TestEvalGolden_Live 取前 N=5（见 eval_test.go 口径注释），与 golden-check.py 一致。
+func checkKeywordRelevance(content string, keywords []string) bool {
+	if len(keywords) == 0 {
+		// 无关键词判据 → 视为不通过（禁止退化为恒真）
+		return false
+	}
+	lower := strings.ToLower(content)
+	for _, kw := range keywords {
+		if strings.Contains(lower, strings.ToLower(kw)) {
+			return true
+		}
+	}
+	return false
+}
+
 type EvalAPI struct {
 	Pipeline *storage.RecallPipeline
 	LanceDB  storage.LanceDB
 }
 
 type EvalRun struct {
-	ID        string    `json:"id"`
-	Model     string    `json:"model"`
-	RecallAt5 float64   `json:"recall_at_5"`
+	ID           string  `json:"id"`
+	Model        string  `json:"model"`
+	RecallAt5    float64 `json:"recall_at_5"`
 	PrecisionAt5 float64 `json:"precision_at_5"`
-	MRR       float64   `json:"mean_reciprocal_rank"`
-	NDCG      float64   `json:"ndcg"`
+	MRR          float64 `json:"mean_reciprocal_rank"`
+	NDCG         float64 `json:"ndcg"`
 
 	// 按标签分组
-	RecallByTag   map[string]float64 `json:"recall_at_5_by_tag,omitempty"`
+	RecallByTag    map[string]float64 `json:"recall_at_5_by_tag,omitempty"`
 	PrecisionByTag map[string]float64 `json:"precision_at_5_by_tag,omitempty"`
 
 	// 按 Agent 分组
@@ -40,17 +185,27 @@ type EvalRun struct {
 }
 
 type EvalQueryDetail struct {
-	Query       string   `json:"query"`
-	ExpectedIDs []string `json:"expected_ids"`
-	Hits        []string `json:"hits"`
-	Misses      []string `json:"misses"`
-	RecallAt5   float64  `json:"recall_at_5"`
-	PrecisionAt5 float64 `json:"precision_at_5"`
+	Query               string   `json:"query"`
+	ExpectedIDs         []string `json:"expected_ids"`
+	MustIncludeKeywords []string `json:"must_include_keywords,omitempty"`
+	Hits                []string `json:"hits"`
+	Misses              []string `json:"misses"`
+	RecallAt5           float64  `json:"recall_at_5"`
+	PrecisionAt5        float64  `json:"precision_at_5"`
+	KeywordOK           bool     `json:"keyword_ok,omitempty"`
+}
+
+// GoldenRunResult 金标 run 的扩展结果（含关键词判据）
+type GoldenRunResult struct {
+	EvalRun
+	KeywordPass  int `json:"keyword_pass"`
+	KeywordFail  int `json:"keyword_fail"`
+	KeywordTotal int `json:"keyword_total"`
 }
 
 type ConsolidationHits struct {
-	TotalDistilledUsed int     `json:"total_distilled_used"`
-	HitsAfterConsolidation int  `json:"hits_after_consolidation"`
+	TotalDistilledUsed     int     `json:"total_distilled_used"`
+	HitsAfterConsolidation int     `json:"hits_after_consolidation"`
 	ConsolidationBenefit   float64 `json:"consolidation_benefit"`
 }
 
@@ -68,7 +223,7 @@ func NewEvalAPI(p *storage.RecallPipeline, ldb storage.LanceDB) *EvalAPI {
 // POST /api/v1/eval/run
 func (ea *EvalAPI) Run(w http.ResponseWriter, r *http.Request) {
 	var req struct {
-		Queries   []struct {
+		Queries []struct {
 			Query       string   `json:"query"`
 			ExpectedIDs []string `json:"expected_ids"`
 		} `json:"queries"`
@@ -85,6 +240,14 @@ func (ea *EvalAPI) Run(w http.ResponseWriter, r *http.Request) {
 		return
 	}
 
+	// 空断言自守：裸前缀 mem_ / 空串 / None → 直接报错，不许静默 PASS
+	for _, q := range req.Queries {
+		if err := validateExpectedIDs(q.Query, q.ExpectedIDs); err != nil {
+			respondError(w, 400, err.Error())
+			return
+		}
+	}
+
 	// 默认 namespace 为 shared，支持传入 hermes 等
 	ns := req.Namespace
 	if ns == "" {
@@ -125,11 +288,15 @@ func (ea *EvalAPI) Run(w http.ResponseWriter, r *http.Request) {
 		}
 
 		pAt5 := 0.0
-		if len(results) > 0 { pAt5 = float64(hits) / float64(len(results)) }
+		if len(results) > 0 {
+			pAt5 = float64(hits) / float64(len(results))
+		}
 		totalPrecision += pAt5
 
 		rAt5 := 0.0
-		if len(q.ExpectedIDs) > 0 { rAt5 = float64(hits) / float64(len(q.ExpectedIDs)) }
+		if len(q.ExpectedIDs) > 0 {
+			rAt5 = float64(hits) / float64(len(q.ExpectedIDs))
+		}
 		totalRecall += rAt5
 
 		// MRR
@@ -144,11 +311,17 @@ func (ea *EvalAPI) Run(w http.ResponseWriter, r *http.Request) {
 		dcg, idcg := 0.0, 0.0
 		for i, res := range results {
 			rel := 0.0
-			if expectedSet[res.ID] { rel = 1.0 }
+			if expectedSet[res.ID] {
+				rel = 1.0
+			}
 			dcg += rel / log2(float64(i+2))
-			if i < len(q.ExpectedIDs) { idcg += 1.0 / log2(float64(i+2)) }
+			if i < len(q.ExpectedIDs) {
+				idcg += 1.0 / log2(float64(i+2))
+			}
+		}
+		if idcg > 0 {
+			totalNDCG += dcg / idcg
 		}
-		if idcg > 0 { totalNDCG += dcg / idcg }
 
 		// 分组统计
 		// (category 从 query 推断)
@@ -183,18 +356,18 @@ func (ea *EvalAPI) Run(w http.ResponseWriter, r *http.Request) {
 	recallByAgent := avgByGroup(recallByAgentMap)
 
 	run := &EvalRun{
-		ID:            time.Now().Format("20060102-150405"),
-		Model:         req.Model,
-		RecallAt5:     totalRecall / float64(totalQueries),
-		PrecisionAt5:  totalPrecision / float64(totalQueries),
-		MRR:           totalMRR / float64(totalQueries),
-		NDCG:          totalNDCG / float64(totalQueries),
-		RecallByTag:   recallByTag,
+		ID:             time.Now().Format("20060102-150405"),
+		Model:          req.Model,
+		RecallAt5:      totalRecall / float64(totalQueries),
+		PrecisionAt5:   totalPrecision / float64(totalQueries),
+		MRR:            totalMRR / float64(totalQueries),
+		NDCG:           totalNDCG / float64(totalQueries),
+		RecallByTag:    recallByTag,
 		PrecisionByTag: precisionByTag,
-		RecallByAgent: recallByAgent,
-		QueryDetails:  queryDetails,
-		Queries:       totalQueries,
-		RanAt:         time.Now(),
+		RecallByAgent:  recallByAgent,
+		QueryDetails:   queryDetails,
+		Queries:        totalQueries,
+		RanAt:          time.Now(),
 	}
 
 	evals.mu.Lock()
@@ -204,6 +377,176 @@ func (ea *EvalAPI) Run(w http.ResponseWriter, r *http.Request) {
 	respond(w, 200, run)
 }
 
+// POST /api/v1/eval/golden — 使用金标集跑评估（含关键词相关性判据）
+// 加载 testdata/golden-set.json，对每条执行 recall + 关键词检查
+// 返回 GoldenRunResult（含 keyword_pass/fail/total）
+func (ea *EvalAPI) RunGolden(w http.ResponseWriter, r *http.Request) {
+	var req struct {
+		Namespace  string `json:"namespace"`
+		GoldenPath string `json:"golden_path"` // 可选，默认 testdata/golden-set.json
+	}
+	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
+		respondError(w, 400, "invalid body")
+		return
+	}
+	ns := req.Namespace
+	if ns == "" {
+		ns = "shared"
+	}
+	gs, err := LoadGoldenSet(req.GoldenPath)
+	if err != nil {
+		respondError(w, 500, "failed to load golden set: "+err.Error())
+		return
+	}
+
+	var totalPrecision, totalRecall, totalMRR, totalNDCG float64
+	totalQueries := 0
+	keywordPass, keywordFail, keywordTotal := 0, 0, 0
+	var queryDetails []EvalQueryDetail
+	var failures []string
+
+	for _, item := range gs.Items {
+		// 空断言自守
+		if err := validateExpectedIDs(item.Query, item.ExpectedIDs); err != nil {
+			respondError(w, 400, err.Error())
+			return
+		}
+		// 无关键词判据 → 禁止（会退化成恒真）
+		if len(item.MustIncludeKeywords) == 0 {
+			failures = append(failures, item.Query+": 无关键词判据（禁止退化成恒真）")
+			continue
+		}
+
+		// per-item namespace：条目自带 ns 优先（金标基线是按 ns 钉的，见 GoldenSetItem.Namespace）
+		itemNS := item.Namespace
+		if itemNS == "" {
+			itemNS = ns
+		}
+		results, err := ea.Pipeline.Recall(item.Query, itemNS, 5, 0.5)
+		if err != nil {
+			failures = append(failures, item.Query+": recall 失败 "+err.Error())
+			keywordFail++
+			keywordTotal++
+			continue
+		}
+
+		// ID 命中检查
+		hits := 0
+		var hitIDs []string
+		var missIDs []string
+		expectedSet := makeSet(item.ExpectedIDs)
+		for i, res := range results {
+			if expectedSet[res.ID] && i < 5 {
+				hits++
+				hitIDs = append(hitIDs, res.ID)
+			}
+		}
+		for _, eid := range item.ExpectedIDs {
+			if !contains(hitIDs, eid) {
+				missIDs = append(missIDs, eid)
+			}
+		}
+
+		// 关键词相关性判据（t_a70b07e1 统一口径）：取 recall 返回的前 N=5 条内容查关键词。
+		// （Pipeline.Recall 传 topK=5；此处扫全部 results 即前 N 条。Python 量具 golden-check.py 同口径。）
+		keywordOK := false
+		for _, res := range results {
+			if checkKeywordRelevance(res.Content, item.MustIncludeKeywords) {
+				keywordOK = true
+				break
+			}
+		}
+
+		pAt5 := 0.0
+		if len(results) > 0 {
+			pAt5 = float64(hits) / float64(len(results))
+		}
+		totalPrecision += pAt5
+
+		rAt5 := 0.0
+		if len(item.ExpectedIDs) > 0 {
+			rAt5 = float64(hits) / float64(len(item.ExpectedIDs))
+		}
+		totalRecall += rAt5
+
+		// MRR
+		for i, res := range results {
+			if expectedSet[res.ID] {
+				totalMRR += 1.0 / float64(i+1)
+				break
+			}
+		}
+
+		// NDCG
+		dcg, idcg := 0.0, 0.0
+		for i, res := range results {
+			rel := 0.0
+			if expectedSet[res.ID] {
+				rel = 1.0
+			}
+			dcg += rel / log2(float64(i+2))
+			if i < len(item.ExpectedIDs) {
+				idcg += 1.0 / log2(float64(i+2))
+			}
+		}
+		if idcg > 0 {
+			totalNDCG += dcg / idcg
+		}
+
+		// 关键词统计
+		keywordTotal++
+		if keywordOK {
+			keywordPass++
+		} else {
+			keywordFail++
+			failures = append(failures, item.Query+": 关键词未命中 "+strings.Join(item.MustIncludeKeywords, ","))
+		}
+
+		queryDetails = append(queryDetails, EvalQueryDetail{
+			Query:               item.Query,
+			ExpectedIDs:         item.ExpectedIDs,
+			MustIncludeKeywords: item.MustIncludeKeywords,
+			Hits:                hitIDs,
+			Misses:              missIDs,
+			RecallAt5:           rAt5,
+			PrecisionAt5:        pAt5,
+			KeywordOK:           keywordOK,
+		})
+
+		totalQueries++
+	}
+
+	if totalQueries == 0 {
+		respondError(w, 500, "all golden queries failed")
+		return
+	}
+
+	run := &EvalRun{
+		ID:           time.Now().Format("20060102-150405"),
+		Model:        "golden-set",
+		RecallAt5:    totalRecall / float64(totalQueries),
+		PrecisionAt5: totalPrecision / float64(totalQueries),
+		MRR:          totalMRR / float64(totalQueries),
+		NDCG:         totalNDCG / float64(totalQueries),
+		QueryDetails: queryDetails,
+		Queries:      totalQueries,
+		RanAt:        time.Now(),
+	}
+
+	evals.mu.Lock()
+	evals.runs = append(evals.runs, run)
+	evals.mu.Unlock()
+
+	result := GoldenRunResult{
+		EvalRun:      *run,
+		KeywordPass:  keywordPass,
+		KeywordFail:  keywordFail,
+		KeywordTotal: keywordTotal,
+	}
+	_ = failures
+	respond(w, 200, result)
+}
+
 // GET /api/v1/eval/history
 func (ea *EvalAPI) History(w http.ResponseWriter, r *http.Request) {
 	evals.mu.RLock()
@@ -233,6 +576,8 @@ func (ea *EvalAPI) History(w http.ResponseWriter, r *http.Request) {
 
 // POST /api/v1/eval/generate — 生成 4类×3级=12条金标查询
 // 对每个 query 调用 recall 获取 expected_ids
+// 注意：此端点用于 bootstrap 基线——首次跑时没有基线，用 recall top-3 填充。
+// 后续应改用 LoadGoldenSet 加载已校准的金标集（含具体 ID + 关键词判据）。
 func (ea *EvalAPI) Generate(w http.ResponseWriter, r *http.Request) {
 	var req struct {
 		Namespace string `json:"namespace"`
@@ -252,9 +597,9 @@ func (ea *EvalAPI) Generate(w http.ResponseWriter, r *http.Request) {
 	// 固定的 12 条金标查询（4类×3级）
 	// 对每条生成 query → recall 获得真实 expected_ids
 	goldenQueries := []struct {
-		cat       string
+		cat        string
 		difficulty string
-		query     string
+		query      string
 	}{
 		// system_fact
 		{"system_fact", "easy", "织忆是什么"},
@@ -284,6 +629,17 @@ func (ea *EvalAPI) Generate(w http.ResponseWriter, r *http.Request) {
 				expectedIDs = append(expectedIDs, r.ID)
 			}
 		}
+		// 空断言自守：如果 recall 返回 0 条，不得写入空 expected_ids（那是 bootstrap，不是校准）
+		if len(expectedIDs) == 0 {
+			results = append(results, map[string]interface{}{
+				"query":        gq.query,
+				"expected_ids": []string{},
+				"category":     gq.cat,
+				"difficulty":   gq.difficulty,
+				"warning":      "recall 返回 0 条——bootstrap 基线为空，需人工校准",
+			})
+			continue
+		}
 		results = append(results, map[string]interface{}{
 			"query":        gq.query,
 			"expected_ids": expectedIDs,
@@ -295,7 +651,7 @@ func (ea *EvalAPI) Generate(w http.ResponseWriter, r *http.Request) {
 	respond(w, 200, map[string]interface{}{
 		"queries": results,
 		"count":   len(results),
-		"note":    "4类别×3难度=12条金标查询，通过 recall 自动获取 expected_ids",
+		"note":    "4类别×3难度=12条金标查询，通过 recall 自动获取 expected_ids。注意：bootstrap 模式下 expected_ids 来自实时 recall，非校准基线。校准后的金标集见 testdata/golden-set.json（LoadGoldenSet）。",
 	})
 }
 
@@ -310,14 +666,21 @@ func makeSet(ids []string) map[string]bool {
 
 func log2(x float64) float64 {
 	result := 0.0
-	for x > 2 { x /= 2; result += 1 }
-	if x > 1 { result += (x - 1) }
+	for x > 2 {
+		x /= 2
+		result += 1
+	}
+	if x > 1 {
+		result += (x - 1)
+	}
 	return result
 }
 
 func contains(list []string, item string) bool {
 	for _, s := range list {
-		if s == item { return true }
+		if s == item {
+			return true
+		}
 	}
 	return false
 }
@@ -327,7 +690,9 @@ func avgByGroup(m map[string][]float64) map[string]float64 {
 	for k, vals := range m {
 		if len(vals) > 0 {
 			sum := 0.0
-			for _, v := range vals { sum += v }
+			for _, v := range vals {
+				sum += v
+			}
 			result[k] = sum / float64(len(vals))
 		}
 	}
@@ -336,15 +701,17 @@ func avgByGroup(m map[string][]float64) map[string]float64 {
 
 func inferCategory(query, defaultCat string) string {
 	for cat, keywords := range map[string][]string{
-		"system_fact": {"os", "系统", "gpu", "内存", "ram", "端口", "配置"},
-		"user_pref":   {"偏好", "风格", "喜欢", "牧尘"},
-		"proj_context":{"项目", "路径", "代码", "设计文档"},
-		"tool_usage":  {"工具", "comfyui", "hermes", "opencode"},
+		"system_fact":  {"os", "系统", "gpu", "内存", "ram", "端口", "配置"},
+		"user_pref":    {"偏好", "风格", "喜欢", "牧尘"},
+		"proj_context": {"项目", "路径", "代码", "设计文档"},
+		"tool_usage":   {"工具", "comfyui", "hermes", "opencode"},
 	} {
 		for _, kw := range keywords {
 			if len(query) >= len(kw) {
 				for i := 0; i <= len(query)-len(kw); i++ {
-					if query[i:i+len(kw)] == kw { return cat }
+					if query[i:i+len(kw)] == kw {
+						return cat
+					}
 				}
 			}
 		}

go/internal/api/routes/eval_test.go

@@ -0,0 +1,493 @@
+// 织忆 MemoryWeave — 评估系统测试
+// 覆盖：金标加载 / 空断言自守 / 关键词判据 / 反证 / 全量金标跑（需活服务）
+package routes
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+)
+
+// liveAPIKey 取活服务密钥：优先 env ZHIYI_TEST_API_KEY；否则读密钥文件
+// /var/lib/memoryweave/.api-key 的主密钥（第 1 个非注释、非 legacy 行）；
+// 再退到旧 dev 键（无生产密钥文件的环境，如 CI）。
+// 2026-09-21 密钥轮换后，旧硬编码键 zhiyi-dev-key-2026 已 401 —— 全量金标实跑
+// 永远 SKIP 的根因（t_fa6c9104 修复量具可用性；不改任何判据）。
+func liveAPIKey() string {
+	if k := os.Getenv("ZHIYI_TEST_API_KEY"); k != "" {
+		return k
+	}
+	if b, err := os.ReadFile("/var/lib/memoryweave/.api-key"); err == nil {
+		for _, line := range strings.Split(string(b), "\n") {
+			line = strings.TrimSpace(line)
+			if line == "" || strings.HasPrefix(line, "#") || strings.HasPrefix(line, "legacy:") {
+				continue
+			}
+			return line
+		}
+	}
+	return "zhiyi-dev-key-2026"
+}
+
+// ─── recall 辅助：liveRecallOnce ─────────────────────────────
+// 一次 recall API 调用同时拿 top-5 的 ids + contents，避免两条 HTTP（限流保护）。
+// 每次调用间加 200ms 间隔，由调用方负责（TestEvalGolden_Live 里 sleep）。
+//
+// 统一判据口径（t_a70b07e1，与 Python 量具 golden-check.py 一致）：
+//
+//	取 recall 返回的前 N=5 条：ID 命中 = expected_ids 任一 ∈ 前 N 条；
+//	关键词命中 = 前 N 条内容任一含 must_include_keywords（大小写不敏感）。
+//	N=5 的来源：recall 请求 top_k=5（liveRecallOnce 固定），与 golden-set.json 说明字段一致。
+type recallResult struct {
+	IDS         []string
+	TopContents []string
+	Error       error
+}
+
+func liveRecallOnce(query string, maxRetries int, retryWait time.Duration) recallResult {
+	payload, _ := json.Marshal(map[string]interface{}{"query": query, "top_k": 5})
+	var lastErr error
+	for attempt := 0; attempt <= maxRetries; attempt++ {
+		if attempt > 0 {
+			time.Sleep(retryWait * time.Duration(attempt))
+		}
+		req, _ := http.NewRequest("POST", "http://127.0.0.1:7821/api/v1/recall", bytes.NewReader(payload))
+		req.Header.Set("Content-Type", "application/json")
+		req.Header.Set("X-API-Key", liveAPIKey())
+		resp, err := http.DefaultClient.Do(req)
+		if err != nil {
+			lastErr = fmt.Errorf("HTTP 请求失败: %v", err)
+			continue
+		}
+		body, _ := io.ReadAll(resp.Body)
+		resp.Body.Close()
+		if resp.StatusCode == http.StatusUnauthorized {
+			return recallResult{Error: fmt.Errorf("HTTP 401 unauthorized（API key 无效）")}
+		}
+		if resp.StatusCode != http.StatusOK {
+			// 限流（429）或服务错误：读 retry_after_s，退避重试
+			var errResp struct {
+				Error struct {
+					Code           string `json:"code"`
+					Message        string `json:"message"`
+					RetryAfterSecs int    `json:"retry_after_s"`
+				} `json:"error"`
+			}
+			_ = json.Unmarshal(body, &errResp)
+			lastErr = fmt.Errorf("HTTP %d (%s): %s", resp.StatusCode, errResp.Error.Code, errResp.Error.Message)
+			if resp.StatusCode == http.StatusTooManyRequests {
+				wait := time.Duration(errResp.Error.RetryAfterSecs) * time.Second
+				if wait == 0 {
+					wait = 2 * time.Second
+				}
+				time.Sleep(wait)
+				continue
+			}
+			return recallResult{Error: lastErr}
+		}
+		var result struct {
+			Results []struct {
+				ID      string `json:"id"`
+				Content string `json:"content"`
+			} `json:"results"`
+		}
+		if err := json.Unmarshal(body, &result); err != nil {
+			return recallResult{Error: fmt.Errorf("JSON 解析失败: %v | body=%s", err, string(body[:min(100, len(body))]))}
+		}
+		rr := recallResult{}
+		for _, r := range result.Results {
+			rr.IDS = append(rr.IDS, r.ID)
+			rr.TopContents = append(rr.TopContents, r.Content)
+		}
+		return rr
+	}
+	return recallResult{Error: lastErr}
+}
+
+// keywordHitAnywhere 检查前 N=5 条结果的 content 是否含任一关键词（统一口径 t_a70b07e1：N=5，与 golden-check.py 一致）
+func keywordHitAnywhere(topContents []string, keywords []string) bool {
+	if len(keywords) == 0 {
+		return false
+	}
+	for _, content := range topContents {
+		if checkKeywordRelevance(content, keywords) {
+			return true
+		}
+	}
+	return false
+}
+
+// ─── AC① 空断言自守：validateExpectedIDs ────────────────────
+
+func TestValidateExpectedIDs_RejectsBareMemPrefix(t *testing.T) {
+	err := validateExpectedIDs("测试查询", []string{"mem_"})
+	if err == nil {
+		t.Fatal("✗ 裸前缀 mem_ 未报错（恒真断言漏检）")
+	}
+	va, ok := err.(*VacuousAssertionError)
+	if !ok {
+		t.Fatalf("✗ 期望 VacuousAssertionError，实际 %T: %v", err, err)
+	}
+	t.Logf("✓ 裸前缀拦截：%s", va.Error())
+}
+
+func TestValidateExpectedIDs_RejectsEmptyString(t *testing.T) {
+	err := validateExpectedIDs("测试查询", []string{""})
+	if err == nil {
+		t.Fatal("✗ 空串期望值未报错")
+	}
+	t.Logf("✓ 空串拦截：%s", err.Error())
+}
+
+func TestValidateExpectedIDs_RejectsNone(t *testing.T) {
+	for _, bad := range []string{"None", "none", "NONE"} {
+		if err := validateExpectedIDs("测试查询", []string{bad}); err == nil {
+			t.Fatalf("✗ None 变体 %q 未报错", bad)
+		}
+	}
+	t.Log("✓ None 各变体全部拦截")
+}
+
+func TestValidateExpectedIDs_AllowsRealID(t *testing.T) {
+	if err := validateExpectedIDs("测试查询", []string{"mem_1789153726378940805"}); err != nil {
+		t.Fatalf("✗ 真实 ID 被误拦: %v", err)
+	}
+	t.Log("✓ 真实 ID 放行")
+}
+
+func TestValidateExpectedIDs_AllowsEmptySlice(t *testing.T) {
+	if err := validateExpectedIDs("测试查询", []string{}); err != nil {
+		t.Fatalf("✗ 空 slice（bootstrap）被误拦: %v", err)
+	}
+	t.Log("✓ 空 slice 放行（bootstrap）")
+}
+
+// ─── 关键词判据：checkKeywordRelevance ──────────────────────
+
+func TestCheckKeywordRelevance(t *testing.T) {
+	if !checkKeywordRelevance("用户偏好：牧尘喜欢直接动手", []string{"牧尘"}) {
+		t.Fatal("✗ 应命中「牧尘」")
+	}
+	if checkKeywordRelevance("用户偏好：牧尘喜欢直接动手", []string{"技术栈"}) {
+		t.Fatal("✗ 不应命中「技术栈」")
+	}
+	if !checkKeywordRelevance("MemoryWeave recall API 文档", []string{"RECALL"}) {
+		t.Fatal("✗ 大小写不敏感匹配失败")
+	}
+	if checkKeywordRelevance("任意内容", []string{}) {
+		t.Fatal("✗ 空关键词判据退化成恒真")
+	}
+	t.Log("✓ 关键词判据全部正确")
+}
+
+// ─── LoadGoldenSet：加载 + 校验 ────────────────────────────
+
+func TestLoadGoldenSet_ValidFile(t *testing.T) {
+	p := filepath.Join("testdata", "golden-set.json")
+	if _, err := os.Stat(p); err != nil {
+		t.Skipf("testdata 不在 CWD（测试需在 go/internal/api/routes/ 下跑）: %v", err)
+	}
+	gs, err := LoadGoldenSet(p)
+	if err != nil {
+		t.Fatalf("✗ 加载金标集失败: %v", err)
+	}
+	if gs.Description == "" {
+		t.Error("✗ description 缺失")
+	}
+	if len(gs.Items) != 14 {
+		t.Fatalf("✗ 期望 14 条（12 基线 + 2 目标），实际 %d", len(gs.Items))
+	}
+	for i, item := range gs.Items {
+		if item.Query == "" {
+			t.Fatalf("✗ item %d query 为空", i)
+		}
+		if len(item.ExpectedIDs) == 0 {
+			t.Fatalf("✗ item %d expected_ids 为空（金标必须有具体 ID）", i)
+		}
+		for _, id := range item.ExpectedIDs {
+			if id == "mem_" || id == "mem" || id == "" {
+				t.Fatalf("✗ item %d expected_ids 含裸前缀/空串 %q", i, id)
+			}
+		}
+		if len(item.MustIncludeKeywords) == 0 {
+			t.Fatalf("✗ item %d must_include_keywords 为空（禁止无关键词判据条目）", i)
+		}
+	}
+	t.Logf("✓ 金标集加载成功：%d 条，全部含具体 expected_ids + must_include_keywords", len(gs.Items))
+}
+
+func TestLoadGoldenSet_RejectsVacuousIDs(t *testing.T) {
+	dir := t.TempDir()
+	bad := `{
+		"description": "bad",
+		"items": [
+			{"query": "q1", "expected_ids": ["mem_"], "must_include_keywords": ["a"]}
+		]
+	}`
+	p := filepath.Join(dir, "bad.json")
+	os.WriteFile(p, []byte(bad), 0644)
+	if _, err := LoadGoldenSet(p); err == nil {
+		t.Fatal("✗ 含裸前缀 mem_ 的金标集未被拒绝")
+	} else {
+		t.Logf("✓ 含裸前缀的金标集被拒：%v", err)
+	}
+}
+
+// t_dfc93ad0 AC5 回归：部署二进制 CWD=/home/muc（无 testdata/），
+// POST /api/v1/eval/golden 不传 golden_path 时，旧版「相对 testdata/ 读不到」⇒ 500。
+// 修复后：缺省路径 / 读不到的路径 ⇒ 回落 **内嵌** 基线（//go:embed），仍可用。
+// 判据（可复算）：在**没有 testdata/ 的 CWD** 里 LoadGoldenSet("") 与
+// LoadGoldenSet("testdata/golden-set.json")（相对路径读不到）都必须成功
+// 且返回 14 条 —— 即部署态「缺省命令可用」。
+func TestLoadGoldenSet_EmbedFallback_DeployCWD(t *testing.T) {
+	empty := t.TempDir() // 模拟部署 CWD：里面没有 testdata/
+	orig, err := os.Getwd()
+	if err != nil {
+		t.Fatalf("取 CWD 失败: %v", err)
+	}
+	if err := os.Chdir(empty); err != nil {
+		t.Fatalf("chdir 到空目录失败: %v", err)
+	}
+	t.Cleanup(func() { _ = os.Chdir(orig) })
+
+	// ① 缺省（path=""）：不传 golden_path 的部署态 ⇒ 必须成功 + 14 条（内嵌兜底）
+	gs, err := LoadGoldenSet("")
+	if err != nil {
+		t.Fatalf("✗ AC5：缺省路径加载金标集失败（旧版这里 500）: %v", err)
+	}
+	if len(gs.Items) != 14 {
+		t.Fatalf("✗ AC5：内嵌基线期望 14 条，实际 %d", len(gs.Items))
+	}
+	// ② 相对路径读不到（部署 CWD 无 testdata/）：也回落内嵌，不 500
+	gs2, err := LoadGoldenSet(filepath.Join("testdata", "golden-set.json"))
+	if err != nil {
+		t.Fatalf("✗ AC5：相对路径读不到时应回落内嵌基线而非报错: %v", err)
+	}
+	if len(gs2.Items) != 14 {
+		t.Fatalf("✗ AC5：回落基线期望 14 条，实际 %d", len(gs2.Items))
+	}
+	t.Logf("✓ AC5：空 CWD（无 testdata/）下 缺省/相对路径 均回落内嵌基线（14 条），部署态缺省命令可用")
+}
+
+// ─── AC① 反证：故意把某条 expected_id 改成错误 ID → 该条必须 FAIL ───
+// 这是本卡核心验收点：证明判据非恒真（改错期望 → 真 FAIL）
+
+func TestEvalGolden_CounterEvidence(t *testing.T) {
+	// 探活 zhiyid（无服务 → 跳过）
+	healthReq, _ := http.NewRequest("GET", "http://127.0.0.1:7821/api/v1/health", nil)
+	healthReq.Header.Set("X-API-Key", liveAPIKey())
+	resp, err := http.DefaultClient.Do(healthReq)
+	if err != nil || resp == nil {
+		t.Skipf("zhiyid 服务不可达（%v），跳过反证测试", err)
+	}
+	resp.Body.Close()
+
+	// 1. 先跑一次正常（期望 PASS）
+	gs, err := LoadGoldenSet(filepath.Join("testdata", "golden-set.json"))
+	if err != nil {
+		t.Fatalf("加载金标集失败: %v", err)
+	}
+	// 统一口径 t_a70b07e1：反证测试也按 golden-set.json 的 expected_ids 判定（前 N=5 条），
+	// 保证「故意改坏输入 → 量具 FAIL」的证明与量具实际判据一致。
+
+	// 取第一条（"织忆是什么"，expected=mem_1789153726378940805）
+	item0 := gs.Items[0]
+	t.Logf("反证前：item[0] query=%q expected=%v", item0.Query, item0.ExpectedIDs)
+
+	// 反证前跑一次
+	rrBefore := liveRecallOnce(item0.Query, 3, 2*time.Second)
+	if rrBefore.Error != nil {
+		t.Skipf("recall API 不可用（%v），跳过反证", rrBefore.Error)
+	}
+	idHitBefore := containsAny(rrBefore.IDS, item0.ExpectedIDs)
+	kwHitBefore := keywordHitAnywhere(rrBefore.TopContents, item0.MustIncludeKeywords)
+	if idHitBefore && kwHitBefore {
+		t.Logf("反证前 ✓ PASS: top1=%s kw_hit=%v", firstID(rrBefore.IDS), kwHitBefore)
+	} else {
+		t.Logf("反证前（基线本身就有问题，跳过反证验证）: idHit=%v kwHit=%v", idHitBefore, kwHitBefore)
+		t.Skipf("反证前基线不 PASS，无法验证反证")
+	}
+
+	// 2. 故意改成错误 ID
+	badID := "mem_0000000000000000000" // 不存在的 ID
+	gs2 := &GoldenSet{Items: []GoldenSetItem{}}
+	for _, it := range gs.Items {
+		if it.Query == item0.Query {
+			it.ExpectedIDs = []string{badID}
+		}
+		gs2.Items = append(gs2.Items, it)
+	}
+
+	// 空断言自守：badID 是合法格式（mem_+数字），不被拦截 ✓
+	if err := validateExpectedIDs(item0.Query, []string{badID}); err != nil {
+		t.Fatalf("✗ 反证用 ID %q 被空断言自守误拦: %v", badID, err)
+	}
+
+	// 3. 反证跑：用错误 ID 必须 FAIL
+	rrAfter := liveRecallOnce(item0.Query, 3, 2*time.Second)
+	if rrAfter.Error != nil {
+		t.Fatalf("反证时 recall 失败: %v", rrAfter.Error)
+	}
+	idHitAfter := containsAny(rrAfter.IDS, []string{badID})
+	if idHitAfter {
+		t.Errorf("✗ 反证失败：错误 ID %q 竟然命中（判据恒真！）", badID)
+	} else {
+		t.Logf("反证后 ✓ FAIL（如预期）: top1=%s 不含 %s", firstID(rrAfter.IDS), badID)
+	}
+
+	// 4. 恢复（golden-set.json 没改，只是内存里改了 gs2）
+	t.Logf("反证测试完成：正常=%v 错误ID=%v", idHitBefore, !idHitAfter)
+}
+
+// ─── 全量金标跑（需活 zhiyid 服务；无服务时 skip）──────────
+// 判据：14 条中 12 基线 + 2 目标，跑真实 recall，统计 pass/fail/warn/error。
+// 统一判据口径（t_a70b07e1，与 Python 量具 golden-check.py + golden-set.json 说明字段三方一致）：
+//   取 recall 返回的前 N=5 条内容查关键词；ID 命中 = expected 任一 ∈ 前 N 条。
+//   判定：PASS = id 命中且 kw 命中；WARN = kw 命中但 id 漂移（降级）；FAIL = 其余；ERROR = recall 调用失败。
+//   验收 = 同一批金标经两量具记录（CALIBRATED 行），喂给同一分类器（classify.py）复算，分类差异 = 0。
+// 校准后真实基线 = 12/14（2 条目标知识在库但召回 top-5 不稳定 → 允许 FAIL，不许抹平）。
+// 每次调用后 200ms 间隔 + 限流退避，避免打满 zhiyid 令牌桶。
+
+func TestEvalGolden_Live(t *testing.T) {
+	// 探活 zhiyid（无服务 → 跳过，不污染 CI）
+	healthReq, _ := http.NewRequest("GET", "http://127.0.0.1:7821/api/v1/health", nil)
+	healthReq.Header.Set("X-API-Key", liveAPIKey())
+	resp, err := http.DefaultClient.Do(healthReq)
+	if err != nil || resp == nil {
+		t.Skipf("zhiyid 服务不可达（%v），跳过全量金标实跑", err)
+	}
+	defer resp.Body.Close()
+	body, _ := io.ReadAll(resp.Body)
+	if resp.StatusCode != http.StatusOK {
+		t.Skipf("zhiyid 服务 HTTP %d: %s", resp.StatusCode, string(body[:min(80, len(body))]))
+	}
+
+	gs, err := LoadGoldenSet(filepath.Join("testdata", "golden-set.json"))
+	if err != nil {
+		t.Fatalf("加载金标集失败: %v", err)
+	}
+
+	pass, fail, warn, errorCount := 0, 0, 0, 0
+	for i, item := range gs.Items {
+		// 空断言自守
+		if err := validateExpectedIDs(item.Query, item.ExpectedIDs); err != nil {
+			t.Fatalf("item %d 空断言自守失败: %v", i, err)
+		}
+
+		// 一次 HTTP 调用同时拿 top-5 ids + contents（限流保护）
+		rr := liveRecallOnce(item.Query, 3, 2*time.Second)
+		if rr.Error != nil {
+			errorCount++
+			t.Logf("ERROR[%d] %q: recall 错误 %v", i, item.Query, rr.Error)
+			t.Logf("CALIBRATED[%d] %s", i, fmt.Sprintf(
+				`{"tool":"go","i":%d,"query":%q,"top1":"","ids":null,"kw_hit":false,"id_hit":false,"error":%q}`,
+				i, item.Query, rr.Error.Error()))
+			continue
+		}
+
+		// 判据① ID 命中（前 N=5 条里有 expected_ids 之一）—— 统一口径 t_a70b07e1
+		idHit := containsAny(rr.IDS, item.ExpectedIDs)
+
+		// 判据② 关键词相关性（前 N=5 条结果 content 任一含关键词）—— 统一口径 t_a70b07e1
+		kwHit := keywordHitAnywhere(rr.TopContents, item.MustIncludeKeywords)
+
+		// 结构化记录：供统一口径分类器 classify.py 复算 go/py 分类差异（AC1 验收 = 差异 0）
+		t.Logf("CALIBRATED[%d] %s", i, fmt.Sprintf(
+			`{"tool":"go","i":%d,"query":%q,"top1":%q,"ids":%v,"kw_hit":%v,"id_hit":%v,"error":null}`,
+			i, item.Query, firstID(rr.IDS), rr.IDS, kwHit, idHit))
+
+		if idHit && kwHit {
+			pass++
+			t.Logf("PASS[%d] %q top1=%s", i, item.Query, firstID(rr.IDS))
+		} else {
+			fail++
+			why := []string{}
+			if !idHit {
+				if kwHit {
+					// ID 漂移但内容切题 → 降级为 WARN（不算硬 FAIL）
+					why = append(why, fmt.Sprintf("ID漂移(基线 %v ∉ top-5 %v, 但内容切题)",
+						item.ExpectedIDs, rr.IDS[:min(3, len(rr.IDS))]))
+					warn++
+				} else {
+					why = append(why, fmt.Sprintf("ID未命中(期望 %v, top-5 %v)",
+						item.ExpectedIDs, rr.IDS[:min(3, len(rr.IDS))]))
+				}
+			}
+			if !kwHit {
+				head := ""
+				if len(rr.TopContents) > 0 {
+					head = rr.TopContents[0][:min(40, len(rr.TopContents[0]))]
+				}
+				why = append(why, fmt.Sprintf("关键词未命中(%v) | top1_content: %s",
+					item.MustIncludeKeywords, head))
+			}
+			// ID 漂移但内容切题 → 降级为 WARN，不算硬 FAIL
+			if kwHit && !idHit {
+				fail--
+				t.Logf("WARN[%d] %q %v", i, item.Query, why)
+			} else {
+				t.Logf("FAIL[%d] %q %v | top1: %s", i, item.Query, why, firstID(rr.IDS))
+			}
+		}
+
+		// 条目间隔 200ms（限流保护，避免连续打满令牌桶）
+		time.Sleep(200 * time.Millisecond)
+	}
+
+	// 结尾打印真实数字（不写死任何期望值）
+	total := len(gs.Items)
+	t.Logf("RESULT: total=%d pass=%d fail=%d warn=%d error=%d（真实数字，无硬编码基线）",
+		total, pass, fail, warn, errorCount)
+
+	// 漏跑检查：pass+fail+warn+error 必须 = 总条数
+	if pass+fail+warn+errorCount != total {
+		t.Errorf("✗ 漏跑：pass+fail+warn+error=%d ≠ 总条数 %d", pass+fail+warn+errorCount, total)
+	}
+
+	// 全灭（pass=0）→ 判据可能已失效或金标集 ID 已过期
+	if pass == 0 {
+		t.Errorf("✗ 全灭（pass=0）——关键词判据可能失效或金标集 ID 已过期")
+	}
+
+	// 校准基线（供人工校准，不写死期望值）
+	t.Logf("校准基线：pass=%d/%d（fail=%d warn=%d error=%d）", pass, total, fail, warn, errorCount)
+
+	// error 条目 → 服务问题，不计入 fail 但必须报告
+	if errorCount > 0 {
+		t.Logf("⚠ %d 条 ERROR（recall API 调用失败，非召回质量问题）", errorCount)
+	}
+}
+
+// containsAny 检查 expectedIDs 中是否有任意一个出现在 recalledIDs 中
+func containsAny(recalledIDs []string, expectedIDs []string) bool {
+	set := makeSet(expectedIDs)
+	for _, id := range recalledIDs {
+		if set[id] {
+			return true
+		}
+	}
+	return false
+}
+
+// firstID 取 slice 第一个元素
+func firstID(ids []string) string {
+	if len(ids) == 0 {
+		return "(无)"
+	}
+	return ids[0]
+}
+
+// min 取两个整数较小值
+func min(a, b int) int {
+	if a < b {
+		return a
+	}
+	return b
+}

