@tea-agent/loop-agent 0.9.0 → 0.10.0-alpha.0

package/docs/init-surface.manifest.json

@@ -10,10 +10,19 @@
     "docs/README.md",
     "docs/init-surface.manifest.json",
     "docs/architecture/runtime-boundaries.md",
+    "docs/architecture/README.md",
+    "docs/architecture/system-overview.md",
+    "docs/architecture/dag-execution.md",
+    "docs/architecture/worker-and-feature.md",
+    "docs/architecture/facts-and-state.md",
+    "docs/architecture/evolution.md",
     "docs/skills/README.md",
     "docs/skills/vetted-skill-registry.md",
     "docs/templates/init-evolution-review.md",
     "docs/templates/production-readiness-checklist.md",
+    "docs/templates/frontend-task-requirement.md",
+    "docs/templates/frontend-task-constraints.md",
+    "docs/templates/frontend-design-contract.md",
     "docs/templates/worker-dogfood-setup.md",
     "docs/templates/worker-dogfood-evidence.md",
     "docs/templates/harness.schema.json",
@@ -32,8 +41,11 @@
     "scripts/check-task-pool-root.sh",
     "docs/templates/agent-dag.schema.json",
     "examples/example-dag.json",
+    "skills/agent-worker/SKILL.md",
+    "skills/agent-worker/references/agent-worker-operator.md",
     "skills/loop-agent/SKILL.md",
     "skills/loop-agent/references/command-reference.md",
+    "skills/loop-agent/references/docs-converge.md",
     "skills/ai-engineering-context/SKILL.md",
     "skills/verification-before-completion/SKILL.md",
     "skills/systematic-debugging/SKILL.md",
@@ -42,7 +54,17 @@
     "skills/test-driven-development/SKILL.md",
     "skills/code-review-core/SKILL.md",
     "skills/init-capability-evolution/SKILL.md",
-    "skills/webapp-testing/SKILL.md"
+    "skills/webapp-testing/SKILL.md",
+    "skills/frontend-implementation/SKILL.md",
+    "skills/frontend-implementation/references/node-contracts.md",
+    "skills/frontend-implementation/references/design-spec.md",
+    "skills/frontend-implementation/references/code-standards.md",
+    "skills/frontend-design-review/SKILL.md",
+    "skills/frontend-design-review/references/review-checklist.md",
+    "skills/frontend-review/SKILL.md",
+    "skills/frontend-review/references/review-findings.md",
+    "skills/frontend-verification/SKILL.md",
+    "skills/frontend-verification/references/verification-checklist.md"
   ],
   "initFullRequired": [
     "README.md",
@@ -56,6 +78,9 @@
     "docs/loop-agent-harness.md",
     "docs/templates/init-evolution-review.md",
     "docs/templates/production-readiness-checklist.md",
+    "docs/templates/frontend-task-requirement.md",
+    "docs/templates/frontend-task-constraints.md",
+    "docs/templates/frontend-design-contract.md",
     "docs/templates/worker-dogfood-setup.md",
     "docs/templates/worker-dogfood-evidence.md",
     "docs/templates/harness.schema.json",
@@ -70,6 +95,7 @@
     "docs/templates/product-line/task.yaml",
     "docs/templates/product-line/closeout.yaml",
     "docs/templates/product-line/links.md",
+    "scripts/check-skill-entry.sh",
     "scripts/check-product-line-docs.sh",
     "scripts/check-repo.sh",
     "scripts/ci-governance.sh",
@@ -78,8 +104,11 @@
     ".harness/prompts/analyze.md",
     ".harness/tasks",
     ".harness/dag-runs/active",
+    "skills/agent-worker/SKILL.md",
+    "skills/agent-worker/references/agent-worker-operator.md",
     "skills/loop-agent/SKILL.md",
     "skills/loop-agent/references/command-reference.md",
+    "skills/loop-agent/references/docs-converge.md",
     "skills/ai-engineering-context/SKILL.md",
     "skills/verification-before-completion/SKILL.md",
     "skills/systematic-debugging/SKILL.md",
@@ -89,8 +118,21 @@
     "skills/code-review-core/SKILL.md",
     "skills/init-capability-evolution/SKILL.md",
     "skills/webapp-testing/SKILL.md",
+    "skills/frontend-implementation/SKILL.md",
+    "skills/frontend-implementation/references/node-contracts.md",
+    "skills/frontend-implementation/references/design-spec.md",
+    "skills/frontend-implementation/references/code-standards.md",
+    "skills/frontend-design-review/SKILL.md",
+    "skills/frontend-design-review/references/review-checklist.md",
+    "skills/frontend-review/SKILL.md",
+    "skills/frontend-review/references/review-findings.md",
+    "skills/frontend-verification/SKILL.md",
+    "skills/frontend-verification/references/verification-checklist.md",
+    ".agents/skills/agent-worker/SKILL.md",
+    ".agents/skills/agent-worker/references/agent-worker-operator.md",
     ".agents/skills/loop-agent/SKILL.md",
     ".agents/skills/loop-agent/references/command-reference.md",
+    ".agents/skills/loop-agent/references/docs-converge.md",
     ".agents/skills/ai-engineering-context/SKILL.md",
     ".agents/skills/verification-before-completion/SKILL.md",
     ".agents/skills/systematic-debugging/SKILL.md",
@@ -113,6 +155,9 @@
     "docs/loop-agent-harness.md": "generated",
     "docs/templates/init-evolution-review.md": "copied",
     "docs/templates/production-readiness-checklist.md": "copied",
+    "docs/templates/frontend-task-requirement.md": "copied",
+    "docs/templates/frontend-task-constraints.md": "copied",
+    "docs/templates/frontend-design-contract.md": "copied",
     "docs/templates/worker-dogfood-setup.md": "copied",
     "docs/templates/worker-dogfood-evidence.md": "copied",
     "