go/internal/api/routes/feedback.go

@@ -12,7 +12,7 @@ import (
 )
 
 type FeedbackAPI struct {
-	LanceDB  storage.LanceDB
+	LanceDB storage.LanceDB
 }
 
 func NewFeedbackAPI(ldb storage.LanceDB) *FeedbackAPI {
@@ -40,7 +40,14 @@ func (fa *FeedbackAPI) MarkUseful(w http.ResponseWriter, r *http.Request) {
 		[]string{req.MemoryID},
 		"user_feedback", "success", "",
 	)
-	fa.LanceDB.IncrementUseful(req.MemoryID)
+	if err := fa.LanceDB.IncrementUseful(req.MemoryID); err != nil {
+		log.Printf("[feedback] IncrementUseful failed %s: %v", req.MemoryID, err)
+	}
+	// GA-P0-4 (t_bb10e8bc)：用户点「有用」= 真实正向 outcome → 追加 hits+ 事件（带时间戳）。
+	// 语义定义见 internal/storage/hit_event.go（hits+ = 被真实使用/认可）。
+	if err := fa.LanceDB.AppendHitEvent(req.MemoryID, "pos", "human_feedback"); err != nil {
+		log.Printf("[feedback] AppendHitEvent(pos) failed %s: %v", req.MemoryID, err)
+	}
 
 	// Phase F: 质量监控 — 每次反馈后检查低质量记忆
 	feedbackCount := selfoptimize.Dash.UsefulCount + selfoptimize.Dash.NotUsefulCount
@@ -78,7 +85,15 @@ func (fa *FeedbackAPI) MarkNotUseful(w http.ResponseWriter, r *http.Request) {
 		[]string{req.MemoryID},
 		"user_feedback", "failure", "",
 	)
-	fa.LanceDB.IncrementNotUseful(req.MemoryID)
+	// GA-P0-1 (t_718d8707)：人工点「没用」→ 写 human_not_useful_count（不再写混写列 not_useful_count）
+	if err := fa.LanceDB.IncrementHumanNotUseful(req.MemoryID); err != nil {
+		log.Printf("[feedback] IncrementHumanNotUseful failed %s: %v", req.MemoryID, err)
+	}
+	// GA-P0-4 (t_bb10e8bc)：用户点「没用」= 真实负向 outcome → 追加 hits− 事件（带时间戳）。
+	// 注意：这只在用户人工点「没用」时追加，dedup 自动降权走 dedup_downweight_count 不追加（见 hit_event.go 排除项）。
+	if err := fa.LanceDB.AppendHitEvent(req.MemoryID, "neg", "human_feedback"); err != nil {
+		log.Printf("[feedback] AppendHitEvent(neg) failed %s: %v", req.MemoryID, err)
+	}
 
 	// Phase F: 质量监控 — negative feedback 触发低质量检测
 	feedbackCount := selfoptimize.Dash.UsefulCount + selfoptimize.Dash.NotUsefulCount
@@ -110,16 +125,18 @@ func (fa *FeedbackAPI) Deprecate(w http.ResponseWriter, r *http.Request) {
 	}
 
 	selfoptimize.Dash.RecordDeprecation()
-	fa.LanceDB.SoftDelete(req.MemoryID, req.Reason)
+	if err := fa.LanceDB.SoftDelete(req.MemoryID, req.Reason); err != nil {
+		log.Printf("[softdelete] failed id: %v", err)
+	}
 	respond(w, 200, map[string]string{"status": "deprecated", "memory_id": req.MemoryID})
 }
 
 // POST /api/v1/feedback/correct
 func (fa *FeedbackAPI) Correct(w http.ResponseWriter, r *http.Request) {
 	var req struct {
-		MemoryID      string `json:"memory_id"`
-		NewContent    string `json:"new_content"`
-		Source        string `json:"source"`
+		MemoryID   string `json:"memory_id"`
+		NewContent string `json:"new_content"`
+		Source     string `json:"source"`
 	}
 	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
 		respondError(w, 400, "invalid body")

go/internal/api/routes/forget_guard_e2e_test.go

@@ -0,0 +1,143 @@
+// Package routes — 遗忘保护守卫端到端回归测试（t_944344b5，AC1/AC2）。
+//
+// 2026-10-16 (t_944344b5)：证明「decay/auto_forget 的判定道（admin Forget 路由）命中保护
+// 规则的候选**不被** SoftDelete，而非保护的衰减候选**照常**被软删」。
+//
+// 必须走**真实判定路径**（routes.NewAdminAPI.Forget），不能只测 memprotect.IsProtected，
+// 因为守卫是在 decay 循环 / admin Forget 路由里 SoftDelete 之前拦截的（见 server.go / admin.go）。
+//
+// 跑法（A/B 纪律：复制数据目录 + 独立 socket，绝不碰生产 /var/lib/memoryweave）：
+//   1. cp -a /var/lib/memoryweave /tmp/mw-prot-copy-<ts>
+//   2. zhiyi-consolidate --mode socket --socket /tmp/mw-prot.sock \
+//        --data-dir /tmp/mw-prot-copy-<ts> &
+//   3. PROT_TEST_SOCK=/tmp/mw-prot.sock \
+//        go test ./internal/api/routes/ -run TestForgetGuardE2E -v -count=1
+//
+// 未设 PROT_TEST_SOCK → 自动跳过（可安全留在 CI）。
+package routes_test
+
+import (
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"testing"
+	"time"
+
+	"github.com/xiaoxue/memoryweave/internal/governance"
+	"github.com/xiaoxue/memoryweave/internal/models"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+
+	"github.com/xiaoxue/memoryweave/internal/api/routes"
+)
+
+func TestForgetGuardE2E(t *testing.T) {
+	sock := os.Getenv("PROT_TEST_SOCK")
+	if sock == "" {
+		t.Skip("PROT_TEST_SOCK 未设置，跳过（需先起 test sidecar 指向复制的数据目录）")
+	}
+
+	// 让遗忘窗口游标指到临时文件（缺失 → 空游标 → 从最小 id 起轮转），
+	// 保证我们构造的 `mem_0...` 行（id 最小）落在首个 2000 窗口内。
+	// 不碰生产 /var/lib/memoryweave/forget-cursor.json。
+	cursorDir := t.TempDir()
+	cursorFile := cursorDir + "/test-forget-cursor.json"
+	t.Setenv("ZHIYI_FORGET_CURSOR", cursorFile)
+
+	// embedder 传 nil：记录自带 1024 维 Vector，走不到编码分支（同 conf_e2e_test.go 做法）。
+	rc := storage.NewRustLanceDBClient(sock, nil)
+	f := governance.NewForgetter()
+	aa := routes.NewAdminAPI(rc, f, nil) // GraphStore 传 nil：extractTopEntityDegree 空安全
+
+	// ── 构造 4 条测试记忆（id 都以 mem_0 开头 → 字典序最小 → 必在首窗口）──
+	vec := make([]float32, 1024)
+	for i := range vec {
+		vec[i] = 0.001
+	}
+	now := time.Now().UTC()
+	mk := func(id, category, content, freshness string, daysAgo int) models.MemoryRecord {
+		lr := now.Add(-time.Duration(daysAgo) * 24 * time.Hour)
+		return models.MemoryRecord{
+			ID: id, AgentID: "hermes-a06", Namespace: "prot-test",
+			Content: content, Category: category, Vector: vec,
+			Tier: "normal", Importance: 0.5, QualityScore: 0.5,
+			Version: 1, Freshness: freshness,
+			Source: "wiki:/tmp/forget-guard-e2e.md",
+			EmbeddingModel: "bge-m3", ScriptVersion: "forget-guard-e2e",
+			CreatedAt: lr, UpdatedAt: lr, LastRecalledAt: lr,
+			RecallVerifiedCount: 0,
+		}
+	}
+
+	// P1：规则3 freshness=verified 保护。content <30 字 + >20 天 ⇒ 本应走碎片快速道被删，
+	//     但 freshness=verified ⇒ 守卫拦截 ⇒ 存活。
+	p1 := mk("mem_0aaaa_p1", "distilled", "verified fact, kept", "verified", 30)
+	// P2：规则1 category=decision 保护。
+	p2 := mk("mem_0aaaa_p2", "decision", "short decision note", "fresh", 30)
+	// P3：规则4 content 关键词「牧尘要求」保护。
+	p3 := mk("mem_0aaaa_p3", "distilled", "牧尘要求：部署前必须先备份", "fresh", 30)
+	// N1：不命中任何规则（distilled / fresh / 无关键词 / 低 V / 不在名单）⇒ 应被碎片快速道软删。
+	n1 := mk("mem_0aaaa_n1", "distilled", "hi", "fresh", 45)
+
+	for _, m := range []models.MemoryRecord{p1, p2, p3, n1} {
+		if err := rc.InsertMemory(m); err != nil {
+			t.Fatalf("InsertMemory(%s) 失败: %v", m.ID, err)
+		}
+	}
+
+	// ── 跑真实判定路径：admin Forget 路由（= decay/auto_forget 决策 + 守卫 + SoftDelete）──
+	req := httptest.NewRequest(http.MethodPost, "/api/v1/admin/forget", nil)
+	w := httptest.NewRecorder()
+	aa.Forget(w, req)
+	if w.Code != http.StatusOK {
+		t.Fatalf("admin forget expected 200, got %d: %s", w.Code, w.Body.String())
+	}
+	var resp map[string]interface{}
+	json.Unmarshal(w.Body.Bytes(), &resp)
+	t.Logf("admin forget 响应: %v", resp)
+
+	// ── 断言 ──
+	// 读路径语义（rust lancedb_ops.rs）：
+	//   scan_all     → 仅 is_deleted=false（存活集，无 5000 上限）
+	//   scan_deleted → 仅 is_deleted=true（已删集，带 delete_reason）
+	// 故 P1/P2/P3 应出现在 ScanAll 里；N1 应不在 ScanAll、而在 ScanDeleted 里。
+	all, err := rc.ScanAll()
+	if err != nil {
+		t.Fatalf("ScanAll 失败: %v", err)
+	}
+	liveSet := map[string]bool{}
+	for _, m := range all {
+		liveSet[m.ID] = true // scan_all 只回 is_deleted=false
+	}
+	for _, id := range []string{p1.ID, p2.ID, p3.ID} {
+		if !liveSet[id] {
+			t.Errorf("❌ 受保护记忆 %s 应存活（ScanAll 命中），实际不在 —— 守卫失效", id)
+		}
+	}
+	if liveSet[n1.ID] {
+		t.Errorf("❌ 非保护衰减记忆 %s 应被软删（不在 ScanAll），实际仍存活 —— 回归异常", n1.ID)
+	}
+
+	deleted, derr := rc.ScanDeleted()
+	if derr != nil {
+		t.Fatalf("ScanDeleted 失败: %v", derr)
+	}
+	deletedSet, n1Reason := map[string]bool{}, ""
+	for _, m := range deleted {
+		deletedSet[m.ID] = true
+		if m.ID == n1.ID {
+			n1Reason = m.DeleteReason
+		}
+	}
+	if !deletedSet[n1.ID] {
+		t.Errorf("❌ 非保护衰减记忆 %s 应出现在 ScanDeleted（is_deleted=true），实际不在", n1.ID)
+	}
+	if t.Failed() {
+		return
+	}
+	// 受保护 3 条存活 + 非保护 1 条被删，打印结论
+	t.Logf("✅ AC1 3 条命中保护规则（freshness=verified / category=decision / 关键词）全部存活于 ScanAll")
+	t.Logf("✅ AC2 非保护衰减碎片 %s 照常软删（在 ScanDeleted）", n1.ID)
+	// 取证：N1 的 delete_reason（auto_forget*/fragment 经 NormalizeDeleteReason → retention）
+	t.Logf("取证: %s delete_reason=%q", n1.ID, n1Reason)
+}

go/internal/api/routes/gap_repair.go

@@ -14,19 +14,19 @@ import (
 
 // GapAutoRepair 缺口自动修复器
 type GapAutoRepair struct {
-	detector  *selfoptimize.GapDetector
-	ldb       storage.LanceDB
-	synonyms  map[string][]string // 同义词映射
+	detector *selfoptimize.GapDetector
+	ldb      storage.LanceDB
+	synonyms map[string][]string // 同义词映射
 }
 
 var GapRepair = &GapAutoRepair{
 	detector: nil, // 由 InitGapRepair 注入
 	synonyms: map[string][]string{
-		"GPU":   {"gpu", "显卡", "graphics"},
-		"OS":    {"os", "操作系统", "系统"},
-		"API":   {"api", "接口", "endpoint"},
-		"SSE":   {"sse", "server-sent events", "事件推送"},
-		"CRDT":  {"crdt", "冲突合并", "merge"},
+		"GPU":  {"gpu", "显卡", "graphics"},
+		"OS":   {"os", "操作系统", "系统"},
+		"API":  {"api", "接口", "endpoint"},
+		"SSE":  {"sse", "server-sent events", "事件推送"},
+		"CRDT": {"crdt", "冲突合并", "merge"},
 	},
 }
 

go/internal/api/routes/gaps.go

@@ -4,6 +4,7 @@ package routes
 import (
 	"encoding/json"
 	"net/http"
+	"strings"
 
 	"github.com/xiaoxue/memoryweave/internal/selfoptimize"
 )
@@ -32,17 +33,31 @@ func (ga *GapAPI) List(w http.ResponseWriter, r *http.Request) {
 	})
 }
 
-// POST /api/v1/gaps/{topic}/close
+// POST /api/v1/gaps/close/{topic}（现行前缀注册 /api/v1/gaps/close/；亦兼容 {topic} 通配注册）
 func (ga *GapAPI) Close(w http.ResponseWriter, r *http.Request) {
 	topic := r.PathValue("topic")
+	if topic == "" {
+		// 现行注册是前缀路由（模式无 {topic} 通配 ⇒ PathValue 恒为 ""，端点此前恒 400）：
+		// 回落剥前缀取 topic。若日后改注册为 /api/v1/gaps/{topic}/close，PathValue 直接命中。
+		const prefix = "/api/v1/gaps/close/"
+		if strings.HasPrefix(r.URL.Path, prefix) {
+			topic = r.URL.Path[len(prefix):]
+		}
+	}
 	if topic == "" {
 		respondError(w, 400, "topic required")
 		return
 	}
 
-	ga.Detector.Close(topic)
-	selfoptimize.Dash.RecordGapClosed()
-	respond(w, 200, map[string]string{"status": "closed", "topic": topic})
+	// t_e5997ca4：计数口径 = 「缺口个数」。+1 只发生在 GapDetector 的 open→closed 迁移点
+	// （Close 内部）；旧的无条件 selfoptimize.Dash.RecordGapClosed() 已删除 —— 那是同一
+	// 缺口可被无限计数、ClosedGaps(9) 越过 TotalGaps(2) 的写侧根因 #1。
+	switch ga.Detector.Close(topic) {
+	case true:
+		respond(w, 200, map[string]string{"status": "closed", "topic": topic})
+	default:
+		respond(w, 200, map[string]string{"status": "already_closed", "topic": topic})
+	}
 }
 
 // POST /api/v1/gaps/detect — 手动触发缺口检测（miss 计数）

go/internal/api/routes/gaps_close_test.go

@@ -0,0 +1,112 @@
+// 织忆 MemoryWeave — gap 关闭端点写侧回归（t_e5997ca4 卡 AC2）
+//
+// POST /close 打两回（httptest）⇒ gap 计数器只 +1；auto 等其余计数器不受影响。
+package routes
+
+import (
+	"encoding/json"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/xiaoxue/memoryweave/internal/selfoptimize"
+)
+
+// makeGapTopic 3 次 miss 命中默认阈值 3 ⇒ 生成缺口（TotalGaps +1）。
+func makeGapTopic(t *testing.T, gd *selfoptimize.GapDetector, topic string) {
+	t.Helper()
+	var gap *selfoptimize.Gap
+	for i := 0; i < 3; i++ {
+		if gap = gd.RecordMiss(topic); gap != nil {
+			break
+		}
+	}
+	if gap == nil {
+		t.Fatalf("expected gap for %q after 3 misses", topic)
+	}
+}
+
+func postGapClose(t *testing.T, url string) (int, map[string]string) {
+	t.Helper()
+	resp, err := http.Post(url, "application/json", nil)
+	if err != nil {
+		t.Fatalf("POST %s: %v", url, err)
+	}
+	defer resp.Body.Close()
+	var body map[string]string
+	if err := json.NewDecoder(resp.Body).Decode(&body); err != nil {
+		b, _ := io.ReadAll(resp.Body)
+		t.Fatalf("POST %s: 非 JSON 响应 %q: %v", url, b, err)
+	}
+	return resp.StatusCode, body
+}
+
+// 卡 AC2：现行注册形态（前缀路由，topic 在 /api/v1/gaps/close/ 之后）打两回 ⇒ 只 +1。
+func TestGapAPI_CloseTwice_CounterOnlyOnce(t *testing.T) {
+	gd := selfoptimize.NewGapDetector(nil, nil)
+	makeGapTopic(t, gd, "close-twice-topic")
+
+	baseClosed := selfoptimize.Dash.ClosedGapsCount()
+	baseTotal := selfoptimize.Dash.TotalGapsCount()
+	baseMetrics := selfoptimize.Dash.Metrics() // auto_* 快照（应不受影响）
+
+	api := NewGapAPI(gd)
+	mux := http.NewServeMux()
+	mux.HandleFunc("/api/v1/gaps/close/", api.Close) // 与 server.go:1205 现行注册同形
+	ts := httptest.NewServer(mux)
+	defer ts.Close()
+
+	code1, body1 := postGapClose(t, ts.URL+"/api/v1/gaps/close/close-twice-topic")
+	if code1 != 200 || body1["status"] != "closed" {
+		t.Fatalf("第 1 次 POST: got %d %v, want 200 closed", code1, body1)
+	}
+	code2, body2 := postGapClose(t, ts.URL+"/api/v1/gaps/close/close-twice-topic")
+	if code2 != 200 || body2["status"] != "already_closed" {
+		t.Fatalf("第 2 次 POST: got %d %v, want 200 already_closed（不得再动计数）", code2, body2)
+	}
+
+	if got := selfoptimize.Dash.ClosedGapsCount() - baseClosed; got != 1 {
+		t.Fatalf("2× POST close 后 ClosedGaps 增量: got %d, want 1（改前为 +3 ⇒ 虚增）", got)
+	}
+	if got := selfoptimize.Dash.TotalGapsCount(); got != baseTotal {
+		t.Errorf("TotalGaps 不应被 close 端点移动: got %d want %d", got, baseTotal)
+	}
+
+	after := selfoptimize.Dash.Metrics()
+	if after["gap_closure_rate"] > 1 {
+		t.Errorf("gap_closure_rate = %f 越界", after["gap_closure_rate"])
+	}
+	// auto（冲突）计数器与 gap 关闭事件无关 ⇒ 快照不变（「auto 不适用」）
+	for _, k := range []string{"auto_resolve_rate", "auto_resolve_samples", "auto_resolve_valid"} {
+		if after[k] != baseMetrics[k] {
+			t.Errorf("auto 计数器被 gap 关闭污染: %s got %f want %f", k, after[k], baseMetrics[k])
+		}
+	}
+}
+
+// {topic} 通配注册形态（若日后把 server.go 注册改成 /api/v1/gaps/{topic}/close）：
+// PathValue 直接命中，计数语义不变（+1 只在首次）。
+func TestGapAPI_CloseWildcardRegistration(t *testing.T) {
+	gd := selfoptimize.NewGapDetector(nil, nil)
+	makeGapTopic(t, gd, "wildcard-topic")
+
+	baseClosed := selfoptimize.Dash.ClosedGapsCount()
+
+	api := NewGapAPI(gd)
+	mux := http.NewServeMux()
+	mux.HandleFunc("/api/v1/gaps/{topic}/close", api.Close)
+	ts := httptest.NewServer(mux)
+	defer ts.Close()
+
+	code, body := postGapClose(t, ts.URL+"/api/v1/gaps/wildcard-topic/close")
+	if code != 200 || body["status"] != "closed" {
+		t.Fatalf("通配注册第 1 次 POST: got %d %v, want 200 closed", code, body)
+	}
+	if _, body2 := postGapClose(t, ts.URL+"/api/v1/gaps/wildcard-topic/close"); body2["status"] != "already_closed" {
+		t.Errorf("通配注册第 2 次 POST: got %v, want already_closed", body2)
+	}
+	if got := selfoptimize.Dash.ClosedGapsCount() - baseClosed; got != 1 {
+		t.Fatalf("通配注册 2× POST 后 ClosedGaps 增量: got %d, want 1", got)
+	}
+}

go/internal/api/routes/graph.go

@@ -66,17 +66,18 @@ func (ga *GraphAPI) Query(w http.ResponseWriter, r *http.Request) {
 
 // POST /api/v1/graph/navigate
 // 支持三种模式：
-//   1. 单实体 BFS: {"entity": "...", "max_hops": 2}
-//   2. 双向 BFS: {"source": "...", "target": "...", "max_hops": 3}
-//   3. 关系过滤 BFS: {"entity": "...", "max_hops": 2, "relation_filter": ["related_to", "uses"]}（E1.5）
+//  1. 单实体 BFS: {"entity": "...", "max_hops": 2}
+//  2. 双向 BFS: {"source": "...", "target": "...", "max_hops": 3}
+//  3. 关系过滤 BFS: {"entity": "...", "max_hops": 2, "relation_filter": ["related_to", "uses"]}（E1.5）
+//
 // 返回新增 grouped_by_relation 字段，按关系类型分组，便于阅读
 func (ga *GraphAPI) Navigate(w http.ResponseWriter, r *http.Request) {
 	var req struct {
-		Entity        string   `json:"entity"`
-		Source        string   `json:"source"`
-		Target        string   `json:"target"`
-		MaxHops       int      `json:"max_hops"`
-		Namespace     string   `json:"namespace"`
+		Entity         string   `json:"entity"`
+		Source         string   `json:"source"`
+		Target         string   `json:"target"`
+		MaxHops        int      `json:"max_hops"`
+		Namespace      string   `json:"namespace"`
 		RelationFilter []string `json:"relation_filter"` // E1.5: 关系类型白名单
 	}
 	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
@@ -87,6 +88,15 @@ func (ga *GraphAPI) Navigate(w http.ResponseWriter, r *http.Request) {
 		req.MaxHops = 2
 	}
 
+	// 响应体积上限：与 /graph/notes 同因同治（2026-09-27 小怡）。
+	// 实测 navigate("牧尘", 2) 服务端吐 34,465 条 ≈3.4MB，再加 grouped_by_relation 同量级复制
+	// ⇒ 单次响应 ~7MB。全仓调用点核对（3 处）都只读样本，paths 全量无人消费：
+	//   - hermes 插件 plugins/zhiyi/__init__.py:911 → for p in paths[:50]
+	//   - web-ui/index.html → 只读 d.count（不渲染 paths）
+	//   - openclaw 客户端 → 已退役
+	// ⇒ paths 截断为可观测样本；count 仍报真实总数（调用方用它显示 "Found N paths"）。
+	const maxNavigatePaths = 200
+
 	// 模式 2: 双向 BFS（source + target）
 	if req.Source != "" && req.Target != "" {
 		source := NormalizeEntity(req.Source)
@@ -96,9 +106,14 @@ func (ga *GraphAPI) Navigate(w http.ResponseWriter, r *http.Request) {
 			respondError(w, 500, "navigate failed: "+err.Error())
 			return
 		}
+		totalPaths := len(paths)
+		if totalPaths > maxNavigatePaths {
+			paths = paths[:maxNavigatePaths]
+		}
 		respond(w, 200, map[string]interface{}{
-			"paths":           paths, "source": req.Source, "target": req.Target,
-			"bidirectional":   true, "count": len(paths),
+			"paths": paths, "source": req.Source, "target": req.Target,
+			"bidirectional": true, "count": totalPaths,
+			"returned": len(paths), "truncated": totalPaths > len(paths),
 			"relation_filter": req.RelationFilter,
 		})
 		return
@@ -116,6 +131,12 @@ func (ga *GraphAPI) Navigate(w http.ResponseWriter, r *http.Request) {
 		return
 	}
 
+	// paths 截断（见函数开头 maxNavigatePaths 说明）；grouped/suggestions 基于截断后样本构建。
+	totalPaths := len(paths)
+	if totalPaths > maxNavigatePaths {
+		paths = paths[:maxNavigatePaths]
+	}
+
 	// 新增：按 relation 分组，更直观
 	grouped := make(map[string][]map[string]interface{})
 	for _, p := range paths {
@@ -146,14 +167,16 @@ func (ga *GraphAPI) Navigate(w http.ResponseWriter, r *http.Request) {
 	}
 
 	respond(w, 200, map[string]interface{}{
-		"paths":                   paths,
-		"grouped_by_relation":     grouped,
-		"entity":                  req.Entity,
-		"normalized_entity":      entity,
-		"count":                   len(paths),
-		"relation_count":          len(grouped),
-		"suggestions":             suggestions,
-		"relation_filter":        req.RelationFilter,
+		"paths":               paths,
+		"grouped_by_relation": grouped,
+		"entity":              req.Entity,
+		"normalized_entity":   entity,
+		"count":               totalPaths,
+		"returned":            len(paths),
+		"truncated":           totalPaths > len(paths),
+		"relation_count":      len(grouped),
+		"suggestions":         suggestions,
+		"relation_filter":     req.RelationFilter,
 	})
 }
 

go/internal/api/routes/health.go

@@ -4,9 +4,19 @@ package routes
 import (
 	"encoding/json"
 	"net/http"
+
+	"github.com/xiaoxue/memoryweave/internal/consolidate"
 )
 
+// sidecarHealthCheck 探测 Rust sidecar（zhiyi-consolidate）可达性。
+// 声明为变量以便测试注入，默认走真实 IPC 探测。
+var sidecarHealthCheck = consolidate.HealthCheck
+
 // HandleHealth GET /health — 无需认证
+//
+// 深健康检查：不仅报告自身存活，还探测 IPC sidecar 是否可达。
+// 背景（2026-10-08 故障）：sidecar 异常时 /health 仍返回 200 —— "假绿"，
+// 使部署失败被静默数小时，只能靠 pgrep 才发现。深检查把依赖故障暴露为 503。
 func HandleHealth(w http.ResponseWriter, r *http.Request) {
 	if r.Method != http.MethodGet {
 		http.Error(w, `{"error":"method not allowed"}`, http.StatusMethodNotAllowed)
@@ -14,10 +24,23 @@ func HandleHealth(w http.ResponseWriter, r *http.Request) {
 	}
 
 	w.Header().Set("Content-Type", "application/json")
-	w.WriteHeader(http.StatusOK)
-	json.NewEncoder(w).Encode(map[string]string{
+
+	body := map[string]string{
 		"status":  "ok",
 		"service": "zhiyid",
 		"version": "0.1.0",
-	})
+	}
+
+	if err := sidecarHealthCheck(); err != nil {
+		body["status"] = "degraded"
+		body["sidecar"] = "unreachable"
+		body["error"] = err.Error()
+		w.WriteHeader(http.StatusServiceUnavailable)
+		json.NewEncoder(w).Encode(body)
+		return
+	}
+	body["sidecar"] = "ok"
+
+	w.WriteHeader(http.StatusOK)
+	json.NewEncoder(w).Encode(body)
 }

go/internal/api/routes/l3.go

@@ -26,17 +26,17 @@ type WorldModel struct {
 
 	Services []ServiceInfo `json:"services"` // ℐ: 服务和端口
 
-	ImplicitRules []string `json:"implicit_rules"`      // C: 软约束
-	HardConstraints []string `json:"hard_constraints"`   // C: 硬约束
-	Capabilities   []string `json:"capabilities"`
+	ImplicitRules   []string `json:"implicit_rules"`   // C: 软约束
+	HardConstraints []string `json:"hard_constraints"` // C: 硬约束
+	Capabilities    []string `json:"capabilities"`
 
 	UpdatedAt string `json:"updated_at"`
 }
 
 type ServiceInfo struct {
-	Name    string   `json:"name"`
-	Port    int      `json:"port"`
-	Status  string   `json:"status"`
+	Name      string   `json:"name"`
+	Port      int      `json:"port"`
+	Status    string   `json:"status"`
 	DependsOn []string `json:"depends_on,omitempty"`
 	Provides  []string `json:"provides,omitempty"`
 }
@@ -111,13 +111,13 @@ func (wm *WorldModel) GetHandler(w http.ResponseWriter, r *http.Request) {
 // POST /api/v1/l3/worldmodel — 更新
 func (wm *WorldModel) UpdateHandler(w http.ResponseWriter, r *http.Request) {
 	var req struct {
-		Environment   *struct {
+		Environment *struct {
 			OS   string `json:"os"`
 			GPU  string `json:"gpu"`
 			VRAM string `json:"vram"`
 			RAM  string `json:"ram"`
 		} `json:"environment"`
-		AddRules    []string `json:"add_rules"`
+		AddRules        []string `json:"add_rules"`
 		AddCapabilities []string `json:"add_capabilities"`
 	}
 	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {

go/internal/api/routes/memory_write_test.go

@@ -0,0 +1,222 @@
+// 织忆 MemoryWeave — /api/v1/memory/write 显式直写 memories 表（主症A，t_fa6c9104，2026-10-16）
+//
+// 锁定行为：
+//
+//	W1 写入落 memories 表：InsertMemory 被调用，id 前缀 mem_，tier=normal，quality=1.0，
+//	    且向量非全 0（嵌入服务故障时不写零向量记录）
+//	W2 category 缺省 distilled（recall Step 2.5 rawCategories 不过滤它）
+//	W3 嵌入失败 ⇒ 502 且不写（零向量 = 主症A 病根，禁静默写）
+//	W4 精确重复 ⇒ merged（不新增行）
+//	W5 短内容 <8 字 ⇒ 400；GET ⇒ 405
+//	W6 零回归：commit 路径仍只写 episodes、不写 memories（AC5）
+package routes
+
+import (
+	"bytes"
+	"encoding/json"
+	"errors"
+	"net/http"
+	"net/http/httptest"
+	"strconv"
+	"strings"
+	"sync"
+	"testing"
+
+	"github.com/xiaoxue/memoryweave/internal/distill"
+	"github.com/xiaoxue/memoryweave/internal/governance"
+	"github.com/xiaoxue/memoryweave/internal/models"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// recLance 记录型 LanceDB mock：断言 InsertMemory/Update 的实际调用与载荷。
+// 复用 zz_commit_conflict_gate_test.go 的 mockGraph / newMockEmbedServer / newMockLLM / commitWith（同包）。
+type recLance struct {
+	storage.LanceDB // 未实现的方法落到接口兜底（panic 即测试失败，不放任）
+
+	mu       sync.Mutex
+	seeds    []models.MemoryRecord // 预置「库里已有」的记忆（Search 返回）
+	inserted []models.MemoryRecord // 实际 InsertMemory 调用（W1/W3/W4 断言）
+	updates  []string              // Update(table,id,...) 的 id 序列（merge 断言）
+	failIns  bool                 // 模拟 IPC 落库失败（500 断言）
+}
+
+func (r *recLance) Search(table string, vector []float32, topK int, ns string) ([]models.MemoryRecord, error) {
+	if table != "memories" {
+		return nil, nil
+	}
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	out := make([]models.MemoryRecord, 0, len(r.seeds))
+	out = append(out, r.seeds...)
+	if topK > 0 && len(out) > topK {
+		out = out[:topK]
+	}
+	return out, nil
+}
+
+func (r *recLance) InsertMemory(m models.MemoryRecord) error {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	if r.failIns {
+		return errors.New("ipc down (simulated)")
+	}
+	r.inserted = append(r.inserted, m)
+	r.seeds = append(r.seeds, m) // 后续 Search 能看见（近重复语义）
+	return nil
+}
+
+func (r *recLance) Update(table, id string, fields map[string]any) error {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	r.updates = append(r.updates, id)
+	return nil
+}
+
+func (r *recLance) InsertEpisode(agentID, namespace, content, category string) (string, error) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	return "ep_rec_" + strconv.Itoa(len(r.inserted)), nil
+}
+
+func newRecAPI(t *testing.T, seeds ...models.MemoryRecord) (*API, *recLance) {
+	t.Helper()
+	emb := newMockEmbedServer()
+	t.Cleanup(emb.Close)
+	ldb := &recLance{seeds: seeds}
+	return NewAPI(ldb, storage.NewEmbedder(emb.URL), nil, governance.NewConflictDetector(), mockGraph{}), ldb
+}
+
+func writeMemRequest(t *testing.T, api *API, body map[string]any) (int, map[string]interface{}) {
+	t.Helper()
+	b, _ := json.Marshal(body)
+	req := httptest.NewRequest("POST", "/api/v1/memory/write", bytes.NewReader(b))
+	req.Header.Set("Content-Type", "application/json")
+	w := httptest.NewRecorder()
+	api.WriteMemory(w, req)
+	var resp map[string]interface{}
+	json.Unmarshal(w.Body.Bytes(), &resp)
+	return w.Code, resp
+}
+
+// W1+W2：写入落 memories 表，mem_ 前缀，category 缺省 distilled，向量非全 0
+func TestWriteMemory_InsertLandsInMemories(t *testing.T) {
+	api, ldb := newRecAPI(t)
+	code, resp := writeMemRequest(t, api, map[string]any{
+		"content":  "牧尘的工作所在省份是河南省",
+		"agent_id": "hermes-a06",
+	})
+	if code != 201 {
+		t.Fatalf("expected 201, got %d: %s", code, resp)
+	}
+	if resp["status"] != "ok" || resp["table"] != "memories" {
+		t.Errorf("expected status=ok table=memories, got %v", resp)
+	}
+	if !strings.HasPrefix(resp["memory_id"].(string), "mem_") {
+		t.Errorf("memory_id 须 mem_ 前缀（落 memories 表），got %v", resp["memory_id"])
+	}
+	if resp["category"] != "distilled" {
+		t.Errorf("category 缺省须 distilled（Step 2.5 不过滤），got %v", resp["category"])
+	}
+	if len(ldb.inserted) != 1 {
+		t.Fatalf("expected exactly 1 InsertMemory, got %d", len(ldb.inserted))
+	}
+	m := ldb.inserted[0]
+	if m.Category != "distilled" || m.Tier != "normal" || m.QualityScore != 1.0 {
+		t.Errorf("row 字段: category=%s tier=%s quality=%v（期望 distilled/normal/1.0）", m.Category, m.Tier, m.QualityScore)
+	}
+	if m.Namespace != "hermes-main" {
+		t.Errorf("namespace 推导: got %s 期望 hermes-main（hermes-a06 ⇒ deriveNamespace）", m.Namespace)
+	}
+	if len(m.Vector) != 1024 {
+		t.Fatalf("vector 维度须 1024，got %d", len(m.Vector))
+	}
+	// W1 核心判据：vector 非全 0
+	nonzero := 0
+	for _, f := range m.Vector {
+		if f != 0 {
+			nonzero++
+		}
+	}
+	if nonzero == 0 {
+		t.Errorf("向量全 0 —— 主症A 病根复现（禁写零向量记录）")
+	}
+}
+
+// W3：嵌入失败 ⇒ 502 且不写库（fail-close，不静默写零向量）
+func TestWriteMemory_EmbedFail_502_NoWrite(t *testing.T) {
+	failEmb := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		http.Error(w, "embed backend down", http.StatusInternalServerError)
+	}))
+	defer failEmb.Close()
+	ldb := &recLance{}
+	api := NewAPI(ldb, storage.NewEmbedder(failEmb.URL), nil, governance.NewConflictDetector(), mockGraph{})
+	code, resp := writeMemRequest(t, api, map[string]any{"content": "牧尘的工作所在省份是河南省"})
+	if code != 502 {
+		t.Fatalf("embed 失败须 502，got %d: %v", code, resp)
+	}
+	if len(ldb.inserted) != 0 {
+		t.Errorf("嵌入失败不得写库（零向量 = 主症A），inserted=%d", len(ldb.inserted))
+	}
+}
+
+// W4：精确重复 ⇒ merged，不新增行
+func TestWriteMemory_ExactDupMerges(t *testing.T) {
+	seed := models.MemoryRecord{ID: "mem_seed", Content: "家用服务器地址是 192.168.123.11", Category: "distilled"}
+	api, ldb := newRecAPI(t, seed)
+	code, resp := writeMemRequest(t, api, map[string]any{"content": seed.Content})
+	if code != 200 || resp["status"] != "merged" {
+		t.Fatalf("expected 200 merged, got %d %v", code, resp)
+	}
+	if len(ldb.inserted) != 0 {
+		t.Errorf("精确重复不得新增行，inserted=%d", len(ldb.inserted))
+	}
+	if len(ldb.updates) == 0 || ldb.updates[0] != "mem_seed" {
+		t.Errorf("merge 须 Update 既有行 mem_seed，updates=%v", ldb.updates)
+	}
+}
+
+// W5：短内容 400；GET 405
+func TestWriteMemory_Validation(t *testing.T) {
+	api, _ := newRecAPI(t)
+	if c, _ := writeMemRequest(t, api, map[string]any{"content": "太短了"}); c != 400 {
+		t.Errorf("短内容(<8字)须 400, got %d", c)
+	}
+	req := httptest.NewRequest("GET", "/api/v1/memory/write", nil)
+	w := httptest.NewRecorder()
+	api.WriteMemory(w, req)
+	if w.Code != 405 {
+		t.Errorf("GET 须 405, got %d", w.Code)
+	}
+}
+
+// 落库失败（IPC down）⇒ 500 不吞错
+func TestWriteMemory_InsertFail_500(t *testing.T) {
+	api, ldb := newRecAPI(t)
+	ldb.failIns = true
+	code, _ := writeMemRequest(t, api, map[string]any{"content": "测试落库失败场景内容足够长"})
+	if code != 500 {
+		t.Fatalf("InsertMemory 失败须 500（不吞错），got %d", code)
+	}
+}
+
+// W6 零回归：commit 路径仍「只写 episodes、不写 memories」（等蒸馏）
+func TestCommit_StillEpisodesOnly_ZeroRegression(t *testing.T) {
+	api, ldb := newRecAPI(t)
+	llm := newMockLLM()
+	defer llm.server.Close()
+	oldRef := DistillEngineRef
+	DistillEngineRef = distill.NewEngine(llm.server.URL, "mock-model", "")
+	defer func() { DistillEngineRef = oldRef }()
+
+	// 复用 commitWith（同包 zz 测试）：普通对话提交
+	resp := commitWith(t, api, "今天天气不错，我们去楼下散步", false)
+	if resp["status"] != "ok" {
+		t.Fatalf("commit 基线须 status=ok，got %v", resp)
+	}
+	if resp["memory_id"] != "" {
+		t.Errorf("commit 不在写入时创建 memory（AC5 零回归），got %v", resp["memory_id"])
+	}
+	if len(ldb.inserted) != 0 {
+		t.Errorf("commit 不得直写 memories 表（零回归），inserted=%d", len(ldb.inserted))
+	}
+}

go/internal/api/routes/metanote_gate_test.go

@@ -0,0 +1,55 @@
+package routes
+
+// t_5ad3f843 AC8：handler 级门禁验证。
+// 门禁开 ⇒ meta-note 在**触碰 storage 之前**被 422 拒绝：
+// 零值 API{}（LanceDB 接口为 nil）足证——若门禁按设计在写路径前生效，
+// 422 先行返回、不会 nil 解引用 panic；若门禁失效走到 InsertEpisode ⇒ panic。
+
+import (
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"strings"
+	"testing"
+
+	"github.com/xiaoxue/memoryweave/internal/metanote"
+)
+
+func TestCommitGate_MetaRejectedBeforeStorage(t *testing.T) {
+	old := os.Getenv("ZHIYI_META_FILTER")
+	os.Setenv("ZHIYI_META_FILTER", "1")
+	defer os.Setenv("ZHIYI_META_FILTER", old)
+
+	a := &API{} // 零值：LanceDB 为 nil —— 422 先于写路径 ⇒ 不 panic
+	body := `{"agent_id":"hermes-a06","namespace":"hermes-main","content":"实测 淇县 ANN 731→15（都进池50）","category":"system_fact"}`
+	req := httptest.NewRequest(http.MethodPost, "/api/v1/commit", strings.NewReader(body))
+	w := httptest.NewRecorder()
+	a.Commit(w, req)
+	if w.Code != 422 {
+		t.Fatalf("commit(meta, 门禁开) 期望 422，实际 %d body=%s", w.Code, w.Body.String())
+	}
+	if !strings.Contains(w.Body.String(), "meta-note") {
+		t.Errorf("拒绝体应说明 meta-note 原因: %s", w.Body.String())
+	}
+
+	// 门禁关 ⇒ 同一内容 ShouldSkip 必须 false（AC6 零行为变化）
+	os.Setenv("ZHIYI_META_FILTER", "")
+	if skip, _ := metanote.ShouldSkip("实测 淇县 ANN 731→15（都进池50）"); skip {
+		t.Fatal("门禁关闭时 ShouldSkip 必须为 false（零行为变化）")
+	}
+}
+
+func TestWriteMemoryGate_MetaRejectedBeforeStorage(t *testing.T) {
+	old := os.Getenv("ZHIYI_META_FILTER")
+	os.Setenv("ZHIYI_META_FILTER", "1")
+	defer os.Setenv("ZHIYI_META_FILTER", old)
+
+	a := &API{}
+	body := `{"agent_id":"hermes-a06","content":"织忆 golden 14/14 不退化，recall@5 = 0.4286"}`
+	req := httptest.NewRequest(http.MethodPost, "/api/v1/memory/write", strings.NewReader(body))
+	w := httptest.NewRecorder()
+	a.WriteMemory(w, req)
+	if w.Code != 422 {
+		t.Fatalf("write_memory(meta, 门禁开) 期望 422，实际 %d body=%s", w.Code, w.Body.String())
+	}
+}

go/internal/api/routes/obsidian.go

@@ -2,6 +2,8 @@
 package routes
 
 import (
+	"crypto/sha256"
+	"encoding/hex"
 	"encoding/json"
 	"fmt"
 	"log"
@@ -11,7 +13,9 @@ import (
 	"strings"
 	"sync"
 	"time"
+	"unicode"
 
+	"github.com/xiaoxue/memoryweave/internal/redact"
 	"github.com/xiaoxue/memoryweave/internal/storage"
 )
 
@@ -52,14 +56,29 @@ func (s *ObsidianSyncer) PushToObsidian(memories []map[string]interface{}, folde
 	}
 
 	count := 0
+	used := make(map[string]int, len(memories))
 	for _, mem := range memories {
 		content := strVal(mem["content"])
+		// 2026-09-10 P0：镜像正文同样脱敏——旧事故里正文与文件名都把
+		// tskey-auth-… 原样落盘（小唯/07-Wiki/织忆/未找到命令.md）。
+		content = redact.RedactSecrets(content)
 		if len(content) < 5 {
 			continue
 		}
 
-		filename := sanitizeFilename(content[:minz(len(content), 30)]) + ".md"
-		filepath := filepath.Join(targetDir, filename)
+		// 2026-09-10 修复：旧实现 content[:30] 按字节切 + sanitizeFilename 把非 ASCII
+		// 全替换成 '_'，中文标题直接变 ____123.md（07-Wiki/织忆 1628 个乱码文件名成因），
+		// 且不同记忆撞名会静默互相覆盖。改为 rune 级截断 + 保留 Unicode + 撞名加短哈希。
+		base := sanitizeFilename(truncateRunes(content, 30))
+		if base == "" {
+			base = "untitled"
+		}
+		if used[base] > 0 {
+			base = base + "-" + shortHash(content)
+		}
+		used[base]++
+		filename := base + ".md"
+		target := filepath.Join(targetDir, filename)
 
 		mdContent := fmt.Sprintf(`---
 id: %s
@@ -80,7 +99,7 @@ sync_time: %s
 			truncate(content, 60),
 			content)
 
-		if err := os.WriteFile(filepath, []byte(mdContent), 0644); err != nil {
+		if err := os.WriteFile(target, []byte(mdContent), 0644); err != nil {
 			log.Printf("[obsidian] 写入失败 %s: %v", filename, err)
 			continue
 		}
@@ -206,29 +225,61 @@ func (s *ObsidianSyncer) StatusHandler(w http.ResponseWriter, r *http.Request) {
 
 // ─── 辅助 ─────────────────────────────────────────────────
 
+// sanitizeFilename 把标题转成文件系统安全的文件名片段，并保留 Unicode（中文）可读性。
+//
+// 2026-09-10 修复：旧实现把所有非 ASCII rune 逐个替换为 '_'，于是中文标题全变成
+// `____123.md`，在 Obsidian 里无法按名浏览（07-Wiki/织忆 1628 个 legacy 文件的成因）。
+// 现在只替换文件系统非法字符（/ \ : * ? " < > | 控制符等），连续的非法字符折叠成一个 '_'。
+// 返回空串表示内容里没有任何可用的安全字符，由调用方兜底。
 func sanitizeFilename(s string) string {
-	s = strings.Map(func(r rune) rune {
-		if (r >= 'a' && r <= 'z') || (r >= 'A' && r <= 'Z') ||
-			(r >= '0' && r <= '9') || r == '_' || r == '-' || r == ' ' {
-			return r
+	// 2026-09-10 P0：入参先脱敏，防任何路径把凭证带进文件名
+	s = redact.RedactSecrets(s)
+	var b strings.Builder
+	prevIllegal := false
+	for _, r := range s {
+		if unicode.IsLetter(r) || unicode.IsDigit(r) ||
+			r == '-' || r == '_' || r == ' ' || r == '.' {
+			prevIllegal = false
+			b.WriteRune(r)
+			continue
 		}
-		return '_'
-	}, s)
-	return strings.TrimSpace(s)
+		// 文件系统非法字符（/ \ : * ? " < > | 控制符等）+ 其它符号 → 折叠成一个 '_'
+		if !prevIllegal {
+			b.WriteRune('_')
+			prevIllegal = true
+		}
+	}
+	return strings.Trim(b.String(), " ._-")
 }
 
+// shortHash 取内容摘要前 8 位十六进制，仅用于同名文件消歧，不做安全用途。
+func shortHash(s string) string {
+	sum := sha256.Sum256([]byte(s))
+	return hex.EncodeToString(sum[:4])
+}
+
+// truncateRunes 按 rune（而非字节）截断，避免把多字节字符切掉一半产生乱码。
+func truncateRunes(s string, maxRunes int) string {
+	if maxRunes <= 0 {
+		return ""
+	}
+	n := 0
+	for i := range s {
+		if n == maxRunes {
+			return s[:i]
+		}
+		n++
+	}
+	return s
+}
+
+// truncate 展示用截断，按 rune 计数（2026-09-10 从字节切改为 rune 切，避免乱码）。
 func truncate(s string, maxLen int) string {
-	if len(s) <= maxLen {
+	t := truncateRunes(s, maxLen)
+	if len(t) == len(s) {
 		return s
 	}
-	return s[:maxLen] + "..."
-}
-
-func minz(a, b int) int {
-	if a < b {
-		return a
-	}
-	return b
+	return t + "..."
 }
 
 func floatVal(v interface{}) float64 {

go/internal/api/routes/obsidian_carrier.go

@@ -103,8 +103,10 @@ func formatEntry(content, category string) string {
 
 func strValSafe(v interface{}) string {
 	switch s := v.(type) {
-	case string: return s
-	case fmt.Stringer: return s.String()
+	case string:
+		return s
+	case fmt.Stringer:
+		return s.String()
 	}
 	return ""
 }

go/internal/api/routes/obsidian_test.go

@@ -0,0 +1,182 @@
+// 织忆 MemoryWeave — Obsidian 同步文件名/slug 单元测试
+// 2026-09-10 新增（kanban t_5f9c56ef）：锁死「中文标题不许变下划线」这个回归。
+package routes
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"unicode/utf8"
+)
+
+func TestSanitizeFilename_KeepsChinese(t *testing.T) {
+	cases := []struct {
+		in   string
+		want string
+	}{
+		{"修复了respond重复声明问题", "修复了respond重复声明问题"},
+		{"110GB 磁盘占用排查", "110GB 磁盘占用排查"},
+		{"E5.3 Web UI 单文件React+静态文件中间件", "E5.3 Web UI 单文件React_静态文件中间件"},
+		{"a/b\\c:d*e?f\"g<h>i|j", "a_b_c_d_e_f_g_h_i_j"},
+		{"  前后空格  ", "前后空格"},
+		{"...隐藏点开头...", "隐藏点开头"},
+		{"///", ""},
+		{"", ""},
+		{"测试::连续非法字符::折叠", "测试_连续非法字符_折叠"},
+	}
+	for _, c := range cases {
+		if got := sanitizeFilename(c.in); got != c.want {
+			t.Errorf("sanitizeFilename(%q) = %q, want %q", c.in, got, c.want)
+		}
+	}
+}
+
+// 回归锁：旧实现（2026-09-10 前）会把中文全部换成 '_'，这条用例就是当时的现场。
+func TestSanitizeFilename_RegressionNoUnderscoreSoup(t *testing.T) {
+	got := sanitizeFilename("____112GB_____")
+	if strings.Contains(got, "____") {
+		t.Fatalf("中文标题不该出现连续下划线乱码: %q", got)
+	}
+	old := sanitizeFilename("修复了respond重复声明问题")
+	if strings.Contains(old, "_") {
+		t.Fatalf("中文/字母不该被替换成下划线: %q", old)
+	}
+}
+
+func TestTruncateRunes_NoBrokenUTF8(t *testing.T) {
+	s := "修复了respond重复声明问题，这是很长的中文内容用来测试截断"
+	got := truncateRunes(s, 7)
+	if got != "修复了resp" {
+		t.Fatalf("truncateRunes = %q", got)
+	}
+	if !utf8.ValidString(got) {
+		t.Fatal("截断结果不是合法 UTF-8（多字节字符被切半）")
+	}
+	if truncateRunes(s, 0) != "" || truncateRunes(s, -1) != "" {
+		t.Fatal("maxRunes<=0 应返回空串")
+	}
+	if truncateRunes("短", 10) != "短" {
+		t.Fatal("短串应原样返回")
+	}
+	// 关键回归：旧实现 content[:30] 是字节切，中文会切出半个字符（utf8.RuneError）
+	if got := truncateRunes("中文标题测试内容很长很长很长很长很长很长", 8); !utf8.ValidString(got) {
+		t.Fatalf("rune 截断后仍不是合法 UTF-8: %q", got)
+	}
+}
+
+func TestTruncate_RuneCounted(t *testing.T) {
+	s := strings.Repeat("记", 40) // 120 字节；旧实现 s[:10] 只能得到 3 个半截汉字
+	got := truncate(s, 10)
+	want := strings.Repeat("记", 10) + "..."
+	if got != want {
+		t.Fatalf("truncate = %q, want %q", got, want)
+	}
+	if !utf8.ValidString(got) {
+		t.Fatal("truncate 结果不是合法 UTF-8")
+	}
+	if truncate("短内容", 60) != "短内容" {
+		t.Fatal("未超长应原样返回")
+	}
+}
+
+func TestShortHash_Deterministic(t *testing.T) {
+	a, b := shortHash("同一段内容"), shortHash("同一段内容")
+	if a != b {
+		t.Fatal("shortHash 必须稳定")
+	}
+	if len(a) != 8 {
+		t.Fatalf("shortHash 长度应为 8，得到 %d", len(a))
+	}
+	if shortHash("内容甲") == shortHash("内容乙") {
+		t.Fatal("不同内容不该同哈希")
+	}
+}
+
+// 端到端：真的落盘，检查文件名可读、无乱码、同名不互相覆盖。
+func TestPushToObsidian_ReadableFilenames(t *testing.T) {
+	dir := t.TempDir()
+	s := NewObsidianSyncer(dir, nil)
+
+	dupPrefix := "同一段开头的内容 abcdefghijklmnopqrstuvwxyz"
+	mems := []map[string]interface{}{
+		{"id": "mem_1", "category": "distilled", "quality_score": 0.9,
+			"content": "修复了 respond 重复声明问题"},
+		{"id": "mem_2", "category": "distilled", "quality_score": 0.8,
+			"content": "LanceDB 已设天花板，不会再膨胀到 35G"},
+		// 前 30 个 rune 完全相同 → 必须加哈希后缀而不是覆盖
+		{"id": "mem_3", "category": "distilled", "content": dupPrefix + " 变体一"},
+		{"id": "mem_4", "category": "distilled", "content": dupPrefix + " 变体二"},
+	}
+
+	if err := s.PushToObsidian(mems, "mirror"); err != nil {
+		t.Fatalf("PushToObsidian 失败: %v", err)
+	}
+
+	entries, err := os.ReadDir(filepath.Join(dir, "mirror"))
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(entries) != len(mems) {
+		t.Fatalf("应写入 %d 个文件（含撞名消歧），实际 %d 个：%v", len(mems), len(entries), names(entries))
+	}
+	t.Logf("实际落盘文件名: %v", names(entries))
+
+	for _, e := range entries {
+		n := e.Name()
+		if strings.Contains(n, "____") {
+			t.Errorf("文件名出现下划线乱码: %q", n)
+		}
+		if !strings.ContainsAny(n, "修复了重复声明问题磁盘排查天花板膨胀同一段") {
+			t.Errorf("文件名丢失中文标题: %q", n)
+		}
+		if !utf8.ValidString(n) {
+			t.Errorf("文件名不是合法 UTF-8: %q", n)
+		}
+		data, err := os.ReadFile(filepath.Join(dir, "mirror", n))
+		if err != nil {
+			t.Fatal(err)
+		}
+		if !utf8.Valid(data) {
+			t.Errorf("文件内容不是合法 UTF-8: %q", n)
+		}
+		if !strings.Contains(string(data), "由织忆 MemoryWeave 同步") {
+			t.Errorf("缺少同步 footer: %q", n)
+		}
+	}
+
+	// 撞名消歧：两条同前缀记忆都要落盘（4 个文件已证明没互相覆盖）
+	var suffixed int
+	for _, e := range entries {
+		if strings.Contains(e.Name(), dupPrefix[:12]) && strings.Contains(e.Name(), "-") {
+			suffixed++
+		}
+	}
+	if suffixed != 1 {
+		t.Errorf("应恰好 1 个文件带哈希后缀消歧，实际 %d：%v", suffixed, names(entries))
+	}
+}
+
+func names(entries []os.DirEntry) []string {
+	out := make([]string, 0, len(entries))
+	for _, e := range entries {
+		out = append(out, e.Name())
+	}
+	return out
+}
+
+// ─── 2026-09-10 P0：文件名不得带凭证 ───────────────────────
+
+func TestSanitizeFilename_RedactsSecret(t *testing.T) {
+	const fakeKey = "tskey-auth-kTESTONLY0000000000000000000000000000000000000"
+	got := sanitizeFilename("authkey " + fakeKey)
+	if got == "" {
+		t.Fatal("脱敏后不应为空（占位符仍可读）")
+	}
+	if strings.Contains(got, "tskey-auth-kTESTONLY") {
+		t.Errorf("文件名仍含凭证: %q", got)
+	}
+	if strings.Contains(got, "/") || strings.Contains(got, string(os.PathSeparator)) {
+		t.Errorf("文件名仍含路径分隔符: %q", got)
+	}
+}

go/internal/api/routes/patterns.go

@@ -0,0 +1,211 @@
+// 织忆 MemoryWeave — L2 模式 API（t_d32e5f8a）
+//
+// 设计 §4.1 端点：
+//   POST /api/v1/patterns/detect  手动触发模式检测（扫 L1 蒸馏记忆 → 工具组合频率 + 路径聚类
+//                                  → 写 graph_nodes type='pattern'）。短命单发：modernc sqlite
+//                                  开/关不占长连接、不碰 state.db（规避当年 daemon main_loop
+//                                  退役根因）—— 可挂 cron 单发。
+//   GET  /api/v1/patterns         列出检测到的模式（按 source_tasks 长度降序）。
+package routes
+
+import (
+	"database/sql"
+	"encoding/json"
+	"net/http"
+	"net/url"
+	"os"
+	"strconv"
+	"strings"
+
+	"github.com/xiaoxue/memoryweave/internal/models"
+	"github.com/xiaoxue/memoryweave/internal/pattern"
+	"github.com/xiaoxue/memoryweave/internal/skillgen"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// PatternAPI L2 模式端点 + L3 skill 草稿端点。
+type PatternAPI struct {
+	Store    *pattern.GraphPatternStore
+	Provider *pattern.PatternProvider
+	LDB      func() storage.LanceDB // 扫 L1 记忆（distilled）
+	Cfg      pattern.Config
+
+	// L3 skill 草稿（t_fcadee04）：草稿生成器 + skills 表存储 + 落盘根目录。
+	// 由 server.go 装配；不装配时 Draft/Promote 返回 503/404（显式 API，不挂 cron）。
+	SkillStore    *skillgen.Store
+	Generator     skillgen.Generator
+	AttachmentsDir string
+}
+
+// NewPatternAPI 构造 L2 模式端点。
+func NewPatternAPI(store *pattern.GraphPatternStore, provider *pattern.PatternProvider, ldb func() storage.LanceDB) *PatternAPI {
+	cfg := pattern.DefaultConfig()
+	// cron 单发跑大批量时可放宽扫描上限
+	if v := os.Getenv("ZHIYI_PATTERN_MAX_RECORDS"); v != "" {
+		if n, err := strconv.Atoi(v); err == nil && n > 0 {
+			cfg.MaxRecords = n
+		}
+	}
+	return &PatternAPI{Store: store, Provider: provider, LDB: ldb, Cfg: cfg}
+}
+
+// GET /api/v1/patterns — 列出所有检测到的模式（AC1：空库返回 count=0 不报错）。
+func (pa *PatternAPI) List(w http.ResponseWriter, r *http.Request) {
+	limit := pa.Cfg.TopKPatternList
+	if v := r.URL.Query().Get("limit"); v != "" {
+		if n, err := strconv.Atoi(v); err == nil && n > 0 {
+			limit = n
+		}
+	}
+	pats, err := pa.Store.ListPatterns(limit)
+	if err != nil {
+		respondError(w, 500, "list patterns: "+err.Error())
+		return
+	}
+	respond(w, 200, map[string]interface{}{
+		"patterns": pats,
+		"count":    len(pats),
+	})
+}
+
+// POST /api/v1/patterns/detect — 手动触发模式检测。
+// body 可选 {"limit": N, "namespace": "hermes-main"}。
+// 流程：LanceDB Scan(limit, ns) → pattern.DetectFromRecords → UpsertPatterns(graph_nodes)。
+// 空库/无命中 → created/updated=0，不报错（AC1）。
+func (pa *PatternAPI) Detect(w http.ResponseWriter, r *http.Request) {
+	var req struct {
+		Limit     int    `json:"limit"`
+		Namespace string `json:"namespace"`
+	}
+	if r.Body != nil {
+		_ = json.NewDecoder(r.Body).Decode(&req) // 空 body 也合法（用默认）
+	}
+	if req.Limit <= 0 {
+		req.Limit = pa.Cfg.MaxRecords
+	}
+	ldb := pa.LDB()
+	if ldb == nil {
+		respondError(w, 503, "lancedb unavailable")
+		return
+	}
+	records, err := ldb.Scan(req.Limit, req.Namespace)
+	if err != nil {
+		respondError(w, 500, "scan memories: "+err.Error())
+		return
+	}
+	// 只拿蒸馏产物（L1 蒸馏记忆）做检测——原始对话/episode 不是「已验证」的解决路径
+	// （与召回管线 2.5 步同源判据：raw/episodes 类不参与模式沉淀）。
+	distilled := make([]models.MemoryRecord, 0, len(records))
+	for _, rec := range records {
+		if rec.Category == "distilled" || rec.Category == "" {
+			distilled = append(distilled, rec)
+		}
+	}
+	pats := pattern.DetectFromRecords(distilled, pa.Cfg)
+	created, updated, uerr := pa.Store.UpsertPatterns(pats)
+	if uerr != nil {
+		respondError(w, 500, "upsert patterns: "+uerr.Error())
+		return
+	}
+	respond(w, 200, map[string]interface{}{
+		"scanned":  len(records),
+		"created":  created,
+		"updated":  updated,
+		"detected": len(pats),
+		"patterns": pats,
+	})
+}
+
+// ── L3 skill 草稿（t_fcadee04，设计 §4.2 SkillGenerator）────────────────────
+
+// patternSubPath 解析 /api/v1/patterns/{id}/{action} 子路径。
+// 返回 (url-unescape 后的 id, action)。格式不符 → ("", action=="")。
+func patternSubPath(path string) (id, action string) {
+	const prefix = "/api/v1/patterns/"
+	if !strings.HasPrefix(path, prefix) {
+		return "", ""
+	}
+	rest := strings.TrimPrefix(path, prefix) // <id>/draft 或 <id>/promote
+	parts := strings.Split(rest, "/")
+	if len(parts) != 2 || parts[0] == "" {
+		return "", ""
+	}
+	dec, err := url.PathUnescape(parts[0])
+	if err != nil {
+		return "", ""
+	}
+	return dec, parts[1]
+}
+
+// Sub 分发 /api/v1/patterns/{id}/draft 与 /api/v1/patterns/{id}/promote。
+// 装配在 server.go 的 /api/v1/patterns/ 前缀路由；List(/patterns) 与 Detect(/patterns/detect)
+// 因更长的精确路由优先，不受影响。
+func (pa *PatternAPI) Sub(w http.ResponseWriter, r *http.Request) {
+	id, action := patternSubPath(r.URL.Path)
+	if id == "" {
+		respondError(w, 404, "pattern id required")
+		return
+	}
+	switch action {
+	case "draft":
+		pa.Draft(w, r, id)
+	case "promote":
+		pa.Promote(w, r, id)
+	default:
+		respondError(w, 404, "unknown action: "+action)
+	}
+}
+
+// POST /api/v1/patterns/{id}/draft — 生成 L3 skill 草稿（conf≥0.85 才生成）。
+// 流程：读 pattern → 校验 conf → DraftSkill（写 skills 表 draft + 落盘 attachments）。
+// 草稿绝不进 skills/ 技能库（红线）。
+func (pa *PatternAPI) Draft(w http.ResponseWriter, r *http.Request, id string) {
+	if pa.SkillStore == nil || pa.Generator == nil {
+		respondError(w, 503, "skillgen 未装配（SkillStore/Generator 为空）")
+		return
+	}
+	pat, err := pa.Store.GetPattern(id)
+	if err != nil {
+		if err == sql.ErrNoRows {
+			respondError(w, 404, "pattern not found: "+id)
+		} else {
+			respondError(w, 500, "get pattern: "+err.Error())
+		}
+		return
+	}
+	attRoot := pa.AttachmentsDir
+	if attRoot == "" {
+		attRoot = skillgen.DefaultAttachmentsRoot()
+	}
+	sk, path, derr := skillgen.DraftSkill(pa.SkillStore, pa.Generator, pat, attRoot)
+	if derr != nil {
+		respondError(w, 400, "draft skill: "+derr.Error())
+		return
+	}
+	respond(w, 200, map[string]interface{}{
+		"skill": sk,
+		"file":  path,
+	})
+}
+
+// POST /api/v1/patterns/{id}/promote — 人工确认后 draft→active。
+func (pa *PatternAPI) Promote(w http.ResponseWriter, r *http.Request, id string) {
+	if pa.SkillStore == nil {
+		respondError(w, 503, "skillgen 未装配（SkillStore 为空）")
+		return
+	}
+	if err := pa.SkillStore.Promote(id); err != nil {
+		if err == sql.ErrNoRows {
+			respondError(w, 404, "no skill draft for pattern: "+id)
+		} else {
+			respondError(w, 500, "promote skill: "+err.Error())
+		}
+		return
+	}
+	sk, gerr := pa.SkillStore.Get(id)
+	if gerr != nil {
+		respondError(w, 500, "read skill after promote: "+gerr.Error())
+		return
+	}
+	respond(w, 200, map[string]interface{}{"skill": sk})
+}

go/internal/api/routes/patterns_skillgen_test.go

@@ -0,0 +1,145 @@
+// 织忆 MemoryWeave — L3 skill 草稿 API 路由测试（t_fcadee04）
+//
+// 覆盖端点：POST /api/v1/patterns/{id}/draft + POST /api/v1/patterns/{id}/promote
+// 共享同一临时 graph.db（pattern 在 graph_nodes、skill 在 skills 表同库），零网络/零 CGO。
+package routes
+
+import (
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"github.com/xiaoxue/memoryweave/internal/pattern"
+	"github.com/xiaoxue/memoryweave/internal/skillgen"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// newSkillgenTestAPI 构造共享同一 temp graph.db 的 PatternAPI（含 L3 装配）。
+func newSkillgenTestAPI(t *testing.T) (*PatternAPI, string) {
+	t.Helper()
+	dbDir := t.TempDir()
+	graphPath := filepath.Join(dbDir, "graph.db")
+	pstore := pattern.NewGraphPatternStore(graphPath)
+	t.Cleanup(pstore.Close)
+	sstore := skillgen.NewStore(graphPath)
+	t.Cleanup(sstore.Close)
+	attRoot := filepath.Join(t.TempDir(), "attachments")
+
+	api := NewPatternAPI(pstore, nil, func() storage.LanceDB { return nil })
+	api.SkillStore = sstore
+	api.Generator = &skillgen.FakeGenerator{Out: "## 适用场景\n重复处理"}
+	api.AttachmentsDir = attRoot
+
+	// 种一个 conf=0.9 的 pattern 进 graph_nodes
+	hi := pattern.Pattern{
+		ID: "pat_tool_combo:curl+git+jq", Type: "tool_combo",
+		Name: "tool_combo:curl+git+jq", Data: "curl+git+jq",
+		Confidence: 0.9, Tools: []string{"curl", "git", "jq"},
+		SourceTasks: []string{"m_a", "m_b", "m_c"},
+	}
+	if _, _, err := pstore.UpsertPatterns([]pattern.Pattern{hi}); err != nil {
+		t.Fatalf("seed pattern: %v", err)
+	}
+	return api, attRoot
+}
+
+// TestSkillgen_DraftRoute —— AC1 API 层：draft 端点 → 200 + skills 表 1 行 draft + 附件文件。
+func TestSkillgen_DraftRoute(t *testing.T) {
+	api, attRoot := newSkillgenTestAPI(t)
+	id := "pat_tool_combo:curl+git+jq"
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest("POST", "/api/v1/patterns/"+id+"/draft", nil)
+	api.Sub(rec, req)
+	if rec.Code != http.StatusOK {
+		t.Fatalf("draft 期望 200，实际 %d body=%s", rec.Code, rec.Body.String())
+	}
+	// skills 表 1 行 draft
+	if n, _ := api.SkillStore.Count("draft"); n != 1 {
+		t.Fatalf("AC1 判据：draft 行数应 1，实际 %d", n)
+	}
+	sk, err := api.SkillStore.Get(id)
+	if err != nil || sk.Status != "draft" {
+		t.Fatalf("skill 应存在且 draft，实际 %v %+v", err, sk)
+	}
+	// 附件文件存在 + 含 [AUTO-DRAFT]
+	path := filepath.Join(attRoot, skillgen.DraftSubDir("tool_combo:curl+git+jq"), "skill-draft.md")
+	data, rerr := os.ReadFile(path)
+	if rerr != nil {
+		t.Fatalf("附件文件不存在 %s: %v", path, rerr)
+	}
+	if !strings.Contains(string(data), skillgen.AUTO_DRAFT_MARKER) {
+		t.Fatalf("附件应含 [AUTO-DRAFT] 标记")
+	}
+	t.Logf("route draft 落盘: %s", path)
+}
+
+// TestSkillgen_PromoteRoute —— AC3 API 层：promote 端点 → 200 + status active。
+func TestSkillgen_PromoteRoute(t *testing.T) {
+	api, _ := newSkillgenTestAPI(t)
+	id := "pat_tool_combo:curl+git+jq"
+	// 先 draft
+	dr := httptest.NewRecorder()
+	api.Sub(dr, httptest.NewRequest("POST", "/api/v1/patterns/"+id+"/draft", nil))
+	if dr.Code != http.StatusOK {
+		t.Fatalf("前置 draft 应 200，实际 %d", dr.Code)
+	}
+	// promote
+	pr := httptest.NewRecorder()
+	api.Sub(pr, httptest.NewRequest("POST", "/api/v1/patterns/"+id+"/promote", nil))
+	if pr.Code != http.StatusOK {
+		t.Fatalf("promote 期望 200，实际 %d body=%s", pr.Code, pr.Body.String())
+	}
+	var resp struct {
+		Skill struct {
+			Status string `json:"status"`
+		} `json:"skill"`
+	}
+	if err := json.Unmarshal(pr.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("promote 解析: %v", err)
+	}
+	if resp.Skill.Status != "active" {
+		t.Fatalf("AC3 判据：promote 后 status 应 active，实际 %s", resp.Skill.Status)
+	}
+	// 库复核（AC3 判据：sqlite 查）
+	if n, _ := api.SkillStore.Count("active"); n != 1 {
+		t.Fatalf("AC3 判据：active 行数应 1，实际 %d", n)
+	}
+}
+
+// TestSkillgen_DraftRoute_NotFound —— 不存在的 pattern → 404。
+func TestSkillgen_DraftRoute_NotFound(t *testing.T) {
+	api, _ := newSkillgenTestAPI(t)
+	rec := httptest.NewRecorder()
+	api.Sub(rec, httptest.NewRequest("POST", "/api/v1/patterns/nope/draft", nil))
+	if rec.Code != http.StatusNotFound {
+		t.Fatalf("不存在 pattern 的 draft 应 404，实际 %d", rec.Code)
+	}
+}
+
+// TestSkillgen_PromoteRoute_NoDraft —— 无草稿即 promote → 404。
+func TestSkillgen_PromoteRoute_NoDraft(t *testing.T) {
+	api, _ := newSkillgenTestAPI(t)
+	rec := httptest.NewRecorder()
+	api.Sub(rec, httptest.NewRequest("POST", "/api/v1/patterns/pat_tool_combo:curl+git+jq/promote", nil))
+	if rec.Code != http.StatusNotFound {
+		t.Fatalf("无草稿 promote 应 404，实际 %d body=%s", rec.Code, rec.Body.String())
+	}
+}
+
+// TestSkillgen_SubPath —— 子路径解析单元（含 URL 编码 id）。
+func TestSkillgen_SubPath(t *testing.T) {
+	id, action := patternSubPath("/api/v1/patterns/pat_tool_combo:curl+git+jq/promote")
+	if id != "pat_tool_combo:curl+git+jq" || action != "promote" {
+		t.Fatalf("解析失败 id=%q action=%q", id, action)
+	}
+	if _, a := patternSubPath("/api/v1/patterns"); a != "" {
+		t.Fatalf("无子路径应 action 空")
+	}
+	if _, a := patternSubPath("/api/v1/patterns/detect"); a != "" {
+		t.Fatalf("/patterns/detect 是精确路由，不应走 Sub")
+	}
+}

go/internal/api/routes/patterns_test.go

@@ -0,0 +1,132 @@
+package routes
+
+import (
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"testing"
+
+	"github.com/xiaoxue/memoryweave/internal/models"
+	"github.com/xiaoxue/memoryweave/internal/pattern"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// 织忆 — L2 模式 API 路由测试（t_d32e5f8a）
+// AC1: 空库 detect/list → 不报错（count=0）
+// AC2: 3 条同工具组合记忆 → detect 后 list 能查到 tool_combo pattern
+
+// newPatternTestAPI 构造一个用临时 graph.db + 假 LanceDB 后端的 PatternAPI（无网络、无 CGO）。
+// 假后端只实现 PatternAPI 实际调用的 Scan（其余方法嵌 nil 接口，测试路径不会触达）。
+func newPatternTestAPI(t *testing.T, mems []models.MemoryRecord) *PatternAPI {
+	t.Helper()
+	store := pattern.NewGraphPatternStore(t.TempDir() + "/graph.db")
+	t.Cleanup(store.Close)
+	mlc := &patternScanLance{rows: mems}
+	pemb := pattern.NewEmbedderProvider(&fakePatternEmbedder{})
+	return NewPatternAPI(store, pattern.NewPatternProvider(store, pemb), func() storage.LanceDB { return mlc })
+}
+
+// patternScanLance 覆写 Scan（PatternAPI.Detect 唯一用到的 LanceDB 方法）；
+// 其余方法嵌 nil 接口 —— 本测试路径不会调用，触达即 panic（暴露误用）。
+type patternScanLance struct {
+	storage.LanceDB
+	rows []models.MemoryRecord
+}
+
+func (f *patternScanLance) Scan(limit int, namespaceFilter string) ([]models.MemoryRecord, error) {
+	if limit > 0 && len(f.rows) > limit {
+		return f.rows[:limit], nil
+	}
+	return f.rows, nil
+}
+
+// fakePatternEmbedder 确定性假向量（pattern.Provider 匹配测试用；注入测试走 storage 侧不走这里）。
+type fakePatternEmbedder struct{}
+
+func (fakePatternEmbedder) EncodeSingle(text string) ([]float32, error) {
+	v := make([]float32, 8)
+	for i := 0; i < 8; i++ {
+		v[i] = 0.1
+	}
+	return v, nil
+}
+
+// TestPatterns_ListEmpty —— AC1：空库 list → 200 + count=0（不报错）。
+func TestPatterns_ListEmpty(t *testing.T) {
+	api := newPatternTestAPI(t, nil)
+	rec := httptest.NewRecorder()
+	api.List(rec, httptest.NewRequest("GET", "/api/v1/patterns", nil))
+	if rec.Code != http.StatusOK {
+		t.Fatalf("空库 list 期望 200，实际 %d body=%s", rec.Code, rec.Body.String())
+	}
+	var resp map[string]interface{}
+	_ = json.Unmarshal(rec.Body.Bytes(), &resp)
+	if int(resp["count"].(float64)) != 0 {
+		t.Fatalf("空库 count 期望 0，实际 %v", resp["count"])
+	}
+}
+
+// TestPatterns_DetectEmpty —— AC1：空库 detect → 200 + scanned=0/created=0（不报错）。
+func TestPatterns_DetectEmpty(t *testing.T) {
+	api := newPatternTestAPI(t, nil)
+	rec := httptest.NewRecorder()
+	api.Detect(rec, httptest.NewRequest("POST", "/api/v1/patterns/detect", nil))
+	if rec.Code != http.StatusOK {
+		t.Fatalf("空库 detect 期望 200，实际 %d body=%s", rec.Code, rec.Body.String())
+	}
+	var resp map[string]interface{}
+	_ = json.Unmarshal(rec.Body.Bytes(), &resp)
+	if int(resp["scanned"].(float64)) != 0 || int(resp["created"].(float64)) != 0 {
+		t.Fatalf("空库 detect 期望 0/0，实际 %v/%v", resp["scanned"], resp["created"])
+	}
+}
+
+// TestPatterns_DetectToolCombo —— AC2：3 条同工具组合蒸馏记忆 → detect 产出 1 个 tool_combo，
+// 且 list 能查到（sqlite graph_nodes 判据）。
+func TestPatterns_DetectToolCombo(t *testing.T) {
+	mems := make([]models.MemoryRecord, 0, 3)
+	for i, id := range []string{"pc_a", "pc_b", "pc_c"} {
+		mems = append(mems, models.MemoryRecord{
+			ID:       id,
+			Content:  "用 jq git curl 处理第" + string(rune('一'+i)) + "项",
+			Category: "distilled",
+		})
+	}
+	api := newPatternTestAPI(t, mems)
+
+	rec := httptest.NewRecorder()
+	api.Detect(rec, httptest.NewRequest("POST", "/api/v1/patterns/detect", nil))
+	if rec.Code != http.StatusOK {
+		t.Fatalf("detect 期望 200，实际 %d body=%s", rec.Code, rec.Body.String())
+	}
+	var dresp map[string]interface{}
+	_ = json.Unmarshal(rec.Body.Bytes(), &dresp)
+	if int(dresp["detected"].(float64)) < 1 {
+		t.Fatalf("3 条同组合应检出 ≥1 个 pattern，实际 detected=%v", dresp["detected"])
+	}
+	// 用 list 再核一遍 sqlite 判据：确有 tool_combo 行
+	lrec := httptest.NewRecorder()
+	api.List(lrec, httptest.NewRequest("GET", "/api/v1/patterns", nil))
+	var lresp struct {
+		Count int `json:"count"`
+		Patterns []struct {
+			Type string `json:"type"`
+			Name string `json:"name"`
+		} `json:"patterns"`
+	}
+	if err := json.Unmarshal(lrec.Body.Bytes(), &lresp); err != nil {
+		t.Fatalf("list 解析: %v (body=%s)", err, lrec.Body.String())
+	}
+	if lresp.Count < 1 {
+		t.Fatalf("AC2 判据：detect 后 list count 应 ≥1，实际 %d (detected=%v)", lresp.Count, dresp["detected"])
+	}
+	var foundToolCombo bool
+	for _, p := range lresp.Patterns {
+		if p.Type == "tool_combo" {
+			foundToolCombo = true
+		}
+	}
+	if !foundToolCombo {
+		t.Fatalf("AC2 判据：list 应含 tool_combo pattern，实际 %d 条均非 tool_combo", lresp.Count)
+	}
+}

go/internal/api/routes/peer.go

@@ -0,0 +1,335 @@
+// 织忆 MemoryWeave — P3 Peer Representation API（t_f8fc97e6，设计 §6.1）
+//
+// 端点：
+//
+//	POST /api/v1/peer/{id}/refresh   手动刷新某实体的四层推理画像（显式触发）
+//	GET  /api/v1/peer/{id}/explain   解释为何这样理解该实体（可审计：画像 + 来源记忆）
+//	POST /api/v1/recall/peer?peer_id=…  召回某 peer 的推理画像（env 门控 ZHIYI_PEER_RECALL）
+//
+// 设计约束（与 L2 pattern / L3 skillgen 同源）：
+//
+//	· refresh/explain 是显式 API，不挂 cron ⇒ 默认零行为变化。
+//	· 召回注入走独立端点 + env 门控 ZHIYI_PEER_RECALL，默认 0 ⇒ 主召回管线逐条不变（AC2）。
+//	· peers 表在 graph.db（短命 modernc），不碰 state.db。
+package routes
+
+import (
+	"database/sql"
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"net/url"
+	"os"
+	"strings"
+
+	"github.com/xiaoxue/memoryweave/internal/governance"
+	"github.com/xiaoxue/memoryweave/internal/models"
+	"github.com/xiaoxue/memoryweave/internal/peer"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// PeerAPI P3 peer 表示端点。
+type PeerAPI struct {
+	Store      *peer.Store
+	Reasoner   peer.Reasoner
+	LDB        func() storage.LanceDB // 扫 L1 记忆（explicit 层来源）
+	GraphStore governance.GraphStore  // 图谱上下文（推理输入）
+}
+
+// NewPeerAPI 构造。
+func NewPeerAPI(store *peer.Store, reasoner peer.Reasoner, ldb func() storage.LanceDB, g governance.GraphStore) *PeerAPI {
+	return &PeerAPI{Store: store, Reasoner: reasoner, LDB: ldb, GraphStore: g}
+}
+
+// peerSubPath 解析 /api/v1/peer/{id}/{action} 子路径。
+func peerSubPath(path string) (id, action string) {
+	const prefix = "/api/v1/peer/"
+	if !strings.HasPrefix(path, prefix) {
+		return "", ""
+	}
+	rest := strings.TrimPrefix(path, prefix) // <id>/refresh 或 <id>/explain
+	parts := strings.Split(rest, "/")
+	if len(parts) != 2 || parts[0] == "" {
+		return "", ""
+	}
+	dec, err := url.PathUnescape(parts[0])
+	if err != nil {
+		return "", ""
+	}
+	return dec, parts[1]
+}
+
+// Sub 分发 /api/v1/peer/{id}/refresh 与 /api/v1/peer/{id}/explain。
+func (pa *PeerAPI) Sub(w http.ResponseWriter, r *http.Request) {
+	id, action := peerSubPath(r.URL.Path)
+	if id == "" {
+		respondError(w, 404, "peer id required")
+		return
+	}
+	switch action {
+	case "refresh":
+		if r.Method != http.MethodPost {
+			http.Error(w, "POST only", http.StatusMethodNotAllowed)
+			return
+		}
+		pa.Refresh(w, r, id)
+	case "explain":
+		if r.Method != http.MethodGet {
+			http.Error(w, "GET only", http.StatusMethodNotAllowed)
+			return
+		}
+		pa.Explain(w, r, id)
+	default:
+		respondError(w, 404, "unknown action: "+action)
+	}
+}
+
+// RefreshAsync 异步刷新某实体的四层画像（推理不阻塞主流程，AC4）。
+// 返回一个 done channel，调用方可不等待（非阻塞）；推理在后台 goroutine 完成。
+// 任何错误只记日志，不影响调用方。
+func (pa *PeerAPI) RefreshAsync(id, namespace string) chan error {
+	done := make(chan error, 1)
+	if pa == nil || pa.Store == nil || pa.Reasoner == nil {
+		done <- fmt.Errorf("peer 未装配")
+		return done
+	}
+	go func() {
+		done <- pa.refresh(id, namespace, 2000)
+	}()
+	return done
+}
+
+// RefreshFromContent 提交新记忆后异步刷新内容里出现的已知实体（设计「新记忆进入 → 不阻塞主流程
+// → 更新该 peer 表示」）。匹配：content 含某 peer_id 子串即触发。全部 goroutine 内执行，立即返回。
+// 由 server.go 在 commit 处理器里以 ZHIYI_PEER_RECALL=1 门控调用；默认关 ⇒ 不触发。
+func (pa *PeerAPI) RefreshFromContent(content, namespace string) {
+	if pa == nil || pa.Store == nil || content == "" {
+		return
+	}
+	peers, err := pa.Store.List(200)
+	if err != nil || len(peers) == 0 {
+		return
+	}
+	for _, p := range peers {
+		if p.PeerID != "" && strings.Contains(content, p.PeerID) {
+			go pa.refresh(p.PeerID, namespace, 2000)
+		}
+	}
+}
+
+// refresh 实际刷新编排（同步执行体；由 RefreshAsync / RefreshFromContent 包 goroutine）。
+func (pa *PeerAPI) refresh(id, namespace string, limit int) error {
+	explicit, _, scanErr := pa.collectExplicit(id, namespace, limit)
+	if scanErr != nil {
+		return scanErr
+	}
+	graphCtx := pa.collectGraphContext(id)
+	if _, err := peer.BuildPeer(pa.Store, pa.Reasoner, id, explicit.facts, explicit.prefs, explicit.skills, graphCtx); err != nil {
+		return err
+	}
+	return nil
+}
+
+// Refresh POST /api/v1/peer/{id}/refresh — 扫 L1 记忆 + 图谱上下文 → 四层推理 → 写 peers 表。
+// body 可选 {"namespace": "...", "limit": N}（namespace 空 = 全库）。
+func (pa *PeerAPI) Refresh(w http.ResponseWriter, r *http.Request, id string) {
+	if pa.Store == nil || pa.Reasoner == nil {
+		respondError(w, 503, "peer 未装配（Store/Reasoner 为空）")
+		return
+	}
+	var req struct {
+		Namespace string `json:"namespace"`
+		Limit     int    `json:"limit"`
+	}
+	if r.Body != nil {
+		_ = json.NewDecoder(r.Body).Decode(&req)
+	}
+	if req.Limit <= 0 {
+		req.Limit = 2000
+	}
+	// 1. 扫 L1 记忆（LanceDB），提取 explicit 层
+	explicit, matched, scanErr := pa.collectExplicit(id, req.Namespace, req.Limit)
+	if scanErr != nil {
+		respondError(w, 500, "scan memories: "+scanErr.Error())
+		return
+	}
+	// 2. 图谱上下文（实体 1 跳邻居名）
+	graphCtx := pa.collectGraphContext(id)
+	// 3. 四层推理 + 写库
+	p, perr := peer.BuildPeer(pa.Store, pa.Reasoner, id, explicit.facts, explicit.prefs, explicit.skills, graphCtx)
+	if perr != nil {
+		respondError(w, 400, "build peer: "+perr.Error())
+		return
+	}
+	respond(w, 200, map[string]interface{}{
+		"peer":       p,
+		"matched":    matched,
+		"scanned":    req.Limit,
+		"graph_ctx":  graphCtx,
+		"has_reason": len(p.InferredTraits) > 0 || len(p.InferredGoals) > 0 || p.HypothesizedContext != "",
+	})
+}
+
+// Explain GET /api/v1/peer/{id}/explain — 可审计解释：画像 + 来源记忆。
+func (pa *PeerAPI) Explain(w http.ResponseWriter, r *http.Request, id string) {
+	if pa.Store == nil {
+		respondError(w, 503, "peer 未装配（Store 为空）")
+		return
+	}
+	p, err := pa.Store.Get(id)
+	if err != nil {
+		if err == sql.ErrNoRows {
+			respondError(w, 404, "peer not found: "+id+"（先 POST /peer/{id}/refresh 生成画像）")
+		} else {
+			respondError(w, 500, "get peer: "+err.Error())
+		}
+		return
+	}
+	// 来源记忆：扫 L1 里含该实体的记忆（仅读，不改库）
+	src := []string{}
+	if pa.LDB != nil {
+		if ldb := pa.LDB(); ldb != nil {
+			if recs, serr := ldb.Scan(2000, ""); serr == nil {
+				for _, m := range recs {
+					if strings.Contains(m.Content, id) {
+						src = append(src, m.Content)
+					}
+				}
+			}
+		}
+	}
+	respond(w, 200, map[string]interface{}{"peer": p, "source_memories": src})
+}
+
+// RecallPeer POST /api/v1/recall/peer?peer_id=… — 召回某 peer 的四层推理画像。
+// env 门控 ZHIYI_PEER_RECALL：默认 0 → 404（零变化）；=1 → 返回画像 JSON（AC3 判据）。
+func (pa *PeerAPI) RecallPeer(w http.ResponseWriter, r *http.Request) {
+	if !peerRecallEnabled() {
+		respondError(w, 404, "peer recall 默认关（ZHIYI_PEER_RECALL 未设/≠1）")
+		return
+	}
+	pid := r.URL.Query().Get("peer_id")
+	if pid == "" {
+		respondError(w, 400, "peer_id required")
+		return
+	}
+	if pa.Store == nil {
+		respondError(w, 503, "peer 未装配（Store 为空）")
+		return
+	}
+	p, err := pa.Store.Get(pid)
+	if err != nil {
+		if err == sql.ErrNoRows {
+			respondError(w, 404, "peer not found: "+pid)
+		} else {
+			respondError(w, 500, "get peer: "+err.Error())
+		}
+		return
+	}
+	respond(w, 200, map[string]interface{}{
+		"peer_id": p.PeerID,
+		"portrait": map[string]interface{}{
+			"explicit":  map[string]interface{}{"facts": p.Facts, "preferences": p.Preferences, "skills": p.Skills},
+			"deductive": map[string]interface{}{"traits": p.InferredTraits, "goals": p.InferredGoals},
+			"inductive": map[string]interface{}{"patterns": p.Patterns},
+			"abductive": map[string]interface{}{"hypothesized_context": p.HypothesizedContext},
+		},
+		"updated_at": p.UpdatedAt,
+	})
+}
+
+// explicitResult explicit 层收集结果。
+type explicitResult struct {
+	facts, prefs, skills []string
+}
+
+// collectExplicit 扫 L1 记忆并提取 explicit 层。namespace 空 = 全库。
+func (pa *PeerAPI) collectExplicit(id, namespace string, limit int) (explicitResult, int, error) {
+	var res explicitResult
+	if pa.LDB == nil {
+		return res, 0, nil
+	}
+	ldb := pa.LDB()
+	if ldb == nil {
+		return res, 0, nil
+	}
+	recs, err := ldb.Scan(limit, namespace)
+	if err != nil {
+		return res, 0, err
+	}
+	facts, prefs, skills, matched := peer.CollectFacts(recs, id)
+	return explicitResult{facts: facts, prefs: prefs, skills: skills}, matched, nil
+}
+
+// collectGraphContext 取实体图谱 1 跳邻居名作为推理上下文（读图，只读不改）。
+func (pa *PeerAPI) collectGraphContext(id string) []string {
+	if pa.GraphStore == nil {
+		return nil
+	}
+	paths, err := pa.GraphStore.Navigate(id, 1, "", nil)
+	if err != nil || len(paths) == 0 {
+		return nil
+	}
+	seen := map[string]bool{}
+	var out []string
+	for _, p := range paths {
+		for _, k := range []string{"to", "target"} {
+			if v, ok := p[k].(string); ok && v != "" && !seen[v] && v != id {
+				seen[v] = true
+				out = append(out, v)
+			}
+		}
+	}
+	if len(out) > 10 {
+		out = out[:10]
+	}
+	return out
+}
+
+// peerRecallEnabled 每次调用读 env（默认关，零行为变化）。
+func peerRecallEnabled() bool {
+	switch strings.ToLower(os.Getenv("ZHIYI_PEER_RECALL")) {
+	case "1", "true", "on":
+		return true
+	}
+	return false
+}
+
+// GetRecallResults 供 storage.PeerRecallProvider 装配（server.go 在 ZHIYI_PEER_RECALL=1
+// 时注册）。query 命中某实体 → 返回该 peer 的四层画像（Category="peer"）。
+// 命中判定：query 里含 peer_id 子串（大小写不敏感）；取最多 topK 个。无命中 → 空。
+func (pa *PeerAPI) GetRecallResults(query, namespace string, topK int) []models.RecallResult {
+	if pa == nil || pa.Store == nil || topK <= 0 {
+		return nil
+	}
+	peers, err := pa.Store.List(200)
+	if err != nil || len(peers) == 0 {
+		return nil
+	}
+	q := strings.ToLower(strings.TrimSpace(query))
+	var out []models.RecallResult
+	for _, p := range peers {
+		if p.PeerID != "" && q != "" && strings.Contains(q, strings.ToLower(p.PeerID)) {
+			content := "peer 画像: " + p.PeerID +
+				" | 事实: " + strings.Join(p.Facts, "; ") +
+				" | 偏好: " + strings.Join(p.Preferences, "; ") +
+				" | 技能: " + strings.Join(p.Skills, "; ") +
+				" | 演绎特质: " + strings.Join(p.InferredTraits, "; ") +
+				" | 演绎目标: " + strings.Join(p.InferredGoals, "; ") +
+				" | 归纳模式: " + strings.Join(p.Patterns, "; ") +
+				" | 最佳解释: " + p.HypothesizedContext
+			out = append(out, models.RecallResult{
+				ID:        "peer:" + p.PeerID,
+				Content:   content,
+				Category:  "peer",
+				Score:     1.0,
+				Namespace: namespace,
+				Source:    "peer_representation",
+			})
+			if len(out) >= topK {
+				break
+			}
+		}
+	}
+	return out
+}

go/internal/api/routes/peer_route_test.go

@@ -0,0 +1,255 @@
+// 织忆 MemoryWeave — P3 Peer Representation API 路由测试（t_f8fc97e6）
+//
+// 覆盖端点：POST /api/v1/peer/{id}/refresh + GET /api/v1/peer/{id}/explain
+//   - POST /api/v1/recall/peer（env 门控）+ storage.PeerRecallProvider 注入。
+//
+// 共享同一临时 graph.db（peers 表），零网络 / CGO / 真模型。
+package routes
+
+import (
+	"encoding/json"
+	"net/http"
+	"net/http/httptest"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/xiaoxue/memoryweave/internal/models"
+	"github.com/xiaoxue/memoryweave/internal/peer"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// peerScanLance 覆写 Scan（PeerAPI.Refresh/Explain 唯一用到的 LanceDB 方法）；
+// 其余方法嵌 nil 接口 —— 本测试路径不会调用，触达即 panic（暴露误用）。
+type peerScanLance struct {
+	storage.LanceDB
+	rows []models.MemoryRecord
+}
+
+func (f *peerScanLance) Scan(limit int, namespaceFilter string) ([]models.MemoryRecord, error) {
+	if limit > 0 && len(f.rows) > limit {
+		return f.rows[:limit], nil
+	}
+	return f.rows, nil
+}
+
+// newPeerTestAPI 构造共享同一 temp graph.db 的 PeerAPI。
+// GraphStore 传 nil（collectGraphContext 处理 nil ⇒ 无图上下文，不炸）。
+func newPeerTestAPI(t *testing.T, mems []models.MemoryRecord) *PeerAPI {
+	t.Helper()
+	dir := t.TempDir()
+	store := peer.NewStore(filepath.Join(dir, "graph.db"))
+	t.Cleanup(store.Close)
+	mlc := &peerScanLance{rows: mems}
+	return NewPeerAPI(store, &peer.FakeReasoner{}, func() storage.LanceDB { return mlc }, nil)
+}
+
+// fiveMems 造 5 条关于「牧尘」的 L1 记忆（AC1 输入构造）。
+func fiveMems() []models.MemoryRecord {
+	return []models.MemoryRecord{
+		{Content: "牧尘选择 Docker 部署"},
+		{Content: "牧尘偏好 SysV 风格"},
+		{Content: "牧尘擅长 Python"},
+		{Content: "牧尘厌恶 systemd"},
+		{Content: "牧尘使用 jq 处理配置"},
+	}
+}
+
+// TestPeer_RefreshRoute —— AC1 API 层：5 条记忆 → refresh → peers 表 1 行四层非空。
+func TestPeer_RefreshRoute(t *testing.T) {
+	api := newPeerTestAPI(t, fiveMems())
+	rec := httptest.NewRecorder()
+	req := httptest.NewRequest("POST", "/api/v1/peer/%E7%89%A7%E5%B0%98/refresh", nil) // 牧尘
+	api.Sub(rec, req)
+	if rec.Code != http.StatusOK {
+		t.Fatalf("refresh 期望 200，实际 %d body=%s", rec.Code, rec.Body.String())
+	}
+	var resp struct {
+		Peer    peer.Peer `json:"peer"`
+		Matched int       `json:"matched"`
+	}
+	if err := json.Unmarshal(rec.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("refresh 解析: %v", err)
+	}
+	if resp.Matched != 5 {
+		t.Errorf("AC1：matched=%d 期望 5", resp.Matched)
+	}
+	if resp.Peer.PeerID != "牧尘" {
+		t.Errorf("peer_id=%q 期望 牧尘", resp.Peer.PeerID)
+	}
+	// 四层非空（AC1 判据）
+	if len(resp.Peer.InferredTraits) == 0 {
+		t.Error("inferred_traits 空")
+	}
+	if len(resp.Peer.InferredGoals) == 0 {
+		t.Error("inferred_goals 空")
+	}
+	if len(resp.Peer.Patterns) == 0 {
+		t.Error("patterns 空")
+	}
+	if resp.Peer.HypothesizedContext == "" {
+		t.Error("hypothesized_context 空")
+	}
+	// peers 表行数（AC1：sqlite 查）
+	if n, _ := api.Store.Count(); n != 1 {
+		t.Errorf("AC1：peers 行数应 1，实际 %d", n)
+	}
+}
+
+// TestPeer_ExplainRoute —— GET explain → 200 + 画像 + 来源记忆。
+func TestPeer_ExplainRoute(t *testing.T) {
+	api := newPeerTestAPI(t, fiveMems())
+	// 先 refresh 种画像
+	rec := httptest.NewRecorder()
+	api.Sub(rec, httptest.NewRequest("POST", "/api/v1/peer/%E7%89%A7%E5%B0%98/refresh", nil))
+	if rec.Code != http.StatusOK {
+		t.Fatalf("前置 refresh 应 200，实际 %d", rec.Code)
+	}
+	// explain
+	er := httptest.NewRecorder()
+	api.Sub(er, httptest.NewRequest("GET", "/api/v1/peer/%E7%89%A7%E5%B0%98/explain", nil))
+	if er.Code != http.StatusOK {
+		t.Fatalf("explain 期望 200，实际 %d body=%s", er.Code, er.Body.String())
+	}
+	var resp struct {
+		Peer           peer.Peer `json:"peer"`
+		SourceMemories []string  `json:"source_memories"`
+	}
+	if err := json.Unmarshal(er.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("explain 解析: %v", err)
+	}
+	if len(resp.SourceMemories) != 5 {
+		t.Errorf("source_memories=%d 期望 5（可审计：画像有来源）", len(resp.SourceMemories))
+	}
+}
+
+// TestPeer_Explain_NotFound —— 未 refresh 直接 explain → 404。
+func TestPeer_Explain_NotFound(t *testing.T) {
+	api := newPeerTestAPI(t, fiveMems())
+	rec := httptest.NewRecorder()
+	api.Sub(rec, httptest.NewRequest("GET", "/api/v1/peer/%E7%89%A7%E5%B0%98/explain", nil))
+	if rec.Code != http.StatusNotFound {
+		t.Fatalf("无画像 explain 应 404，实际 %d", rec.Code)
+	}
+}
+
+// TestPeer_RecallPeer_DefaultOff —— AC3：ZHIYI_PEER_RECALL=0（默认）→ recall/peer 404 零变化。
+func TestPeer_RecallPeer_DefaultOff(t *testing.T) {
+	t.Setenv("ZHIYI_PEER_RECALL", "0")
+	api := newPeerTestAPI(t, fiveMems())
+	rec := httptest.NewRecorder()
+	api.RecallPeer(rec, httptest.NewRequest("POST", "/api/v1/recall/peer?peer_id=牧尘", nil))
+	if rec.Code != http.StatusNotFound {
+		t.Fatalf("默认关 recall/peer 应 404，实际 %d body=%s", rec.Code, rec.Body.String())
+	}
+}
+
+// TestPeer_RecallPeer_Enabled —— AC3：ZHIYI_PEER_RECALL=1 → 返回四层画像 JSON。
+func TestPeer_RecallPeer_Enabled(t *testing.T) {
+	t.Setenv("ZHIYI_PEER_RECALL", "1")
+	api := newPeerTestAPI(t, fiveMems())
+	// 先 refresh 种画像
+	rrec := httptest.NewRecorder()
+	api.Sub(rrec, httptest.NewRequest("POST", "/api/v1/peer/%E7%89%A7%E5%B0%98/refresh", nil))
+	if rrec.Code != http.StatusOK {
+		t.Fatalf("前置 refresh 应 200，实际 %d", rrec.Code)
+	}
+	// recall/peer
+	rec := httptest.NewRecorder()
+	api.RecallPeer(rec, httptest.NewRequest("POST", "/api/v1/recall/peer?peer_id=牧尘", nil))
+	if rec.Code != http.StatusOK {
+		t.Fatalf("recall/peer 期望 200，实际 %d body=%s", rec.Code, rec.Body.String())
+	}
+	var resp struct {
+		Portrait map[string]interface{} `json:"portrait"`
+	}
+	if err := json.Unmarshal(rec.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("recall/peer 解析: %v", err)
+	}
+	// 四层各有内容（AC3 判据）
+	for _, layer := range []string{"explicit", "deductive", "inductive", "abductive"} {
+		v, ok := resp.Portrait[layer]
+		if !ok {
+			t.Errorf("portrait 缺 %s 层: %v", layer, resp.Portrait)
+			continue
+		}
+		s, _ := json.Marshal(v)
+		if string(s) == "{}" || strings.Contains(string(s), "[]\"") {
+			t.Errorf("portrait.%s 层空: %s", layer, s)
+		}
+	}
+	t.Logf("AC3 recall/peer 画像: %s", rec.Body.String())
+}
+
+// TestPeer_GetRecallResults —— storage.PeerRecallProvider 注入：query 含实体 → 返回画像条目。
+func TestPeer_GetRecallResults(t *testing.T) {
+	api := newPeerTestAPI(t, fiveMems())
+	// 种画像
+	api.Sub(httptest.NewRecorder(),
+		httptest.NewRequest("POST", "/api/v1/peer/%E7%89%A7%E5%B0%98/refresh", nil))
+	res := api.GetRecallResults("牧尘怎么部署", "", 1)
+	if len(res) != 1 {
+		t.Fatalf("GetRecallResults 应返回 1 条，实际 %d", len(res))
+	}
+	if res[0].Category != "peer" || res[0].ID != "peer:牧尘" {
+		t.Errorf("返回条目 Category=%q ID=%q 期望 peer / peer:牧尘", res[0].Category, res[0].ID)
+	}
+	if !strings.Contains(res[0].Content, "演绎特质") || !strings.Contains(res[0].Content, "最佳解释") {
+		t.Errorf("画像内容缺四层: %s", res[0].Content)
+	}
+	// 无命中 → 空
+	if got := api.GetRecallResults("如何优化 Redis", "", 1); len(got) != 0 {
+		t.Errorf("不相关 query 应空，实际 %d", len(got))
+	}
+}
+
+// TestPeer_SubPath —— 子路径解析单元。
+func TestPeer_SubPath(t *testing.T) {
+	id, action := peerSubPath("/api/v1/peer/%E7%89%A7%E5%B0%98/refresh")
+	if id != "牧尘" || action != "refresh" {
+		t.Fatalf("解析失败 id=%q action=%q", id, action)
+	}
+	if _, a := peerSubPath("/api/v1/peer"); a != "" {
+		t.Fatalf("无子路径应 action 空")
+	}
+}
+
+// blockingReasoner 会阻塞的推理器（测非阻塞：RefreshAsync 不等它完成）。
+type blockingReasoner struct{ release chan struct{} }
+
+func (b *blockingReasoner) Reason(facts, prefs, skills, graphCtx []string) (*peer.ReasoningOutput, error) {
+	<-b.release
+	return &peer.ReasoningOutput{InferredTraits: []string{"T"}}, nil
+}
+
+// TestPeer_RefreshAsync_NonBlocking —— AC4：推理不阻塞主流程。
+// RefreshAsync 立即返回（后台 goroutine 推理），即使推理器被阻塞也绝不等它。
+func TestPeer_RefreshAsync_NonBlocking(t *testing.T) {
+	dir := t.TempDir()
+	store := peer.NewStore(filepath.Join(dir, "graph.db"))
+	t.Cleanup(store.Close)
+	br := &blockingReasoner{release: make(chan struct{})}
+	mlc := &peerScanLance{rows: fiveMems()}
+	api := NewPeerAPI(store, br, func() storage.LanceDB { return mlc }, nil)
+
+	done := api.RefreshAsync("牧尘", "")
+	select {
+	case <-done:
+		t.Fatal("RefreshAsync 应在推理器仍阻塞时立即返回（非阻塞），却等到了结果")
+	case <-time.After(200 * time.Millisecond):
+		// 非阻塞 ✓：200ms 内未等推理完成
+	}
+	// 释放推理器，后台完成
+	close(br.release)
+	select {
+	case <-done:
+		// 后台完成
+	case <-time.After(2 * time.Second):
+		t.Fatal("释放后后台推理未完成")
+	}
+	// 画像已落库
+	if n, _ := store.Count(); n != 1 {
+		t.Errorf("非阻塞刷新后 peers 行数应 1，实际 %d", n)
+	}
+}

go/internal/api/routes/provenance_observe.go

@@ -0,0 +1,238 @@
+// 织忆 MemoryWeave — commit 写入的「引用可复算」旁路观测（阶段 1→2：只观测，不拦截）
+//
+// 动机（t_3fbb6cbd）：Akasha 的确定性校验器（sourceRange + quoteHash，fail-closed）
+// 是唯一值得移植的东西。织忆现在的写入无法回证（09-25 实测卡点）。
+// 但现有写入方 payload 里没有 source_uri/quote/quote_sha256，直接上硬闸 = 拒掉 100%
+// 写入 = 掐死织忆全量摄取。所以本卡只做观测，先拿真实分布，再决定阈值/形态。
+//
+// 开关：ZHIYI_COMMIT_PROVENANCE_OBSERVE，默认关闭。关闭时行为与改前完全一致（零影响）。
+// 开启时对每次 commit 做旁路观测（异步 goroutine，不阻塞、不改响应、不改写入路径），
+// 结果追加到 ~/.hermes/logs/zhiyi-commit-provenance/YYYY-MM-DD.jsonl。
+//
+// 判定降级（AC2 重定义，t_dc412532 阶段2 设计输入）：
+//
+//	(a) metadata 有 source_uri + quote + quote_sha256 → 复算 sha256(quote) 比对；
+//	    不等 observable=false, reason=hash_mismatch
+//	(b) 无上述字段但 source=="session-compress" 且 session_id 存在 →
+//	    去 ~/.hermes/state.db（只读）取该 session 消息文本做回证。
+//	    观察期实测（n=205, 24h）：true=1 / false=8 / unknown=193；全部 8 条 false
+//	    均来自 session-compress —— 判据「前 60 字逐字命中 / LCS≥60」对「摘要/派生内容」
+//	    结构性必 false（摘要不会逐字出现在原文里）。AC2 修正：
+//	      * 原文不可得（session 已压缩/清除）→ observable=unknown, reason=source_not_available
+//	        （空原文 = 无法证伪，不是证伪失败 —— 旧逻辑误标 false/not_in_source）
+//	      * 前 60 字逐字命中 或 LCS≥60 → observable=true（引用片段可溯源）
+//	      * 其余（摘要低 LCS，属正常改写）→ observable=unknown, reason=derived_low_lcs
+//	        （对派生内容无法证伪 —— 摘要重述≠不可溯源）
+//	    同时把实测数值 lcs_len / prefix_hit60 写进 jsonl（AC1 先量后判），
+//	    供阶段3 转硬闸时按分布定档。
+//	(c) 其他 source → observable=unknown, reason=no_derivation（不要瞎猜成 false）
+package routes
+
+import (
+	"crypto/sha256"
+	"database/sql"
+	"encoding/hex"
+	"encoding/json"
+	"fmt"
+	"os"
+	"path/filepath"
+	"strings"
+	"sync"
+	"time"
+
+	_ "modernc.org/sqlite"
+)
+
+// observeMu 保护 jsonl 追加（commit 并发时防交错写）
+var observeMu sync.Mutex
+
+// provenanceObserveEnabled 开关（默认关闭，读一次即可）
+var provenanceObserveEnabled = os.Getenv("ZHIYI_COMMIT_PROVENANCE_OBSERVE") == "1" ||
+	strings.EqualFold(os.Getenv("ZHIYI_COMMIT_PROVENANCE_OBSERVE"), "true")
+
+// stateDBPath 只读来源库（session-to-zhiyi.py 等写入方读的同一个 state.db）
+const stateDBPath = "/home/muc/.hermes/state.db"
+
+// provenanceLogDir 观测结果落盘目录
+const provenanceLogDir = "/home/muc/.hermes/logs/zhiyi-commit-provenance"
+
+// minCommonRun 判定「可溯源」的最长公共子串下界（字符）。
+// 依据（t_dc412532 阶段2 实测分布）：观察期 session-compress 摘要 content_len 均值 ~450，
+// 要求 LCS≥60 取「摘要中至少出现一段可对回原文的片段」，过低会把无关噪声当溯源。
+const minCommonRun = 60
+
+// observeCommitProvenance 旁路观测一次 commit（异步调用，永不阻塞写路径）
+func observeCommitProvenance(agentID, content string, metadata map[string]any) {
+	if !provenanceObserveEnabled {
+		return
+	}
+	source := ""
+	sessionID := ""
+	if metadata != nil {
+		if s, ok := metadata["source"].(string); ok {
+			source = s
+		}
+		if sid, ok := metadata["session_id"].(string); ok {
+			sessionID = sid
+		}
+	}
+
+	observable := "unknown"
+	reason := "no_derivation"
+	lcsLen := 0
+	prefixHit60 := false
+
+	// (a) 显式 quote + hash
+	if metadata != nil {
+		uri, _ := metadata["source_uri"].(string)
+		quote, _ := metadata["quote"].(string)
+		qh, _ := metadata["quote_sha256"].(string)
+		if uri != "" && quote != "" && qh != "" {
+			observable, reason = decideQuoteHash(quote, qh)
+		} else if source == "session-compress" && sessionID != "" {
+			// (b) session-compress：去 state.db 只读回证
+			observable, reason, lcsLen, prefixHit60 = observeViaSession(content, sessionID)
+		}
+	} else if source == "session-compress" && sessionID != "" {
+		observable, reason, lcsLen, prefixHit60 = observeViaSession(content, sessionID)
+	}
+
+	// (c) 其余保持 unknown / no_derivation
+
+	line := map[string]interface{}{
+		"ts":           time.Now().Format(time.RFC3339),
+		"agent_id":     agentID,
+		"source":       source,
+		"session_id":   sessionID,
+		"observable":   observable,
+		"reason":       reason,
+		"content_len":  len(content),
+		"lcs_len":      lcsLen,      // AC1: content 与原文的最长公共子串长度（实测数值）
+		"prefix_hit60": prefixHit60, // AC1: 前 60 字是否逐字命中原文
+	}
+	appendProvenanceLog(line)
+}
+
+// decideQuoteHash 复算 sha256(quote) 与调用方声明的 quote_sha256 比对。
+func decideQuoteHash(quote, quoteSHA256 string) (string, string) {
+	sum := sha256.Sum256([]byte(quote))
+	if hex.EncodeToString(sum[:]) == quoteSHA256 {
+		return "true", ""
+	}
+	return "false", "hash_mismatch"
+}
+
+// observeViaSession 去 state.db（只读）取 session 消息文本，回证 content 可溯源。
+// 返回 (observable, reason, lcs_len, prefix_hit60)。
+// AC2 语义：对派生/摘要内容不做逐字强判 —— 原文不可得→unknown（无法证伪），
+// 命中→true（片段可溯源），低 LCS→unknown（摘要改写正常，不可证伪）。
+func observeViaSession(content, sessionID string) (string, string, int, bool) {
+	db, err := sql.Open("sqlite", "file:"+stateDBPath+"?_pragma=query_only(1)&_pragma=busy_timeout(3000)")
+	if err != nil {
+		return "unknown", "state_db_open_failed: " + err.Error(), 0, false
+	}
+	defer db.Close()
+	rows, err := db.Query(
+		"SELECT content FROM messages WHERE session_id=? AND role IN ('user','assistant') ORDER BY id",
+		sessionID)
+	if err != nil {
+		return "unknown", "state_db_query_failed: " + err.Error(), 0, false
+	}
+	defer rows.Close()
+
+	var sb strings.Builder
+	for rows.Next() {
+		var msg string
+		if err := rows.Scan(&msg); err != nil {
+			continue
+		}
+		sb.WriteString(msg)
+		sb.WriteByte('\n')
+	}
+	srcText := sb.String()
+	if srcText == "" {
+		// 原文不可得：无法证伪，不是证伪失败（旧逻辑误标 false/not_in_source）
+		return "unknown", "source_not_available", 0, false
+	}
+	match, lcsLen, prefixHit60 := provenanceMeasure(content, srcText)
+	if match {
+		return "true", "", lcsLen, prefixHit60
+	}
+	// 摘要/派生内容低 LCS 属正常改写，无法证伪 —— 不算 false
+	return "unknown", "derived_low_lcs", lcsLen, prefixHit60
+}
+
+// provenanceMeasure 计算 content 相对 srcText 的可溯源实测指标：
+//
+//	prefixHit60 = content 前 60 字在 srcText 中逐字命中
+//	lcsLen      = content 与 srcText 的最长公共子串长度
+//	match       = prefixHit60 || lcsLen >= minCommonRun
+func provenanceMeasure(content, srcText string) (match bool, lcsLen int, prefixHit60 bool) {
+	if content == "" || srcText == "" {
+		return false, 0, false
+	}
+	first := content
+	if len(first) > minCommonRun {
+		first = first[:minCommonRun]
+	}
+	prefixHit60 = strings.Contains(srcText, first)
+	// LCS 兜底（受限输入防打爆）
+	c := content
+	if len(c) > 2000 {
+		c = c[:2000]
+	}
+	s := srcText
+	if len(s) > 20000 {
+		s = s[:20000]
+	}
+	lcsLen = lcsLength(c, s)
+	return prefixHit60 || lcsLen >= minCommonRun, lcsLen, prefixHit60
+}
+
+// lcsLength 两个字符串的最长公共子串长度（O(n*m) 动态规划，取较短者做列）。
+func lcsLength(a, b string) int {
+	if len(a) == 0 || len(b) == 0 {
+		return 0
+	}
+	// 让 b 为较短者（节省列空间）
+	if len(b) > len(a) {
+		a, b = b, a
+	}
+	prev := make([]int, len(b)+1)
+	cur := make([]int, len(b)+1)
+	best := 0
+	for i := 1; i <= len(a); i++ {
+		for j := 1; j <= len(b); j++ {
+			if a[i-1] == b[j-1] {
+				cur[j] = prev[j-1] + 1
+				if cur[j] > best {
+					best = cur[j]
+				}
+			} else {
+				cur[j] = 0
+			}
+		}
+		prev, cur = cur, prev
+	}
+	return best
+}
+
+// appendProvenanceLog 把一行观测结果追加到当日 jsonl。
+func appendProvenanceLog(line map[string]interface{}) {
+	observeMu.Lock()
+	defer observeMu.Unlock()
+	data, err := json.Marshal(line)
+	if err != nil {
+		return
+	}
+	if err := os.MkdirAll(provenanceLogDir, 0o755); err != nil {
+		return
+	}
+	path := filepath.Join(provenanceLogDir, time.Now().Format("2006-01-02")+".jsonl")
+	f, err := os.OpenFile(path, os.O_APPEND|os.O_CREATE|os.O_WRONLY, 0o644)
+	if err != nil {
+		return
+	}
+	defer f.Close()
+	fmt.Fprintf(f, "%s\n", data)
+}

go/internal/api/routes/provenance_observe_test.go

@@ -0,0 +1,49 @@
+package routes
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"strings"
+	"testing"
+)
+
+func TestProvenanceMatch(t *testing.T) {
+	src := strings.Repeat("源文本", 50) // 150 chars
+	// 前60字符逐字命中
+	content := strings.Repeat("源文本", 25) // 75 chars
+	if m, _, _ := provenanceMeasure(content, src); !m {
+		t.Fatal("expect first-60-chars substring match")
+	}
+	// 完全无关 → false
+	if m, _, _ := provenanceMeasure(strings.Repeat("完全无关", 20), src); m {
+		t.Fatal("expect no match for unrelated")
+	}
+	// LCS >= 60 兜底（content 中间有长公共段）
+	content2 := strings.Repeat("P", 10) + strings.Repeat("源文本", 40) + strings.Repeat("Q", 5)
+	src2 := strings.Repeat("A", 30) + strings.Repeat("源文本", 40) + strings.Repeat("B", 30)
+	if m, _, _ := provenanceMeasure(content2, src2); !m {
+		t.Fatal("expect LCS>=60 fallback match")
+	}
+}
+
+func TestLCSLength(t *testing.T) {
+	if got := lcsLength("abcde", "abfce"); got != 2 { // "ab"
+		t.Fatalf("lcs=2 expect, got %d", got)
+	}
+	if got := lcsLength("", "abc"); got != 0 {
+		t.Fatalf("lcs empty=0 expect, got %d", got)
+	}
+}
+
+func TestDecideQuoteHash(t *testing.T) {
+	quote := "这段引用内容"
+	sum := sha256.Sum256([]byte(quote))
+	observable, reason := decideQuoteHash(quote, hex.EncodeToString(sum[:]))
+	if observable != "true" || reason != "" {
+		t.Fatalf("expect true/'' got %s/%s", observable, reason)
+	}
+	observable, reason = decideQuoteHash(quote, strings.Repeat("0", 64))
+	if observable != "false" || reason != "hash_mismatch" {
+		t.Fatalf("expect false/hash_mismatch got %s/%s", observable, reason)
+	}
+}

go/internal/api/routes/recall_error_shape_test.go

@@ -0,0 +1,319 @@
+// 织忆 MemoryWeave — recall 错误形状统一测试（t_2d4e3b92, 2026-09-21）
+//
+// 契约：
+//   - 200 + "results" 键存在 = 成功（results 可为空数组，属正常"无匹配"）
+//   - 非 2xx = 失败（body 为统一错误体 {"error":{"code","message","retry_after_s"}}）
+//   - 调用方绝不可用 len(results)==0 区分成功/失败
+package routes
+
+import (
+	"encoding/json"
+	"net/http/httptest"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/xiaoxue/memoryweave/internal/governance"
+	"github.com/xiaoxue/memoryweave/internal/models"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// ─── 测试桩 ────────────────────────────────────────────────
+
+// fakeLanceDB 实现 storage.LanceDB，Search 可配置返回结果或错误。
+type fakeLanceDB struct {
+	searchResults []models.MemoryRecord
+	searchErr     error
+}
+
+func (f *fakeLanceDB) InsertEpisode(agentID, namespace, content, category string) (string, error) {
+	return "ep_test", nil
+}
+func (f *fakeLanceDB) GetTopByQuality(agentID string, limit int) ([]models.MemoryRecord, error) {
+	return nil, nil
+}
+func (f *fakeLanceDB) Stats() (map[string]interface{}, error) {
+	return map[string]interface{}{}, nil
+}
+func (f *fakeLanceDB) SoftDelete(id, reason string) error { return nil }
+func (f *fakeLanceDB) PurgeInvalid() (int64, error)       { return 0, nil }
+func (f *fakeLanceDB) SoftDeleteBatch(table string, ids []string, updatedAt string) (int64, error) {
+	return 0, nil
+}
+func (f *fakeLanceDB) GetVersionHistory(id string) ([]map[string]interface{}, error) {
+	return nil, nil
+}
+func (f *fakeLanceDB) GetCandidatesForForgetting() ([]map[string]interface{}, error) {
+	return nil, nil
+}
+func (f *fakeLanceDB) GetSkillCandidates(minRecalls, limit int) ([]models.MemoryRecord, error) {
+	return nil, nil
+}
+func (f *fakeLanceDB) Backup(path string) error { return nil }
+func (f *fakeLanceDB) GetAuditLog(limit int) ([]map[string]interface{}, error) {
+	return nil, nil
+}
+func (f *fakeLanceDB) IncrementUseful(id string) error { return nil }
+func (f *fakeLanceDB) IncrementNotUseful(id string) error {
+	return nil
+}
+
+// GA-P0-1 (t_718d8707)：拆分字段
+func (f *fakeLanceDB) IncrementHumanNotUseful(id string) error  { return nil }
+func (f *fakeLanceDB) AppendHitEvent(memoryID, outcome, reason string) error { return nil }
+func (f *fakeLanceDB) CountHitEvents(memoryID, outcome string, since time.Time) (int, error) { return 0, nil }
+func (f *fakeLanceDB) GetHitEvents(memoryID string, limit int) ([]models.HitEvent, error) { return nil, nil }
+func (f *fakeLanceDB) IncrementDedupDownweight(id string) error { return nil }
+func (f *fakeLanceDB) UpdateMemoryContent(id, newContent, source string) error {
+	return nil
+}
+func (f *fakeLanceDB) InsertMemory(m models.MemoryRecord) error { return nil }
+func (f *fakeLanceDB) Search(table string, vector []float32, topK int, namespaceFilter string) ([]models.MemoryRecord, error) {
+	if f.searchErr != nil {
+		return nil, f.searchErr
+	}
+	if len(f.searchResults) > topK {
+		return f.searchResults[:topK], nil
+	}
+	return f.searchResults, nil
+}
+func (f *fakeLanceDB) Scan(limit int, namespaceFilter string) ([]models.MemoryRecord, error) {
+	return f.searchResults, nil
+}
+func (f *fakeLanceDB) Insert(table string, record any) error { return nil }
+func (f *fakeLanceDB) Update(table, id string, fields map[string]any) error {
+	return nil
+}
+
+// ─── 测试用例 ─────────────────────────────────────────────
+
+// TestRecallSuccess_Returns200WithResults 验证成功路径：200 + results 键存在 + candidates_scanned。
+func TestRecallSuccess_Returns200WithResults(t *testing.T) {
+	db := &fakeLanceDB{
+		searchResults: []models.MemoryRecord{
+			{ID: "mem_1", Content: "test memory one", Category: "distilled"},
+			{ID: "mem_2", Content: "test memory two", Category: "distilled"},
+		},
+	}
+	gs := governance.NewInMemoryGraph()
+	emb := storage.NewEmbedder("")
+	rerank := storage.NewReranker("", emb)
+	api := NewAPI(db, emb, rerank, nil, gs)
+
+	body := `{"query":"test query","limit":5}`
+	req := httptest.NewRequest("POST", "/api/v1/recall", strings.NewReader(body))
+	w := httptest.NewRecorder()
+	api.Recall(w, req)
+
+	if w.Code != 200 {
+		t.Fatalf("expected 200, got %d; body=%s", w.Code, w.Body.String())
+	}
+	var resp map[string]interface{}
+	if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("failed to parse response: %v", err)
+	}
+	if _, hasResults := resp["results"]; !hasResults {
+		t.Error("expected 'results' key in success response, missing — contract violated")
+	}
+	if _, hasCount := resp["count"]; !hasCount {
+		t.Error("expected 'count' key in success response")
+	}
+	if _, hasScanned := resp["candidates_scanned"]; !hasScanned {
+		t.Error("expected 'candidates_scanned' key in success response (t_2d4e3b92; t_a70b07e1 语义修订)")
+	}
+	t.Logf("✓ success: code=200, keys=%v", mapKeys(resp))
+}
+
+// TestRecallSuccess_EmptyResults_Still200WithCandidatesScanned 验证"真无匹配"：
+// results=[] 但 candidates_scanned 字段存在，与依赖故障（非 2xx）可区分。
+func TestRecallSuccess_EmptyResults_Still200WithCandidatesScanned(t *testing.T) {
+	db := &fakeLanceDB{searchResults: nil} // ANN 返回 0 条
+	gs := governance.NewInMemoryGraph()
+	emb := storage.NewEmbedder("")
+	rerank := storage.NewReranker("", emb)
+	api := NewAPI(db, emb, rerank, nil, gs)
+
+	body := `{"query":"nonsense query with no match","limit":5}`
+	req := httptest.NewRequest("POST", "/api/v1/recall", strings.NewReader(body))
+	w := httptest.NewRecorder()
+	api.Recall(w, req)
+
+	// embedder 无真实 endpoint → encode 失败 → 主路径 err != nil
+	// InMemoryGraph FallbackTextSearch 返回空 → 走 502 失败路径
+	// 所以这里验证的是失败路径（非 2xx），而非"200 空结果"
+	if w.Code >= 200 && w.Code < 300 {
+		// 如果是 200，验证 schema
+		var resp map[string]interface{}
+		if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+			t.Fatalf("failed to parse response: %v", err)
+		}
+		if _, hasResults := resp["results"]; !hasResults {
+			t.Error("expected 'results' key even when empty (contract: 200+results=success)")
+		}
+		if _, hasScanned := resp["candidates_scanned"]; !hasScanned {
+			t.Error("expected 'candidates_scanned' key to distinguish no-match from failure (t_a70b07e1 语义：主路径召回前候选数，非最终返回数)")
+		}
+		t.Logf("✓ empty 200: candidates_scanned=%v", resp["candidates_scanned"])
+	} else {
+		// 非 2xx：验证统一错误体
+		var resp map[string]interface{}
+		if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+			t.Fatalf("failed to parse error response: %v", err)
+		}
+		errObj, hasErr := resp["error"].(map[string]interface{})
+		if !hasErr {
+			t.Fatalf("expected nested error object, got %T", resp["error"])
+		}
+		if _, hasCode := errObj["code"]; !hasCode {
+			t.Error("expected 'error.code' in unified error schema")
+		}
+		if _, hasRetry := errObj["retry_after_s"]; !hasRetry {
+			t.Error("expected 'error.retry_after_s' in unified error schema")
+		}
+		t.Logf("✓ failure: code=%d, error=%v", w.Code, errObj)
+	}
+}
+
+// fakeEmbedder 实现 *storage.Embedder 的行为：EncodeSingle 可配置成功/失败。
+// 用于测试 recall 失败路径（模拟依赖不可用）。
+type fakeEmbedder struct {
+	encodeErr error
+	dim       int
+}
+
+// 由于 *storage.Embedder 是具体类型而非接口，我们通过 NewAPI 注入真实 Embedder，
+// 并在测试中利用其内部 endpoint 为空的行为。
+// 这里提供一个辅助方法：通过给 fakeLanceDB 返回 0 条 + 真实 embedder 无 endpoint
+// 来触发 Pipeline.Recall 失败。
+//
+// 实际上 storage.Embedder.Encode 在 endpoint="" 且无 apiKey 时会走 HTTP 请求到 ""，
+// 导致失败。这正是我们需要的失败行为。
+
+// TestRecallFailure_Non2xx 验证失败路径：依赖不可用时返回非 2xx + 统一错误体。
+// 模拟：embedder endpoint="" → encode 必失败（HTTP 到空 URL）⇒ 主路径 err != nil
+//   - InMemoryGraph FallbackTextSearch 返回空 ⇒ 502 失败路径。
+func TestRecallFailure_Non2xx(t *testing.T) {
+	db := &fakeLanceDB{searchResults: nil}
+	gs := governance.NewInMemoryGraph()
+	emb := storage.NewEmbedder("") // 无 endpoint → encode 必失败
+	rerank := storage.NewReranker("", emb)
+	api := NewAPI(db, emb, rerank, nil, gs)
+
+	body := `{"query":"test","limit":5,"mode":"hybrid"}`
+	req := httptest.NewRequest("POST", "/api/v1/recall", strings.NewReader(body))
+	w := httptest.NewRecorder()
+	api.Recall(w, req)
+
+	// 注意：Pipeline.Recall 内部有缓存和降级逻辑，encode 失败时可能返回空结果而非错误。
+	// 如果 Pipeline 自身降级成功（返回 0 条 + nil err），则走"真无匹配"路径（200）。
+	// 要真正触发 502 需要 Pipeline 返回 error。
+	// 这里我们验证两种情况都符合契约：
+	if w.Code >= 200 && w.Code < 300 {
+		// 成功路径：必须带 results 键 + candidates_scanned
+		var resp map[string]interface{}
+		if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+			t.Fatalf("parse failed: %v", err)
+		}
+		if _, hasResults := resp["results"]; !hasResults {
+			t.Error("200 response must have 'results' key (contract)")
+		}
+		if _, hasScanned := resp["candidates_scanned"]; !hasScanned {
+			t.Error("200 response must have 'candidates_scanned' key (t_2d4e3b92; t_a70b07e1 语义修订)")
+		}
+		t.Logf("✓ 200 path: candidates_scanned=%v, count=%v", resp["candidates_scanned"], resp["count"])
+	} else {
+		// 非 2xx 路径：必须带统一错误体
+		var resp map[string]interface{}
+		if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+			t.Fatalf("parse failed: %v", err)
+		}
+		errObj, ok := resp["error"].(map[string]interface{})
+		if !ok {
+			t.Fatalf("expected nested error object, got %T", resp["error"])
+		}
+		if _, hasCode := errObj["code"]; !hasCode {
+			t.Error("expected 'error.code' in unified error schema")
+		}
+		if _, hasRetry := errObj["retry_after_s"]; !hasRetry {
+			t.Error("expected 'error.retry_after_s' in unified error schema")
+		}
+		t.Logf("✓ non-2xx path: code=%d, error=%v", w.Code, errObj)
+	}
+}
+
+// TestRecallDirect502 直接构造 502 场景：用 reflect 注入 Pipeline 使其 Recall 返回错误。
+// 由于 RecallPipeline 是具体类型，我们通过其内部 embedder 的 endpoint 不可用来触发。
+// 这里直接调用 respondErrorSchema 验证输出格式。
+func TestRecallDirect502(t *testing.T) {
+	w := httptest.NewRecorder()
+	respondErrorSchema(w, 502, "recall_backend_error",
+		"recall failed: encode error; graph fallback also empty", 5)
+
+	if w.Code != 502 {
+		t.Fatalf("expected 502, got %d", w.Code)
+	}
+	var resp map[string]interface{}
+	if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+		t.Fatalf("parse failed: %v", err)
+	}
+	errObj, _ := resp["error"].(map[string]interface{})
+	if errObj == nil {
+		t.Fatal("expected nested error object in 502 response")
+	}
+	if errObj["code"] != "recall_backend_error" {
+		t.Errorf("expected code=recall_backend_error, got %v", errObj["code"])
+	}
+	if errObj["retry_after_s"] != float64(5) {
+		t.Errorf("expected retry_after_s=5, got %v", errObj["retry_after_s"])
+	}
+	t.Logf("✓ 502 body: %v", resp)
+}
+
+// TestRecallKeywordFailure_Non2xx 验证 keyword 模式主路径失败时也返回非 2xx。
+func TestRecallKeywordFailure_Non2xx(t *testing.T) {
+	db := &fakeLanceDB{searchResults: nil}
+	gs := governance.NewInMemoryGraph()
+	emb := storage.NewEmbedder("") // 无 endpoint → encode 必失败
+	rerank := storage.NewReranker("", emb)
+	api := NewAPI(db, emb, rerank, nil, gs)
+
+	body := `{"query":"test","limit":5,"mode":"keyword"}`
+	req := httptest.NewRequest("POST", "/api/v1/recall", strings.NewReader(body))
+	w := httptest.NewRecorder()
+	api.Recall(w, req)
+
+	// 同 TestRecallFailure_Non2xx：Pipeline 内部降级可能使 err==nil（返回 0 条），
+	// 此时走 200 路径。验证两种情况都符合契约。
+	if w.Code >= 200 && w.Code < 300 {
+		var resp map[string]interface{}
+		if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+			t.Fatalf("parse failed: %v", err)
+		}
+		if _, hasResults := resp["results"]; !hasResults {
+			t.Error("keyword 200 response must have 'results' key")
+		}
+		if _, hasScanned := resp["candidates_scanned"]; !hasScanned {
+			t.Error("keyword 200 response must have 'candidates_scanned' key")
+		}
+		t.Logf("✓ keyword 200: candidates_scanned=%v", resp["candidates_scanned"])
+	} else {
+		var resp map[string]interface{}
+		if err := json.Unmarshal(w.Body.Bytes(), &resp); err != nil {
+			t.Fatalf("parse failed: %v", err)
+		}
+		errObj, _ := resp["error"].(map[string]interface{})
+		if errObj == nil {
+			t.Error("keyword non-2xx must have unified error object")
+		}
+		t.Logf("✓ keyword non-2xx: code=%d", w.Code)
+	}
+}
+
+// helper
+func mapKeys(m map[string]interface{}) []string {
+	keys := make([]string, 0, len(m))
+	for k := range m {
+		keys = append(keys, k)
+	}
+	return keys
+}

go/internal/api/routes/recall_namespace_test.go

@@ -0,0 +1,62 @@
+package routes
+
+import "testing"
+
+// TestResolveRecallNamespace 覆盖 2026-09-21 五修（t_24eceed0 返修）的 ns 解析规则。
+//
+// 病根：修前「无 namespace 且无 agent_id」会推导成 default-main，导致默认召回
+// 只搜 default-main + shared，而真实记忆散在 hermes-main / wiki-curator-main / wiki /
+// soulful-xiaowei / a06-main / domain:* 等 19 个命名空间（全表 10172 条，实测只拿到 11 个候选）。
+//
+// 判据：两者都空必须返回 ""（= 全库跨 ns），而不是任何具体 ns。
+func TestResolveRecallNamespace(t *testing.T) {
+	cases := []struct {
+		name      string
+		namespace string
+		agentID   string
+		query     string
+		want      string
+		why       string
+	}{
+		{
+			name: "无ns无agentID_必须全库", namespace: "", agentID: "",
+			query: "蒸馏模型日限计数器", want: "",
+			why: "核心判据：默认调用必须跨全部 ns 检索（返回空串 = 不过滤 = 全库）",
+		},
+		{
+			name: "显式ns_原样作用域", namespace: "hermes-main", agentID: "",
+			query: "任意", want: "hermes-main",
+			why: "调用方显式指定时不得改写语义",
+		},
+		{
+			name: "显式ns优先于agentID", namespace: "shared", agentID: "hermes-a06",
+			query: "任意", want: "shared",
+			why: "显式 ns 优先级最高",
+		},
+		{
+			name: "只有agentID_推导ns", namespace: "", agentID: "hermes-a06",
+			query: "任意", want: "hermes-main",
+			why: "有 agent_id 时按 §1.3 推导，保持既有作用域行为",
+		},
+		{
+			name: "只有agentID_openclaw", namespace: "", agentID: "openclaw",
+			query: "任意", want: "openclaw-main",
+			why: "openclaw 前缀推导",
+		},
+	}
+
+	for _, c := range cases {
+		got := resolveRecallNamespace(c.namespace, c.agentID, c.query)
+		if got != c.want {
+			t.Errorf("✗ %s: resolveRecallNamespace(%q,%q) = %q, want %q —— %s",
+				c.name, c.namespace, c.agentID, got, c.want, c.why)
+			continue
+		}
+		t.Logf("✓ %s: got=%q", c.name, got)
+	}
+
+	// 独立强调：全库判据不能被"看起来像 ns"的默认值替代
+	if got := resolveRecallNamespace("", "", "x"); got != "" {
+		t.Fatalf("✗ 回归：默认调用被解析成 %q，又只能看到极小一片记忆", got)
+	}
+}

go/internal/api/routes/recall_truncated_test.go

@@ -0,0 +1,144 @@
+package routes
+
+// t_3aa2a8c1：`/api/v1/recall` 响应暴露 `truncated`（Rust 侧单次重读行数上限截断标志，
+// 引擎已算、此前只打日志不上抛）。选定的省略语义 = **键缺失即未截断（omitempty）**。
+// 本测试经路由层机械判定：
+// ① 粗检索被截断（后端实现 SearchTruncationReporter 且上报 true）⇒ 响应含 "truncated":true
+// ② 未截断 / 纯 Go 后端（不实现 reporter）⇒ 键缺失（不出现 truncated:false 以外的值）
+// ③ 回归：既有字段 results/count/candidates_scanned/mode/cache_hit 行为不变
+
+import (
+	"encoding/json"
+	"net/http/httptest"
+	"strings"
+	"testing"
+
+	"github.com/xiaoxue/memoryweave/internal/governance"
+	"github.com/xiaoxue/memoryweave/internal/models"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// truncReportingLDB 复用同包 fakeLanceDB（recall_error_shape_test.go），
+// 额外实现可选扩展接口 SearchWithTruncationReporter（t_3aa2a8c1）。
+type truncReportingLDB struct {
+	*fakeLanceDB
+	reportTruncated bool
+}
+
+func (f *truncReportingLDB) SearchWithTruncation(table string, vector []float32, topK int, namespaceFilter string) ([]models.MemoryRecord, bool, error) {
+	if f.searchErr != nil {
+		return nil, false, f.searchErr
+	}
+	if len(f.searchResults) > topK {
+		return f.searchResults[:topK], f.reportTruncated, nil
+	}
+	return f.searchResults, f.reportTruncated, nil
+}
+
+func recallBody(t *testing.T, db storage.LanceDB, body string) (int, map[string]interface{}, string) {
+	t.Helper()
+	// 测试须可复跑（t_3aa2a8c1 实测踩坑）：SearchCacheInstance 由 storage 包 init() 恒建
+	// （L2 = 共享 Redis，TTL 60s）⇒ 上一次 run 的 recall 结果会进 L2，下一次 run 同 key
+	// 命中缓存路径（Truncated 恒 false）⇒ 断言 flaky。逐测试摘除缓存（测后恢复），
+	// 恒走未缓存路径。包内无 t.Parallel ⇒ 改全局安全。
+	saved := storage.SearchCacheInstance
+	storage.SearchCacheInstance = nil
+	t.Cleanup(func() { storage.SearchCacheInstance = saved })
+
+	emb := storage.NewEmbedder("")
+	rerank := storage.NewReranker("", emb)
+	api := NewAPI(db, emb, rerank, nil, governance.NewInMemoryGraph())
+	req := httptest.NewRequest("POST", "/api/v1/recall", strings.NewReader(body))
+	w := httptest.NewRecorder()
+	api.Recall(w, req)
+	var m map[string]interface{}
+	_ = json.Unmarshal(w.Body.Bytes(), &m)
+	return w.Code, m, w.Body.String()
+}
+
+// 既有响应键回归基线（t_2d4e3b92 成功 schema + t_1c12e000 cache_hit + t_949840b1 candidates_scanned）
+var recallBaselineKeys = []string{"results", "count", "candidates_scanned", "mode", "cache_hit"}
+
+func assertBaseline(t *testing.T, m map[string]interface{}, body string) {
+	t.Helper()
+	for _, k := range recallBaselineKeys {
+		if _, ok := m[k]; !ok {
+			t.Fatalf("既有字段 %q 缺失（回归破坏），body=%s", k, body)
+		}
+	}
+	if res, ok := m["results"].([]interface{}); ok {
+		if cnt, _ := m["count"].(float64); int(cnt) != len(res) {
+			t.Fatalf("count=%v 与 len(results)=%d 不一致（回归破坏）", m["count"], len(res))
+		}
+	}
+}
+
+// ① 粗检索被行数上限截断 ⇒ 响应暴露 truncated:true + 既有字段回归不变
+func TestRecall_TruncatedExposedWhenHitCap(t *testing.T) {
+	db := &truncReportingLDB{
+		fakeLanceDB: &fakeLanceDB{searchResults: []models.MemoryRecord{
+			{ID: "mem_1", Content: "test memory one", Category: "distilled"},
+			{ID: "mem_2", Content: "test memory two", Category: "distilled"},
+		}},
+		reportTruncated: true,
+	}
+	code, m, body := recallBody(t, db, `{"query":"test query","limit":5,"namespace":"shared"}`)
+	if code != 200 {
+		t.Fatalf("expected 200, got %d; body=%s", code, body)
+	}
+	if v, ok := m["truncated"]; !ok || v != true {
+		t.Fatalf("粗检索已截断，响应必须含 truncated:true；实际 key 存在=%v 值=%v body=%s", ok, m["truncated"], body)
+	}
+	assertBaseline(t, m, body)
+}
+
+// ①b keyword 主路径（kwMeta 透传）同判据
+func TestRecall_Keyword_TruncatedExposed(t *testing.T) {
+	db := &truncReportingLDB{
+		fakeLanceDB: &fakeLanceDB{searchResults: []models.MemoryRecord{
+			{ID: "mem_1", Content: "keyword memory", Category: "distilled"},
+		}},
+		reportTruncated: true,
+	}
+	code, m, body := recallBody(t, db, `{"query":"keyword query","limit":5,"mode":"keyword","namespace":"shared"}`)
+	if code != 200 {
+		t.Fatalf("expected 200, got %d; body=%s", code, body)
+	}
+	if v, ok := m["truncated"]; !ok || v != true {
+		t.Fatalf("keyword 路径粗检索已截断，响应必须含 truncated:true；body=%s", body)
+	}
+	assertBaseline(t, m, body)
+}
+
+// ② 未截断 ⇒ 键缺失（选定语义：省略=未截断），不得出现 truncated:false 以外的值
+func TestRecall_NotTruncated_KeyOmitted(t *testing.T) {
+	db := &truncReportingLDB{
+		fakeLanceDB: &fakeLanceDB{searchResults: []models.MemoryRecord{
+			{ID: "mem_1", Content: "test memory one", Category: "distilled"},
+		}},
+		reportTruncated: false,
+	}
+	code, m, body := recallBody(t, db, `{"query":"test query","limit":5,"namespace":"shared"}`)
+	if code != 200 {
+		t.Fatalf("expected 200, got %d; body=%s", code, body)
+	}
+	if v, ok := m["truncated"]; ok {
+		t.Fatalf("未截断场景键应缺失（省略=未截断）；实际 truncated=%v body=%s", v, body)
+	}
+	assertBaseline(t, m, body)
+}
+
+// ②b 纯 Go 后端（不实现 SearchTruncationReporter）⇒ 键缺失，零值语义、行为不变
+func TestRecall_PureGoBackend_NoTruncatedKey(t *testing.T) {
+	db := &fakeLanceDB{searchResults: []models.MemoryRecord{
+		{ID: "mem_1", Content: "test memory one", Category: "distilled"},
+	}}
+	code, m, body := recallBody(t, db, `{"query":"test query","limit":5,"namespace":"shared"}`)
+	if code != 200 {
+		t.Fatalf("expected 200, got %d; body=%s", code, body)
+	}
+	if _, ok := m["truncated"]; ok {
+		t.Fatalf("纯 Go 后端无行数上限，truncated 键必须缺失；body=%s", body)
+	}
+	assertBaseline(t, m, body)
+}

go/internal/api/routes/routes_test.go

@@ -4,6 +4,8 @@ package routes
 import (
 	"bytes"
 	"encoding/json"
+	"errors"
+	"net/http"
 	"net/http/httptest"
 	"testing"
 	"time"
@@ -15,6 +17,11 @@ import (
 // ─── 核心 API（mock LanceDB）───────────────────────────────
 
 func TestHealthEndpoint(t *testing.T) {
+	// 侧车可达 ⇒ 200 + status ok
+	orig := sidecarHealthCheck
+	defer func() { sidecarHealthCheck = orig }()
+	sidecarHealthCheck = func() error { return nil }
+
 	req := httptest.NewRequest("GET", "/health", nil)
 	w := httptest.NewRecorder()
 	HandleHealth(w, req)
@@ -27,6 +34,38 @@ func TestHealthEndpoint(t *testing.T) {
 	if resp["status"] != "ok" {
 		t.Errorf("expected ok status, got %s", resp["status"])
 	}
+	if resp["sidecar"] != "ok" {
+		t.Errorf("expected sidecar ok, got %s", resp["sidecar"])
+	}
+}
+
+// TestHealthEndpoint_SidecarDown 回归项（2026-10-08 故障）：
+// sidecar 不可达时 /health 必须返回 503，而不是 200 —— 防"假绿"。
+func TestHealthEndpoint_SidecarDown(t *testing.T) {
+	orig := sidecarHealthCheck
+	defer func() { sidecarHealthCheck = orig }()
+	sidecarHealthCheck = func() error {
+		return errors.New("zhiyi-consolidate not reachable: dial unix /tmp/zhiyi-ipc.sock: connect: no such file or directory")
+	}
+
+	req := httptest.NewRequest("GET", "/health", nil)
+	w := httptest.NewRecorder()
+	HandleHealth(w, req)
+
+	if w.Code != http.StatusServiceUnavailable {
+		t.Errorf("expected 503 when sidecar down, got %d", w.Code)
+	}
+	var resp map[string]string
+	json.NewDecoder(w.Body).Decode(&resp)
+	if resp["status"] != "degraded" {
+		t.Errorf("expected degraded status, got %s", resp["status"])
+	}
+	if resp["sidecar"] != "unreachable" {
+		t.Errorf("expected sidecar unreachable, got %s", resp["sidecar"])
+	}
+	if resp["error"] == "" {
+		t.Error("expected error detail in body, got empty")
+	}
 }
 
 // ─── 知识图谱路由 ────────────────────────────────────────
@@ -128,7 +167,7 @@ func TestConflictAPI_Resolve_Validation(t *testing.T) {
 // ─── 知识缺口路由 ────────────────────────────────────────
 
 func TestGapAPI_List(t *testing.T) {
-	gd := selfoptimize.NewGapDetector()
+	gd := selfoptimize.NewGapDetector(nil, nil)
 	ai := NewGapAPI(gd)
 
 	req := httptest.NewRequest("GET", "/api/v1/gaps", nil)
@@ -141,7 +180,7 @@ func TestGapAPI_List(t *testing.T) {
 }
 
 func TestGapAPI_Detect(t *testing.T) {
-	gd := selfoptimize.NewGapDetector()
+	gd := selfoptimize.NewGapDetector(nil, nil)
 	ai := NewGapAPI(gd)
 
 	body, _ := json.Marshal(map[string]string{"topic": "test_gap"})
@@ -199,8 +238,8 @@ func TestTriggers_List(t *testing.T) {
 	}
 	var resp map[string]interface{}
 	json.NewDecoder(w.Body).Decode(&resp)
-	if resp["count"].(float64) != 4 {
-		t.Errorf("expected 4 triggers, got %v", resp["count"])
+	if resp["count"].(float64) != 8 {
+		t.Errorf("expected 8 triggers, got %v", resp["count"])
 	}
 }
 
@@ -306,18 +345,19 @@ func TestAuxFunctions(t *testing.T) {
 	}
 }
 
-// ─── SSE 管理器 ───────────────────────────────────────────
+// ─── WebSocket 事件推送 ───────────────────────────────────
 
-func TestSSEManager_Push(t *testing.T) {
-	// SSEBus.Push should not panic with no clients
-	SSEBus.Push(SSEMessage{Type: "test", Payload: "hello"})
-}
-
-func TestSSEPushHelpers(t *testing.T) {
-	// All push helpers should not panic
-	PushMemoryCommitted("agent", "ns", "mem1")
-	PushConflictDetected("Docker", "facts conflict")
-	PushGapFound("topic", "A")
+func TestWSPushHelpers_NoPanic(t *testing.T) {
+	// 所有推送 helper 在无订阅者时都不该 panic
+	// 2026-09-10 修复：原用例引用的 SSEBus/SSEMessage/PushMemoryCommitted/PushGapFound
+	// 已随 SSE→WebSocket 重构删除，改用当前 API（PushGapDetected/PushConflict* 新签名）。
+	PushPrefetch("agent", []interface{}{})
+	PushGapDetected("agent", "topic", "knowledge", "补一下")
+	PushGapFilled("agent", 1)
+	PushMemoryUpdated("mem1", 2, "test")
+	PushConflictDetected("agent", "c1", "Docker", map[string]string{"a": "b"})
+	PushConflictResolved("agent", "c1", "keep_left")
 	PushConsolidationDone("done")
-	PushConflictResolved("c1", "keep_left")
+	PushQualityDrop("agent", "mem1", 0.2, "补证据")
+	PushDistillationComplete(1, 0)
 }

go/internal/api/routes/skill_bayes.go

@@ -17,18 +17,18 @@ const skillRedisKey = "zhiyi:skills"
 
 // BetaSkill 贝叶斯 Skill 评分（G7 扩展版）
 type BetaSkill struct {
-	Name            string    `json:"name"`
-	Alpha           float64   `json:"alpha"` // α = successes + 1
-	Beta            float64   `json:"beta"` // β = failures + 1
-	Trials          int       `json:"trials"`
-	Successes       int       `json:"successes"`
-	ETA             float64   `json:"eta"` // α/(α+β) 贝叶斯均值
-	Status          string    `json:"status"` // active / probation / retired
-	LastUpdated     time.Time `json:"last_updated"`
+	Name        string    `json:"name"`
+	Alpha       float64   `json:"alpha"` // α = successes + 1
+	Beta        float64   `json:"beta"`  // β = failures + 1
+	Trials      int       `json:"trials"`
+	Successes   int       `json:"successes"`
+	ETA         float64   `json:"eta"`    // α/(α+β) 贝叶斯均值
+	Status      string    `json:"status"` // active / probation / retired
+	LastUpdated time.Time `json:"last_updated"`
 	// G7 扩展字段
-	PromptTemplate  string   `json:"prompt_template,omitempty"`  // 可执行 prompt（含 {param} 占位符）
-	LinkedMemoryIDs []string `json:"linked_memory_ids,omitempty"` // 关联的记忆 ID
-	LinkedEntities  []string `json:"linked_entities,omitempty"` // 关联的图谱实体
+	PromptTemplate  string    `json:"prompt_template,omitempty"`   // 可执行 prompt（含 {param} 占位符）
+	LinkedMemoryIDs []string  `json:"linked_memory_ids,omitempty"` // 关联的记忆 ID
+	LinkedEntities  []string  `json:"linked_entities,omitempty"`   // 关联的图谱实体
 	CreatedAt       time.Time `json:"created_at"`
 }
 
@@ -146,12 +146,12 @@ func (bsm *BayesianSkillManager) Register(name, promptTemplate string, linkedEnt
 	skill, exists := bsm.skills[name]
 	if !exists {
 		skill = &BetaSkill{
-			Name:            name,
-			Alpha:           1.0,
-			Beta:            1.0,
-			CreatedAt:       time.Now(),
-			PromptTemplate:  promptTemplate,
-			LinkedEntities:  linkedEntities,
+			Name:           name,
+			Alpha:          1.0,
+			Beta:           1.0,
+			CreatedAt:      time.Now(),
+			PromptTemplate: promptTemplate,
+			LinkedEntities: linkedEntities,
 		}
 		bsm.skills[name] = skill
 	} else {
@@ -306,11 +306,11 @@ func (bsm *BayesianSkillManager) Stats() map[string]interface{} {
 		}
 	}
 	return map[string]interface{}{
-		"total":      len(bsm.skills),
-		"active":      active,
-		"probation":   probation,
-		"retired":     retired,
-		"persisted":   bsm.persisted,
+		"total":           len(bsm.skills),
+		"active":          active,
+		"probation":       probation,
+		"retired":         retired,
+		"persisted":       bsm.persisted,
 		"redis_connected": bsm.redisClient != nil,
 	}
 }
@@ -371,10 +371,10 @@ func (sm *SkillManager) Trial(w http.ResponseWriter, r *http.Request) {
 // POST /api/v1/skills — 注册新 skill（G7 新增）
 func (sm *SkillManager) Register(w http.ResponseWriter, r *http.Request) {
 	var req struct {
-		Name            string   `json:"name"`
-		PromptTemplate  string   `json:"prompt_template"`
-		LinkedEntities  []string `json:"linked_entities"`
-		Description     string   `json:"description"`
+		Name           string   `json:"name"`
+		PromptTemplate string   `json:"prompt_template"`
+		LinkedEntities []string `json:"linked_entities"`
+		Description    string   `json:"description"`
 	}
 	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
 		respondError(w, 400, "invalid body: "+err.Error())
@@ -427,4 +427,4 @@ func (sm *SkillManager) GetSkillETA(name string) float64 {
 		return 0.5
 	}
 	return skill.ETA
-}
+}

go/internal/api/routes/skill_crystallize.go

@@ -153,8 +153,8 @@ func crystallizeFromRecord(mem *models.MemoryRecord) (*BetaSkill, error) {
 	if err == nil && llmResult != "" {
 		var parsed struct {
 			PromptTemplate string   `json:"prompt_template"`
-			Entities      []string `json:"entities"`
-			SkillName     string   `json:"skill_name"`
+			Entities       []string `json:"entities"`
+			SkillName      string   `json:"skill_name"`
 		}
 		if err2 := json.Unmarshal([]byte(llmResult), &parsed); err2 == nil {
 			if parsed.PromptTemplate != "" {
@@ -318,4 +318,4 @@ func getLDB() storage.LanceDB {
 		return ldbGetter()
 	}
 	return nil
-}
+}

go/internal/api/routes/skill_execute.go

@@ -240,4 +240,3 @@ func ExecuteSkill(w http.ResponseWriter, r *http.Request) {
 		"eta":             skill.ETA,
 	})
 }
-

go/internal/api/routes/standing_answers.go

@@ -0,0 +1,72 @@
+package routes
+
+import (
+	"encoding/json"
+	"net/http"
+
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// 2026-10-13 (t_e9a4d18f P2) 常设答案路由（Hindsight mental models 直读/写入）。
+// 读取 = 一次 DB 读（question_id 主键），零检索零 LLM。写入由生成脚本
+// zhiyi-standing-answers.py 调用（复用本地 2B，免费）。
+
+// standingStore 返回全局常设答案库（server.go 注入）。未初始化 → nil。
+func standingStore() *storage.StandingStore { return storage.StandingStoreInstance }
+
+// HandleStandingAnswers GET/POST /api/v1/standing-answers
+//   GET ?id=<question_id> → 直读一条；无 id → 列全部
+//   POST {"question_id","question","answer","source_memories":[]} → 写入/改写
+func HandleStandingAnswers(w http.ResponseWriter, r *http.Request) {
+	s := standingStore()
+	if s == nil {
+		respondError(w, http.StatusServiceUnavailable, "standing answers store not initialized")
+		return
+	}
+	switch r.Method {
+	case http.MethodGet:
+		id := r.URL.Query().Get("id")
+		if id == "" {
+			all, err := s.List()
+			if err != nil {
+				respondError(w, http.StatusInternalServerError, err.Error())
+				return
+			}
+			respond(w, http.StatusOK, map[string]interface{}{"answers": all, "count": len(all)})
+			return
+		}
+		a, err := s.Get(id)
+		if err != nil {
+			respondError(w, http.StatusInternalServerError, err.Error())
+			return
+		}
+		if a == nil {
+			respondError(w, http.StatusNotFound, "standing answer not found: "+id)
+			return
+		}
+		respond(w, http.StatusOK, a)
+	case http.MethodPost:
+		var in struct {
+			QuestionID     string   `json:"question_id"`
+			Question       string   `json:"question"`
+			Answer         string   `json:"answer"`
+			SourceMemories []string `json:"source_memories"`
+		}
+		if err := json.NewDecoder(r.Body).Decode(&in); err != nil {
+			respondError(w, http.StatusBadRequest, "bad json: "+err.Error())
+			return
+		}
+		if in.QuestionID == "" || in.Answer == "" {
+			respondError(w, http.StatusBadRequest, "question_id and answer required")
+			return
+		}
+		a, err := s.Upsert(in.QuestionID, in.Question, in.Answer, in.SourceMemories)
+		if err != nil {
+			respondError(w, http.StatusInternalServerError, err.Error())
+			return
+		}
+		respond(w, http.StatusOK, a)
+	default:
+		http.Error(w, "method not allowed", http.StatusMethodNotAllowed)
+	}
+}

go/internal/api/routes/system_template.go

@@ -0,0 +1,25 @@
+// 织忆 MemoryWeave — 系统通知/提示样板文本判定（被动注入入口噪音闸门）
+//
+// 背景（t_689b6d38，2026-10-11 实测）：织忆被动注入入口不区分「真实用户消息」与
+// 「系统样板文本」。近 7 天 [graph/notes] 通道 806 次注入里 543 次（67%）的入参是
+// 系统通知/提示样板（形如 "[IMPORTANT: You are running as a scheduled cron job...",
+// "[IMPORTANT: Background process ... completed", "[System note: ..."），它们照常走
+// 完整实体拆解并返回 3 篇笔记注入上下文 —— 纯浪费 + 稀释有效注入信噪比。
+//
+// 判定优先用【结构化信号】（消息来源标记头，如 "[IMPORTANT:"、"[HEARTBEAT]"、
+// "[System note:"、"Cronjob Response:"、role/来源字段），而不是纯文本长度启发式。
+// 命中 ⇒ 调用方应跳过语义召回与笔记返回。白名单：真实用户提问（含飞书/QQ/CLI 入站
+// 消息）不命中任何前缀/子串 ⇒ 照常召回，不误杀。
+package routes
+
+import "github.com/xiaoxue/memoryweave/internal/textfilter"
+
+// IsSystemTemplateInput 判定一段文本是否系统通知/提示样板（非真实用户提问）。
+// 命中 ⇒ 调用方应跳过语义召回与笔记返回（t_689b6d38 噪音闸门）。
+//
+// 2026-10-12：判定清单已抽到叶子包 internal/textfilter（蒸馏侧准入闸门也要用，
+// 而 distill 不能反向 import routes）。本函数保留为薄委托，调用点零改动。
+// 清单只维护 textfilter 一处。
+func IsSystemTemplateInput(s string) bool {
+	return textfilter.IsSystemTemplate(s)
+}

go/internal/api/routes/testdata/golden-set.json

@@ -0,0 +1,187 @@
+{
+  "description": "织忆召回金标检查集（单一真相源，t_949840b1）。统一判据口径 N=5：取 recall 返回前 N=5 条查关键词，ID 命中 = expected_ids 任一 ∈ 前 N=5 条。基线已刷新（golden-check.py --refresh，两量具同源）。",
+  "items": [
+    {
+      "query": "织忆是什么",
+      "expected_ids": [
+        "mem_1791666702950691634"
+      ],
+      "must_include_keywords": [
+        "织忆"
+      ],
+      "category": "system_fact",
+      "difficulty": "easy",
+      "id_baseline": "mem_1791666702950691634",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "织忆的图谱扩展机制是什么",
+      "expected_ids": [
+        "mem_1791899478673190284"
+      ],
+      "must_include_keywords": [
+        "图谱"
+      ],
+      "category": "system_fact",
+      "difficulty": "medium",
+      "id_baseline": "mem_1791899478673190284",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "织忆如何通过 E1 提取和 MMR 实现多样去重",
+      "expected_ids": [
+        "mem_1791666890486125605"
+      ],
+      "must_include_keywords": [
+        "MMR"
+      ],
+      "category": "system_fact",
+      "difficulty": "hard",
+      "id_baseline": "mem_1791666890486125605",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "牧尘的偏好是什么",
+      "expected_ids": [
+        "mem_1790448247125520424"
+      ],
+      "must_include_keywords": [
+        "牧尘"
+      ],
+      "category": "user_pref",
+      "difficulty": "easy",
+      "id_baseline": "mem_1790448247125520424",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "牧尘喜欢什么样的工作方式",
+      "expected_ids": [
+        "mem_1790448247125520424"
+      ],
+      "must_include_keywords": [
+        "牧尘"
+      ],
+      "category": "user_pref",
+      "difficulty": "medium",
+      "id_baseline": "mem_1790448247125520424",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "如何根据牧尘的偏好调整记忆检索策略",
+      "expected_ids": [
+        "mem_1791747597032219499"
+      ],
+      "must_include_keywords": [
+        "记忆"
+      ],
+      "category": "user_pref",
+      "difficulty": "hard",
+      "id_baseline": "mem_1791747597032219499",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "当前项目的技术栈是什么",
+      "expected_ids": [
+        "mem_1790455676397796671"
+      ],
+      "must_include_keywords": [
+        "技术栈"
+      ],
+      "category": "proj_context",
+      "difficulty": "easy",
+      "id_baseline": "mem_1790455676397796671",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "织忆项目有哪些核心组件",
+      "expected_ids": [
+        "mem_1791906354849281414"
+      ],
+      "must_include_keywords": [
+        "织忆"
+      ],
+      "category": "proj_context",
+      "difficulty": "medium",
+      "id_baseline": "mem_1791906354849281414",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "织忆和其他记忆系统相比有什么架构优势",
+      "expected_ids": [
+        "mem_1791666702950691634"
+      ],
+      "must_include_keywords": [
+        "织忆"
+      ],
+      "category": "proj_context",
+      "difficulty": "hard",
+      "id_baseline": "mem_1791666702950691634",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "如何使用 recall 接口",
+      "expected_ids": [
+        "mem_1791796178328736793"
+      ],
+      "must_include_keywords": [
+        "recall"
+      ],
+      "category": "tool_usage",
+      "difficulty": "easy",
+      "id_baseline": "mem_1791796178328736793",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "织忆有哪些管理工具",
+      "expected_ids": [
+        "mem_1791666702950691634"
+      ],
+      "must_include_keywords": [
+        "织忆"
+      ],
+      "category": "tool_usage",
+      "difficulty": "medium",
+      "id_baseline": "mem_1791666702950691634",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "如何通过 API 扩展织忆功能",
+      "expected_ids": [
+        "mem_1791666188168772049"
+      ],
+      "must_include_keywords": [
+        "织忆"
+      ],
+      "category": "tool_usage",
+      "difficulty": "hard",
+      "id_baseline": "mem_1791666188168772049",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "本地 llama 服务 8080 端口",
+      "expected_ids": [
+        "mem_1791423503359721152"
+      ],
+      "must_include_keywords": [
+        "8080"
+      ],
+      "category": "target_knowledge",
+      "difficulty": "easy",
+      "id_baseline": "mem_1791423503359721152",
+      "namespace": "hermes-main"
+    },
+    {
+      "query": "session-to-zhiyi.py 脚本名",
+      "expected_ids": [
+        "mem_1791688470278719760"
+      ],
+      "must_include_keywords": [
+        "session-to-zhiyi"
+      ],
+      "category": "target_knowledge",
+      "difficulty": "easy",
+      "id_baseline": "mem_1791688470278719760",
+      "namespace": "hermes-main"
+    }
+  ]
+}

go/internal/api/routes/testmain_test.go

@@ -0,0 +1,52 @@
+// 织忆 MemoryWeave — routes 包测试隔离（两个护栏合并 · t_e5997ca4 + t_b7a8ed9c）
+//
+// ⚠️ 本文件是**两个 TestMain 的合并体**。两卡分支各自加了一个 TestMain：
+//   · t_e5997ca4 → testmain_test.go（落盘点护栏）
+//   · t_b7a8ed9c → scorecache_guard_test.go（打分缓存护栏）
+// 在各自分支上互不干扰，合并到 main 后同包两个 TestMain **编译冲突**
+// （multiple definitions of TestMain）⇒ 必须合进一个文件。
+// （此点由 t_b7a8ed9c 在 scorecache_guard_test.go 顶部预先标注；小怡合并时未读文件
+//  而漏看，由全套测试的 setup failed 拦下 —— 见 worklog 教训。）
+//
+// 护栏①：单元测试不得写到生产落盘点（t_e5997ca4）
+// ---------------------------------------------------------------
+// selfoptimize 包已有同款 TestMain 先例（曾实测：整套测试跑完把
+// /var/lib/memoryweave/dashboard.json 写成测试数字）。
+// gap close/repair 端点会触发 Dash.persist()（本地 JSON 兜底）⇒ 统一重定向到临时目录。
+//
+// 护栏②：默认关闭 distill 打分缓存（t_b7a8ed9c）
+// ---------------------------------------------------------------
+// 背景：distill 打分记忆化（0dc571d）+ Redis L2（36db9bb，键 zhiyi:distill-score:*，
+// TTL 168h）后，**本包 E2E 测试（zz_commit_conflict_gate_test.go）会读写生产 Redis**：
+// 第一次 run L2 miss → 真调 mock LLM 并把「测试内容哈希 → 结果」写进生产；
+// 7 天内后续 run L2 hit → LLM 被静默短路 → 数 LLM 调用次数的断言（base+1）必挂。
+// 实测 2026-10-17：TestCommitConflictGate_E2E 16.5s 固定 FAIL，
+// 「expected >= 1 LLM distill calls, got 0」。
+//
+// 治法（对齐项目门禁规矩③：加缓存必须给该包 TestMain 加关闭开关）：
+// Setenv ZHIYI_SCORE_CACHE=0 —— distill.getScoreCache 的 Once 闭包**活读 env**
+// （包级变量在 init 时已定型，跨包设不了），本进程内缓存整体禁用：
+// 不读 L2（LLM 真被调用，计数断言有覆盖）、不写 L2（不再污染生产）。
+package routes
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+)
+
+func TestMain(m *testing.M) {
+	// 护栏①：落盘点重定向到临时目录（宁可测试直接失败，也不写生产落盘点）
+	dir, err := os.MkdirTemp("", "routes-test-dash-*")
+	if err != nil {
+		panic(err)
+	}
+	_ = os.Setenv("ZHIYI_DASHBOARD_PATH", filepath.Join(dir, "dashboard.json"))
+
+	// 护栏②：关闭 distill 打分缓存（不读/不写生产 Redis L2）
+	_ = os.Setenv("ZHIYI_SCORE_CACHE", "0")
+
+	code := m.Run()
+	_ = os.RemoveAll(dir)
+	os.Exit(code)
+}

go/internal/api/routes/triggers.go

@@ -42,8 +42,8 @@ type Trigger struct {
 	Condition   string      `json:"condition"`
 	Urgency     float64     `json:"urgency"` // 0-1，越大越紧急
 	Active      bool        `json:"active"`
-	Paused      bool        `json:"paused"`    // kill-switch
-	Cooldown    string      `json:"cooldown"`  // 冷却时间
+	Paused      bool        `json:"paused"`     // kill-switch
+	Cooldown    string      `json:"cooldown"`   // 冷却时间
 	FailCount   int         `json:"fail_count"` // 连续失败计数
 	LastFiredAt time.Time   `json:"last_fired_at,omitempty"`
 	Description string      `json:"description"`
@@ -189,5 +189,3 @@ func (tm *TriggerManager) RecordFail(id string) {
 		}
 	}
 }
-
-

go/internal/api/routes/tuning.go

@@ -2,11 +2,16 @@
 package routes
 
 import (
+	"encoding/json"
+	"log"
 	"net/http"
+	"os"
+	"strings"
 	"sync"
 	"time"
 
 	"github.com/xiaoxue/memoryweave/internal/selfoptimize"
+	"github.com/xiaoxue/memoryweave/internal/storage"
 )
 
 // TuningConfig 可自动调优的参数
@@ -20,7 +25,16 @@ type TuningConfig struct {
 	ConsolidateAfter int     `json:"consolidate_after"` // N 条蒸馏后触发整合
 
 	// 历史调参记录
-	history  []TuningEvent
+	history []TuningEvent
+
+	// lastSkipped 上一次 AutoTune 里「因指标不可用而跳过」的规则与原因。
+	// 让外部能区分「健康所以没调」与「数据不可用所以没调」——后者需要修数据源，不是好事。
+	lastSkipped []string
+
+	// lastMoveAt 每个参数上一次被改动的时间 —— 用于冷却（治棘轮）。
+	lastMoveAt map[string]time.Time
+	// persistEnabled 仅全局 Tuner 为 true；测试里构造的实例为 false ⇒ 不写盘。
+	persistEnabled bool
 }
 
 type TuningEvent struct {
@@ -32,11 +46,149 @@ type TuningEvent struct {
 }
 
 var Tuner = &TuningConfig{
-	Diversity:        0.5,
+	// ⚠️ 初值必须等于**真实消费方的默认值**，否则接线当天就改变生产行为：
+	// Diversity 对齐 storage.DefaultMMRDiversity() 的 0.3（core.go 原硬编码），原为 0.5。
+	Diversity:        0.3,
 	DecayRate:        0.015,
 	GapThreshold:     3,
 	DistillInterval:  5,
 	ConsolidateAfter: 50,
+
+	lastMoveAt:     map[string]time.Time{},
+	persistEnabled: true,
+}
+
+// tuneCooldown 同一参数的调参冷却时间。
+// 为什么需要：**AutoTune 不幂等** —— 同一判据每调一次就推一步（实测连跑两次：
+// decay_rate 0.015→0.018→0.0216、distill_interval 5→4→3），直到撞上限。
+// 闭环每天跑一次 ⇒ 几天内把参数顶到极值。冷却把「持续推进」限制成
+// 「同一指标须持续越界多个窗口，才多次调整」。
+const tuneCooldown = 6 * time.Hour
+
+// minTuningSamples 调参判据的最小样本数。
+// 为什么需要：小样本的比率噪声极大 —— 实测现网 recall_usefulness_rate 只有 9 个样本，
+// 8 好 1 坏 = 0.89，再来 1 个好评就 > 0.9 ⇒ 触发「降低 diversity」直接改生产参数。
+// 少样本不是证据。
+const minTuningSamples = 30
+
+// ratioFor 取一个「可用于调参」的比率指标，不可用则返回 ok=false。
+//
+// 约定（internal/selfoptimize/selfoptimize.go:248-322）：
+//
+//	<name>_rate    —— 比率值；**分母退化时返回 -1**（metricNA），既不是 0 也不是 1
+//	<name>_samples —— 该比率的分母样本数
+//	<name>_valid   —— 分母是否非退化
+//
+// ⚠️ AutoTune 原先只比较数值，无视这套约定 ⇒ `-1` 会让规则对「无数据」**恒真**：
+//
+//	gap_closure_rate = -1 < 0.5  ⇒「降低 gap 阈值」每轮都触发（按垃圾指标改生产参数）
+//	recall_usefulness_rate 样本 9 条 ⇒ 一次偶然就改 Diversity
+//
+// 本函数把「无数据 / 退化 / 样本不足」三类一律判为不可用 —— 宁可不动。
+// ⚠️ 命名不是统一的：比率键为 `<base>_rate`，而 *样本/有效* 键挂在 **base** 上
+// 而不是带 _rate 的那个串上 —— 实际键是 `recall_usefulness_samples`，
+// 不是 `recall_usefulness_rate_samples`（我第一版就按后者拼，护栏静默失效、
+// 参数照样被改；新加的 tuning_guard_test.go 当场逮住）。
+// 且 `avg_distill_loss` 这类**值键本身不带 _rate 后缀** ⇒ 统一用 TrimSuffix 推 base。
+func ratioFor(metrics map[string]float64, valueKey string) (float64, bool) {
+	base := strings.TrimSuffix(valueKey, "_rate")
+	v, ok := metrics[valueKey]
+	if !ok || v < 0 {
+		return 0, false // 缺失 或 -1 哨兵 = 无数据
+	}
+	if valid, ok := metrics[base+"_valid"]; ok && valid == 0 {
+		return 0, false // 分母退化
+	}
+	if samples, ok := metrics[base+"_samples"]; ok && samples < minTuningSamples {
+		return 0, false // 样本不足以支撑判断
+	}
+	return v, true
+}
+
+// canMove 判断某参数当前是否允许改动（冷却窗口内不许重复动 —— 治棘轮）。
+func (tc *TuningConfig) canMove(param string) bool {
+	last, ok := tc.lastMoveAt[param]
+	return !ok || time.Since(last) >= tuneCooldown
+}
+
+func (tc *TuningConfig) markMove(param string) {
+	if tc.lastMoveAt == nil {
+		tc.lastMoveAt = map[string]time.Time{}
+	}
+	tc.lastMoveAt[param] = time.Now()
+}
+
+// pushToConsumers 把当前参数推给**真实消费方**。
+// 🔴 这一步是「调参有没有用」的分水岭：2026-10-14 实测，此前 Tuner 的全部字段
+// 没有任何读者（全仓只有 3 处路由注册引用它）⇒ 调参改了数字、行为零变化。
+func (tc *TuningConfig) pushToConsumers() {
+	storage.SetDefaultMMRDiversity(tc.Diversity)
+	selfoptimize.SetGapThreshold(tc.GapThreshold)
+}
+
+// tuningStatePath 持久化路径（可用 ZHIYI_TUNING_STATE 覆盖，便于测试/隔离）。
+func tuningStatePath() string {
+	if p := os.Getenv("ZHIYI_TUNING_STATE"); p != "" {
+		return p
+	}
+	return "/var/lib/memoryweave/tuning-state.json"
+}
+
+type tuningState struct {
+	Diversity    float64           `json:"diversity"`
+	GapThreshold int               `json:"gap_threshold"`
+	LastMoveAt   map[string]string `json:"last_move_at"`
+	UpdatedAt    string            `json:"updated_at"`
+}
+
+// persist 落盘（失败只告警，不影响调参本身）。
+func (tc *TuningConfig) persist() {
+	if !tc.persistEnabled {
+		return
+	}
+	st := tuningState{Diversity: tc.Diversity, GapThreshold: tc.GapThreshold,
+		LastMoveAt: map[string]string{}, UpdatedAt: time.Now().Format(time.RFC3339)}
+	for k, v := range tc.lastMoveAt {
+		st.LastMoveAt[k] = v.Format(time.RFC3339)
+	}
+	data, err := json.MarshalIndent(st, "", " ")
+	if err != nil {
+		log.Printf("[tuning] 序列化失败: %v", err)
+		return
+	}
+	if err := os.WriteFile(tuningStatePath(), data, 0o644); err != nil {
+		log.Printf("[tuning] 落盘失败: %v", err)
+	}
+}
+
+// load 启动时载入持久化参数（无文件则用默认值），并立即接线到消费方。
+// 没有这一步，重启即回默认 ⇒ 调参成果每次重启清零。
+func (tc *TuningConfig) load() {
+	data, err := os.ReadFile(tuningStatePath())
+	if err == nil {
+		var st tuningState
+		if json.Unmarshal(data, &st) == nil {
+			if st.Diversity > 0 {
+				tc.Diversity = st.Diversity
+			}
+			if st.GapThreshold > 0 {
+				tc.GapThreshold = st.GapThreshold
+			}
+			for k, v := range st.LastMoveAt {
+				if t, e := time.Parse(time.RFC3339, v); e == nil {
+					if tc.lastMoveAt == nil {
+						tc.lastMoveAt = map[string]time.Time{}
+					}
+					tc.lastMoveAt[k] = t
+				}
+			}
+			log.Printf("[tuning] 已载入持久化参数 diversity=%.3f gap_threshold=%d（%s）",
+				tc.Diversity, tc.GapThreshold, st.UpdatedAt)
+		}
+	}
+	tc.mu.Lock()
+	tc.pushToConsumers()
+	tc.mu.Unlock()
 }
 
 // AutoTune 基于仪表盘指标自动调参
@@ -45,61 +197,103 @@ func (tc *TuningConfig) AutoTune(metrics map[string]float64) []TuningEvent {
 	defer tc.mu.Unlock()
 
 	var events []TuningEvent
+	skipped := make([]string, 0, 5)
+	// 记一次「本规则为什么没跑」—— 让调用方能解释「为什么没调参」，
+	// 而不是只回 no_changes_needed（后者无法区分「健康」与「数据不可用」）。
+	note := func(rule, why string) {
+		skipped = append(skipped, rule+": "+why)
+	}
 
-	// 1. 召回命中率 < 0.5 → 提高 diversity（扩大搜索范围）
-	if hitRate, ok := metrics["recall_hit_rate"]; ok && hitRate < 0.5 {
+	// 1. 召回质量低 → 提高 diversity（扩大搜索范围）。
+	// 2026-10-17：原信号 recall_hit_rate 已退役（恒=1、从不及 0.5 ⇒ 本安全阀长期死寂）。
+	// 改用 recall_usefulness_rate（返回且被使用）作为召回质量的实时信号：< 0.5 ⇒ 召回太窄、扩大范围。
+	hitRate, hitOK := ratioFor(metrics, "recall_usefulness_rate")
+	if !hitOK {
+		note("rule1_usefulness", "指标不可用（无数据/退化/样本<20）")
+	}
+	if hitOK && hitRate < 0.5 && !tc.canMove("diversity") {
+		note("rule1_usefulness", "冷却中（距上次调整 <"+tuneCooldown.String()+"）")
+	}
+	if hitOK && hitRate < 0.5 && tc.canMove("diversity") {
 		newDiversity := tc.Diversity + 0.1
 		if newDiversity > 1.0 {
 			newDiversity = 1.0
 		}
-		events = append(events, recordTuning(tc, "diversity", tc.Diversity, newDiversity, "recall_hit_rate < 0.5"))
+		events = append(events, recordTuning(tc, "diversity", tc.Diversity, newDiversity, "recall_usefulness_rate < 0.5"))
 		tc.Diversity = newDiversity
+		tc.markMove("diversity")
 	}
 
 	// 2. 召回有用率 > 0.9 → 降低 diversity（更精准）
-	if usefulRate, ok := metrics["recall_usefulness_rate"]; ok && usefulRate > 0.9 {
+	usefulRate, usefulOK := ratioFor(metrics, "recall_usefulness_rate")
+	if !usefulOK {
+		note("rule2_usefulness", "指标不可用（无数据/退化/样本<30）")
+	}
+	if usefulOK && usefulRate > 0.9 && !tc.canMove("diversity") {
+		note("rule2_usefulness", "冷却中（距上次调整 <"+tuneCooldown.String()+"）")
+	}
+	if usefulOK && usefulRate > 0.9 && tc.canMove("diversity") {
 		newDiversity := tc.Diversity - 0.05
 		if newDiversity < 0.1 {
 			newDiversity = 0.1
 		}
 		events = append(events, recordTuning(tc, "diversity", tc.Diversity, newDiversity, "recall_usefulness_rate > 0.9"))
 		tc.Diversity = newDiversity
+		tc.markMove("diversity")
 	}
 
-	// 3. 废弃率 > 5/day → 提高衰减率（加快清理）
-	if deprecated, ok := metrics["deprecated_per_day"]; ok && deprecated > 5 {
-		newDecay := tc.DecayRate * 1.2
-		if newDecay > 0.05 {
-			newDecay = 0.05
-		}
-		events = append(events, recordTuning(tc, "decay_rate", tc.DecayRate, newDecay, "deprecated_per_day > 5"))
-		tc.DecayRate = newDecay
-	}
+	// 3.（2026-10-14 移除）废弃率 → 提高衰减率。
+	// 移除理由：这个方向危险。governance.NewForgetter 上方注释记载：
+	// 「2026-09-07 方案A: 0.015→0.03 曾致误删(36天库龄下 0-recall 长记忆全被清, 405字CNB经验被删)」。
+	// 自动调参朝这个方向推 = 把那条事故路重新铺上；而收益只是「加快清理」。
+	// 若要恢复，必须先有「误删可回滚 + 删除前抽样人工确认」的保护，不能只靠阈值。
 
 	// 4. 缺口闭环率 < 0.5 → 降低 gap 阈值
-	if gapRate, ok := metrics["gap_closure_rate"]; ok && gapRate < 0.5 {
+	gapRate, gapOK := ratioFor(metrics, "gap_closure_rate")
+	if !gapOK {
+		note("rule4_gap_threshold", "指标不可用（无数据/退化/样本<30）")
+	}
+	if gapOK && gapRate < 0.5 && !tc.canMove("gap_threshold") {
+		note("rule4_gap_threshold", "冷却中（距上次调整 <"+tuneCooldown.String()+"）")
+	}
+	if gapOK && gapRate < 0.5 && tc.canMove("gap_threshold") {
 		newThreshold := tc.GapThreshold - 1
 		if newThreshold < 2 {
 			newThreshold = 2
 		}
 		events = append(events, recordTuningInt(tc, "gap_threshold", tc.GapThreshold, newThreshold, "gap_closure_rate < 0.5"))
 		tc.GapThreshold = newThreshold
+		tc.markMove("gap_threshold")
 	}
 
 	// 5. 蒸馏损失 > 0.3 → 更频繁蒸馏
-	if loss, ok := metrics["avg_distill_loss"]; ok && loss > 0.3 {
-		newInterval := tc.DistillInterval - 1
-		if newInterval < 1 {
-			newInterval = 1
-		}
-		events = append(events, recordTuningInt(tc, "distill_interval", tc.DistillInterval, newInterval, "avg_distill_loss > 0.3"))
-		tc.DistillInterval = newInterval
-	}
+	// 5.（2026-10-14 移除）蒸馏损失 → 缩短蒸馏间隔。
+	// 移除理由：**没有正确的消费方**。distill/engine.go:265 的 batchTimeout（5min）是
+	// 「队列冻结兜底 ticker」（2026-10-06 修的「生产者停止即永久冻结」安全网），
+	// **不是蒸馏节奏**。调安全网的周期治不了蒸馏质量 —— 接到错的地方比不接更坏。
+	// 若将来真有「蒸馏节奏」消费者，再恢复本规则并接对目标。
 
 	tc.history = append(tc.history, events...)
+	tc.lastSkipped = skipped
+	if len(events) > 0 {
+		tc.pushToConsumers() // 改了就必须接线，否则等于没改
+		tc.persist()         // 不落盘则重启即丢
+	}
 	return events
 }
 
+// init 启动时载入持久化参数并接线（没有它：重启回默认 + 调参不生效）。
+func init() { Tuner.load() }
+
+// LastSkipped 返回上一次调参里被跳过的规则及原因（只读快照）。
+func (tc *TuningConfig) LastSkipped() []string {
+	tc.mu.RLock()
+	defer tc.mu.RUnlock()
+	out := make([]string, len(tc.lastSkipped))
+	copy(out, tc.lastSkipped)
+	return out
+}
+
 func recordTuning(tc *TuningConfig, param string, oldVal, newVal float64, reason string) TuningEvent {
 	return TuningEvent{
 		Parameter: param,
@@ -135,7 +329,8 @@ func (tc *TuningConfig) StatusHandler(w http.ResponseWriter, r *http.Request) {
 			"distill_interval":  tc.DistillInterval,
 			"consolidate_after": tc.ConsolidateAfter,
 		},
-		"history": tc.history,
+		"history":      tc.history,
+		"last_skipped": tc.lastSkipped,
 	})
 }
 
@@ -145,7 +340,11 @@ func (tc *TuningConfig) RunHandler(w http.ResponseWriter, r *http.Request) {
 	events := tc.AutoTune(metrics)
 
 	if len(events) == 0 {
-		respond(w, 200, map[string]string{"status": "no_changes_needed"})
+		respond(w, 200, map[string]interface{}{
+			"status": "no_changes_needed",
+			// 区分「健康」与「数据不可用」：后者要修数据源，而不是当作一切正常。
+			"skipped": tc.lastSkipped,
+		})
 	} else {
 		respond(w, 200, map[string]interface{}{
 			"status": "tuned",
@@ -157,25 +356,34 @@ func (tc *TuningConfig) RunHandler(w http.ResponseWriter, r *http.Request) {
 
 // POST /api/v1/tuning/analytics — 运行因果分析
 func (tc *TuningConfig) AnalyticsHandler(w http.ResponseWriter, r *http.Request) {
-	// 基于指标间的关系给出建议
 	m := selfoptimize.Dash.Metrics()
-
-	suggestions := []string{}
-	if m["recall_hit_rate"] < 0.5 && m["avg_distill_loss"] > 0.3 {
-		suggestions = append(suggestions, "蒸馏质量低导致召回率低 — 建议检查 Embedding 模型")
-	}
-	if m["deprecated_per_day"] > 10 {
-		suggestions = append(suggestions, "废弃率过高 — 正在流失大量记忆，检查冲突检测阈值")
-	}
-	if m["gap_closure_rate"] < 0.3 {
-		suggestions = append(suggestions, "缺口闭环率过低 — Agent 未回应缺口学习请求")
-	}
-	if m["auto_resolve_rate"] < 0.5 {
-		suggestions = append(suggestions, "自动裁决率低 — 规则覆盖不全，考虑扩展策略")
-	}
-
+	suggestions := AnalyticsSuggestions(m)
 	respond(w, 200, map[string]interface{}{
 		"metrics":     m,
 		"suggestions": suggestions,
 	})
 }
+
+// AnalyticsSuggestions 依据指标间关系给出调优建议（纯函数，便于单测反例：N/A 不产生建议）。
+//
+// ⚠️ 每条建议的比率比较都必须先判 N/A（ratioFor 已封装：-1 哨兵 / valid=0 / 样本不足 ⇒ 不可用）。
+// 若直接 `m[key] < 阈值`，-1 会恒真 ⇒ 无数据也误报建议（本卡 t_c70f385c 事故面）。
+// deprecated_per_day 是计数（恒 ≥0，非比率），不走 ratioFor。
+func AnalyticsSuggestions(m map[string]float64) []string {
+	suggestions := []string{}
+	if hr, ok := ratioFor(m, "recall_usefulness_rate"); ok && hr < 0.5 {
+		if dl, ok2 := ratioFor(m, "avg_distill_loss"); ok2 && dl > 0.3 {
+			suggestions = append(suggestions, "蒸馏质量低导致召回率低 — 建议检查 Embedding 模型")
+		}
+	}
+	if m["deprecated_per_day"] > 10 {
+		suggestions = append(suggestions, "废弃率过高 — 正在流失大量记忆，检查冲突检测阈值")
+	}
+	if g, ok := ratioFor(m, "gap_closure_rate"); ok && g < 0.3 {
+		suggestions = append(suggestions, "缺口闭环率过低 — Agent 未回应缺口学习请求")
+	}
+	if a, ok := ratioFor(m, "auto_resolve_rate"); ok && a < 0.5 {
+		suggestions = append(suggestions, "自动裁决率低 — 规则覆盖不全，考虑扩展策略")
+	}
+	return suggestions
+}

go/internal/api/routes/tuning_analytics_test.go

@@ -0,0 +1,38 @@
+// AnalyticsSuggestions N/A 哨兵护栏 —— 反例 + 正向对照。
+//
+// 本卡 t_c70f385c：AnalyticsHandler 的建议 builder 原为裸 `m[key] < 阈值`，
+// -1 哨兵（样本不足/越界被拦）会让 `-1 < 0.5` 恒真 ⇒ 无数据也误报建议。
+// 现统一走 ratioFor（已封装 N/A + valid=0 + 样本不足判定）。
+package routes
+
+import "testing"
+
+// TestAnalyticsSuggestions_NA_NoFiring 反例：全 N/A 哨兵（-1）不产生任何建议（避免误报）。
+func TestAnalyticsSuggestions_NA_NoFiring(t *testing.T) {
+	na := map[string]float64{
+		"recall_usefulness_rate":  -1, "recall_usefulness_samples": 0, "recall_usefulness_valid": 0,
+		"avg_distill_loss":        -1, "avg_distill_loss_samples": 0, "avg_distill_loss_valid": 0,
+		"gap_closure_rate":        -1, "gap_closure_samples": 0, "gap_closure_valid": 0,
+		"auto_resolve_rate":       -1, "auto_resolve_samples": 0, "auto_resolve_valid": 0,
+		"deprecated_per_day":      0,
+	}
+	if s := AnalyticsSuggestions(na); len(s) != 0 {
+		t.Fatalf("全 N/A 不应产生任何建议，实际 %d 条: %v", len(s), s)
+	}
+}
+
+// TestAnalyticsSuggestions_ValidLow 正向对照：有效且低于阈值时产生建议（护栏没把功能焊死）。
+func TestAnalyticsSuggestions_ValidLow(t *testing.T) {
+	m := map[string]float64{
+		"recall_usefulness_rate":  0.2, "recall_usefulness_samples": 100, "recall_usefulness_valid": 1,
+		"gap_closure_rate":        0.1, "gap_closure_samples": 100, "gap_closure_valid": 1,
+		"auto_resolve_rate":       0.2, "auto_resolve_samples": 100, "auto_resolve_valid": 1,
+		"avg_distill_loss":        0.1, "avg_distill_loss_samples": 100, "avg_distill_loss_valid": 1,
+		"deprecated_per_day":      0,
+	}
+	s := AnalyticsSuggestions(m)
+	// 至少触发「gap 闭环率过低」+「自动裁决率低」两条
+	if len(s) < 2 {
+		t.Fatalf("有效且低于阈值应产生 ≥2 条建议，实际 %d 条: %v", len(s), s)
+	}
+}

go/internal/api/routes/tuning_guard_test.go

@@ -0,0 +1,125 @@
+package routes
+
+import (
+	"testing"
+	"time"
+
+	"github.com/xiaoxue/memoryweave/internal/selfoptimize"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// TestAutoTuneGuard_SentinelAndSamples 双向验证护栏：
+// ① 无数据（-1 哨兵）不许改参数 —— 这是事故面：gap_closure_rate=-1 会让
+//
+//	「降低 gap 阈值」每轮恒真。
+//
+// ② 样本不足（<30）不许改参数 —— 少样本比率噪声极大。
+// ③ 数据健康时仍然要能调 —— 否则等于把功能焊死。
+func TestAutoTuneGuard_SentinelAndSamples(t *testing.T) {
+	// ① 哨兵值：全 -1 ⇒ 一个事件都不许有
+	tc := &TuningConfig{Diversity: 0.5, DecayRate: 0.015, GapThreshold: 3, DistillInterval: 5, ConsolidateAfter: 50}
+	na := map[string]float64{
+		"gap_closure_rate":        -1,
+		"gap_closure_samples":     0,
+		"gap_closure_valid":       0,
+		"recall_usefulness_rate":  -1,
+		"avg_distill_loss":        -1,
+		"recall_usefulness_valid": 0,
+		"avg_distill_loss_valid":  0,
+	}
+	if ev := tc.AutoTune(na); len(ev) != 0 {
+		t.Fatalf("哨兵值(-1)不应触发任何调参，实际触发 %d 条: %+v", len(ev), ev)
+	}
+	if tc.GapThreshold != 3 || tc.Diversity != 0.5 {
+		t.Fatalf("参数被改动：GapThreshold=%d Diversity=%v", tc.GapThreshold, tc.Diversity)
+	}
+	if len(tc.LastSkipped()) == 0 {
+		t.Fatal("应记录跳过原因（否则无法区分『健康』与『数据不可用』）")
+	}
+
+	// ② 样本不足：比率本身达标（>0.9）但只有 9 个样本 ⇒ 不许动
+	tc2 := &TuningConfig{Diversity: 0.5, DecayRate: 0.015, GapThreshold: 3, DistillInterval: 5, ConsolidateAfter: 50}
+	low := map[string]float64{
+		"recall_usefulness_rate":    0.95,
+		"recall_usefulness_samples": 9,
+		"recall_usefulness_valid":   1,
+	}
+	if ev := tc2.AutoTune(low); len(ev) != 0 {
+		t.Fatalf("样本 9<30 不应触发调参，实际 %d 条: %+v", len(ev), ev)
+	}
+
+	// ③ 数据健康 ⇒ 必须能调（证明护栏没把功能焊死）
+	tc3 := &TuningConfig{Diversity: 0.5, DecayRate: 0.015, GapThreshold: 3, DistillInterval: 5, ConsolidateAfter: 50}
+	good := map[string]float64{
+		"recall_usefulness_rate":    0.95,
+		"recall_usefulness_samples": 100,
+		"recall_usefulness_valid":   1,
+	}
+	ev := tc3.AutoTune(good)
+	if len(ev) != 1 {
+		t.Fatalf("健康数据应触发 1 条调参，实际 %d 条", len(ev))
+	}
+	if tc3.Diversity >= 0.5 {
+		t.Fatalf("Diversity 应被降低，实际 %v", tc3.Diversity)
+	}
+}
+
+// TestAutoTuneCooldown 冷却：同一参数在冷却窗口内不许被重复推进（治棘轮）。
+// 事故面（实测）：AutoTune 不幂等 —— 连跑两次 decay 0.015→0.018→0.0216。
+func TestAutoTuneCooldown(t *testing.T) {
+	tc := &TuningConfig{Diversity: 0.5, DecayRate: 0.015, GapThreshold: 3, DistillInterval: 5, ConsolidateAfter: 50}
+	m := map[string]float64{
+		"recall_usefulness_rate":    0.95,
+		"recall_usefulness_samples": 100,
+		"recall_usefulness_valid":   1,
+	}
+	if ev := tc.AutoTune(m); len(ev) != 1 {
+		t.Fatalf("第一次应调 1 条，实际 %d", len(ev))
+	}
+	first := tc.Diversity
+	if ev := tc.AutoTune(m); len(ev) != 0 {
+		t.Fatalf("冷却窗口内不应再调，实际 %d 条: %+v", len(ev), ev)
+	}
+	if tc.Diversity != first {
+		t.Fatalf("冷却期内参数被改动：%v → %v（棘轮复发）", first, tc.Diversity)
+	}
+}
+
+// TestAutoTunePushesToConsumers 推送：改完参数必须真的到消费方手里，
+// 否则就是「调参改了没人读的数字」（2026-10-14 实测的原状）。
+func TestAutoTunePushesToConsumers(t *testing.T) {
+	orig := storage.DefaultMMRDiversity()
+	defer storage.SetDefaultMMRDiversity(orig)
+	tc := &TuningConfig{Diversity: 0.5, GapThreshold: 3}
+	m := map[string]float64{
+		"gap_closure_rate":    0.1,
+		"gap_closure_samples": 100,
+		"gap_closure_valid":   1,
+	}
+	before := selfoptimize.GapThreshold()
+	ev := tc.AutoTune(m)
+	if len(ev) != 1 {
+		t.Fatalf("应触发 1 条（gap 阈值），实际 %d", len(ev))
+	}
+	if tc.GapThreshold != before-1 {
+		t.Fatalf("期望阈值 %d→%d，实际 %d", before, before-1, tc.GapThreshold)
+	}
+	if got := selfoptimize.GapThreshold(); got != tc.GapThreshold {
+		t.Fatalf("没推到消费方：selfoptimize.GapThreshold()=%d，Tuner=%d", got, tc.GapThreshold)
+	}
+}
+
+// TestTuningPersistence 持久化：换实例后能从盘上读回（否则重启即回默认）。
+func TestTuningPersistence(t *testing.T) {
+	dir := t.TempDir()
+	t.Setenv("ZHIYI_TUNING_STATE", dir+"/tuning-state.json")
+
+	a := &TuningConfig{Diversity: 0.42, GapThreshold: 4, persistEnabled: true, lastMoveAt: map[string]time.Time{}}
+	a.persist()
+
+	b := &TuningConfig{Diversity: 0.3, GapThreshold: 3, lastMoveAt: map[string]time.Time{}}
+	b.load()
+	if b.Diversity != 0.42 || b.GapThreshold != 4 {
+		t.Fatalf("未读回持久化参数：diversity=%v gap=%v", b.Diversity, b.GapThreshold)
+	}
+}

go/internal/api/routes/ws.go

@@ -15,9 +15,9 @@ var upgrader = websocket.Upgrader{
 
 // WSManager 管理所有 WebSocket 连接，支持按 AgentID 定向推送
 type WSManager struct {
-	mu       sync.RWMutex
-	clients  map[string]map[*websocket.Conn]bool        // agentID → 连接集合
-	connMu   sync.Map // per-conn mutex: *websocket.Conn → *sync.Mutex
+	mu      sync.RWMutex
+	clients map[string]map[*websocket.Conn]bool // agentID → 连接集合
+	connMu  sync.Map                            // per-conn mutex: *websocket.Conn → *sync.Mutex
 }
 
 // 全局 WebSocket 管理器

go/internal/api/routes/zz_commit_conflict_gate_test.go

@@ -0,0 +1,286 @@
+// 织忆 MemoryWeave — Commit 冲突门控端到端回归 (2026-09-18 t_a54aae79 #14)
+//
+// 锁定 commit → 冲突检测 → AutoDistillTrigger 门控行为:
+//
+//	A. 真矛盾(否定差异+高重叠) 且无 resolve_conflict → 仍跳过蒸馏(保护不松)
+//	B. 真矛盾 + resolve_conflict=true                → 触发蒸馏(覆盖开关真存在)
+//	C. 单字"别"误报句对(旧逻辑 0.42>0.3 误杀)        → 新逻辑放行, 触发蒸馏
+//	D. 无矛盾                                          → 触发蒸馏(基线)
+package routes
+
+import (
+	"bytes"
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"net/http/httptest"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/xiaoxue/memoryweave/internal/distill"
+	"github.com/xiaoxue/memoryweave/internal/governance"
+	"github.com/xiaoxue/memoryweave/internal/models"
+	"github.com/xiaoxue/memoryweave/internal/storage"
+)
+
+// ─── mock: storage.LanceDB ───────────────────────────────────
+
+type mockLance struct {
+	memories []models.MemoryRecord
+}
+
+// Scan mock：返回按 namespace 过滤后的记忆（测试列表/导出路径用）。
+func (m *mockLance) Scan(limit int, ns string) ([]models.MemoryRecord, error) {
+	if limit <= 0 {
+		limit = 100
+	}
+	var out []models.MemoryRecord
+	for _, r := range m.memories {
+		if ns != "" && r.Namespace != ns {
+			continue
+		}
+		out = append(out, r)
+		if len(out) >= limit {
+			break
+		}
+	}
+	return out, nil
+}
+
+func (m *mockLance) Search(table string, vector []float32, topK int, ns string) ([]models.MemoryRecord, error) {
+	if table != "memories" {
+		return nil, nil
+	}
+	if topK < len(m.memories) {
+		return m.memories[:topK], nil
+	}
+	return m.memories, nil
+}
+func (m *mockLance) InsertEpisode(agentID, namespace, content, category string) (string, error) {
+	return fmt.Sprintf("ep_test_%d", time.Now().UnixNano()), nil
+}
+func (m *mockLance) GetTopByQuality(agentID string, limit int) ([]models.MemoryRecord, error) {
+	return nil, nil
+}
+func (m *mockLance) Stats() (map[string]interface{}, error) { return map[string]interface{}{}, nil }
+func (m *mockLance) SoftDelete(id, reason string) error     { return nil }
+func (m *mockLance) PurgeInvalid() (int64, error)           { return 0, nil }
+func (m *mockLance) SoftDeleteBatch(table string, ids []string, updatedAt string) (int64, error) {
+	return 0, nil
+}
+func (m *mockLance) GetVersionHistory(id string) ([]map[string]interface{}, error) { return nil, nil }
+func (m *mockLance) GetCandidatesForForgetting() ([]map[string]interface{}, error) { return nil, nil }
+func (m *mockLance) GetSkillCandidates(minRecalls, limit int) ([]models.MemoryRecord, error) {
+	return nil, nil
+}
+func (m *mockLance) Backup(path string) error                                { return nil }
+func (m *mockLance) GetAuditLog(limit int) ([]map[string]interface{}, error) { return nil, nil }
+func (m *mockLance) IncrementUseful(id string) error                         { return nil }
+func (m *mockLance) IncrementNotUseful(id string) error                      { return nil }
+
+// GA-P0-1 (t_718d8707)：拆分字段
+func (m *mockLance) IncrementHumanNotUseful(id string) error                 { return nil }
+func (m *mockLance) AppendHitEvent(memoryID, outcome, reason string) error   { return nil }
+func (m *mockLance) CountHitEvents(memoryID, outcome string, since time.Time) (int, error) { return 0, nil }
+func (m *mockLance) GetHitEvents(memoryID string, limit int) ([]models.HitEvent, error)    { return nil, nil }
+func (m *mockLance) IncrementDedupDownweight(id string) error                { return nil }
+func (m *mockLance) UpdateMemoryContent(id, newContent, source string) error { return nil }
+func (m *mockLance) InsertMemory(rec models.MemoryRecord) error              { return nil }
+func (m *mockLance) Insert(table string, record any) error                   { return nil }
+func (m *mockLance) Update(table, id string, fields map[string]any) error    { return nil }
+
+// ─── mock: governance.GraphStore (全 no-op) ──────────────────
+
+type mockGraph struct{}
+
+func (mockGraph) AddNode(id, name, nodeType, namespace string) error { return nil }
+func (mockGraph) AddEdge(id, source, target, relation, namespace string, weight float64) error {
+	return nil
+}
+func (mockGraph) Navigate(entity string, maxHops int, namespace string, relationFilter []string) ([]map[string]interface{}, error) {
+	return nil, nil
+}
+func (mockGraph) NavigateBiDir(source, target string, maxHops int, namespace string, relationFilter []string) ([]map[string]interface{}, error) {
+	return nil, nil
+}
+func (mockGraph) Query(entity, relation, namespace string) []map[string]interface{} {
+	return nil
+}
+func (mockGraph) AllEdges(namespace string) []governance.GraphEdge { return nil }
+func (mockGraph) SearchNodes(label string, namespace string) []map[string]interface{} {
+	return nil
+}
+func (mockGraph) ListNodesByType(nodeType string, namespace string) []map[string]interface{} {
+	return nil
+}
+func (mockGraph) ListNodes(namespace string) []map[string]interface{} { return nil }
+func (mockGraph) Stats() (int, int, float64)                          { return 0, 0, 0 }
+func (mockGraph) Prune(minWeight float64)                             {}
+func (mockGraph) ExpandFromResults(results []models.RecallResult, namespace string, maxHops int) []models.RecallResult {
+	return results
+}
+func (mockGraph) ExpandWithSummary(results []models.RecallResult, namespace string, maxHops int) models.GraphBFSResult {
+	return models.GraphBFSResult{}
+}
+func (mockGraph) PageRank(damping float64, iterations int) map[string]float64 { return nil }
+func (mockGraph) EvidenceCount(entity string) int                             { return 0 }
+func (mockGraph) GetEntityDegree(entity string) int                           { return 0 }
+func (mockGraph) GetGraph(namespace string, limit int) ([]map[string]interface{}, []map[string]interface{}) {
+	return nil, nil
+}
+func (mockGraph) CleanupNoiseNodes(dryRun bool) (int, []string, error) { return 0, nil, nil }
+func (mockGraph) FallbackTextSearch(query, namespace string, limit int) []map[string]interface{} {
+	return nil
+}
+func (mockGraph) AddEdgeFeedback(edgeID string, helpful bool) error { return nil }
+func (mockGraph) IncrementEdgeRetrieval(edgeID string) error        { return nil }
+func (mockGraph) UpdateEdgeTrustScores() error                      { return nil }
+
+// ─── mock: embedding 服务 (OpenAI 兼容, 每文本一个确定性独热向量) ────
+
+func newMockEmbedServer() *httptest.Server {
+	return httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		var reqBody struct {
+			Input []string `json:"input"`
+		}
+		json.NewDecoder(r.Body).Decode(&reqBody)
+		data := make([]map[string]any, 0, len(reqBody.Input))
+		for _, text := range reqBody.Input {
+			v := make([]float64, 1024)
+			// 确定性散列: 每个 rune 决定一个维度
+			sum := 0
+			for i, r := range text {
+				dim := (int(r)*31 + i*7) % 1024
+				v[dim] += 1
+				sum += int(r)
+			}
+			// 保证不同文本向量不同 (至少一个维度不重合的概率极高; 追加 sum 维度兜底)
+			v[sum%1024] += 2
+			data = append(data, map[string]any{"embedding": v})
+		}
+		w.Header().Set("Content-Type", "application/json")
+		json.NewEncoder(w).Encode(map[string]any{"data": data})
+	}))
+}
+
+// ─── mock: LLM 服务 (记录调用次数, 返回合法 5D JSON) ──────────────
+
+type mockLLM struct {
+	server *httptest.Server
+	calls  int32
+}
+
+func newMockLLM() *mockLLM {
+	m := &mockLLM{}
+	m.server = httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		// 2026-09-22 修：槽位探测（probeLLMSlots，GET /props）不是"蒸馏 LLM 调用"，
+		// 不能计入 calls —— 否则 A/B 批次断言（期望 base+1）会被 flush 前后的
+		// 两次探测撑成 base+3。根因：5e2cc63 引入并发槽位探测后，本 mock 未按 path 区分。
+		if r.URL.Path == "/props" {
+			w.Header().Set("Content-Type", "application/json")
+			json.NewEncoder(w).Encode(map[string]any{"total_slots": 1})
+			return
+		}
+		atomic.AddInt32(&m.calls, 1)
+		w.Header().Set("Content-Type", "application/json")
+		json.NewEncoder(w).Encode(map[string]any{
+			"choices": []map[string]any{{
+				"message": map[string]any{
+					"content": `{"facts":["测试事实：mock LLM 已收到蒸馏请求并返回一条合法事实。"],"entities":["测试"],"is":0.9,"su":0.8,"pa":0.8,"vd":0.8,"ru":0.8}`,
+				},
+			}},
+		})
+	}))
+	return m
+}
+
+func (m *mockLLM) count() int { return int(atomic.LoadInt32(&m.calls)) }
+
+// ─── 测试主体 ─────────────────────────────────────────────────
+
+func commitWith(t *testing.T, api *API, content string, resolve bool) map[string]interface{} {
+	t.Helper()
+	body, _ := json.Marshal(map[string]any{
+		"agent_id":         "test-agent",
+		"namespace":        "shared",
+		"category":         "general",
+		"content":          content,
+		"resolve_conflict": resolve,
+	})
+	req := httptest.NewRequest("POST", "/api/v1/commit", bytes.NewReader(body))
+	req.Header.Set("Content-Type", "application/json")
+	w := httptest.NewRecorder()
+	api.Commit(w, req)
+	if w.Code != 201 {
+		t.Fatalf("commit expected 201, got %d: %s", w.Code, w.Body.String())
+	}
+	var resp map[string]interface{}
+	json.Unmarshal(w.Body.Bytes(), &resp)
+	return resp
+}
+
+func waitLLMCalls(t *testing.T, m *mockLLM, min int) {
+	t.Helper()
+	deadline := time.Now().Add(15 * time.Second)
+	for time.Now().Before(deadline) {
+		if m.count() >= min {
+			return
+		}
+		time.Sleep(200 * time.Millisecond)
+	}
+	t.Fatalf("expected >= %d LLM distill calls, got %d (AutoDistillTrigger 未触发?)", min, m.count())
+}
+
+func TestCommitConflictGate_E2E(t *testing.T) {
+	emb := newMockEmbedServer()
+	defer emb.Close()
+	llm := newMockLLM()
+	defer llm.server.Close()
+
+	// 已有记忆 E: "用户昨天买了苹果"
+	ldb := &mockLance{memories: []models.MemoryRecord{
+		{ID: "mem_e", Content: "用户昨天买了苹果", Category: "distilled"},
+	}}
+	api := NewAPI(ldb, storage.NewEmbedder(emb.URL), nil, governance.NewConflictDetector(), mockGraph{})
+
+	oldRef := DistillEngineRef
+	DistillEngineRef = distill.NewEngine(llm.server.URL, "mock-model", "")
+	defer func() { DistillEngineRef = oldRef }()
+
+	// A. 真矛盾 (没买 vs 买了, 重叠 0.875) 无 flag → 跳过蒸馏
+	base := llm.count()
+	resp := commitWith(t, api, "用户昨天没买苹果", false)
+	if resp["distill_skipped"] != "conflict" {
+		t.Errorf("A: expected distill_skipped=conflict, got %v (resp=%s)", resp["distill_skipped"], resp)
+	}
+	time.Sleep(1500 * time.Millisecond) // 给(误)触发的蒸馏留机会, 应没调用
+	if llm.count() != base {
+		t.Errorf("A: contradiction without flag should NOT trigger distill, LLM calls %d→%d", base, llm.count())
+	}
+
+	// B. 同句 + resolve_conflict=true → 触发蒸馏
+	// (内容需 ≥10 rune 才过蒸馏硬规则 HardRulesPass; 用 10 字版, 与 E 重叠 8/10=0.8)
+	resp = commitWith(t, api, "用户昨天没有购买苹果", true)
+	if _, skipped := resp["distill_skipped"]; skipped {
+		t.Errorf("B: resolve_conflict=true must not skip, resp=%v", resp)
+	}
+	waitLLMCalls(t, llm, base+1)
+	time.Sleep(2 * time.Second) // 同一 flush 批次内多 item 会连续调用 LLM; 等批次排空
+	if got := llm.count(); got != base+1 {
+		t.Errorf("A/B: flush 批次应恰含 B 一条(A 被矛盾门拦下), 实际 LLM 调用 %d 次", got)
+	}
+
+	// C. 单字"别"误报句对 (旧逻辑: negdiff(别)+0.42>0.3 → 误杀; 新逻辑: 放行)
+	// 分支断言: 不 skip 即走了 Enqueue 路径 (可达 LLM 由 B 证明, 同一 Enqueue→flush 通道)
+	resp = commitWith(t, api, "用户昨天买了苹果和梨，别的事情先放一放", false)
+	if _, skipped := resp["distill_skipped"]; skipped {
+		t.Errorf("C: single-char 别 false positive must NOT skip, resp=%v", resp)
+	}
+
+	// D. 无矛盾基线 → 不 skip
+	resp = commitWith(t, api, "nginx 反向代理端口是 80", false)
+	if _, skipped := resp["distill_skipped"]; skipped {
+		t.Errorf("D: no-conflict baseline must NOT skip, resp=%v", resp)
+	}
+}

go/internal/api/routes/zz_ns_probe_test.go

@@ -0,0 +1,10 @@
+package routes
+
+import "testing"
+
+func TestClassifyNamespaceForMuchen(t *testing.T) {
+	for _, c := range []string{"牧尘在招商局工作", "牧尘不在招商局工作", "画像规则: 牧尘在招商局工作"} {
+		got := classifyNamespace("profiles", "a06", c)
+		t.Logf("content=%q → namespace=%q", c, got)
+	}
+}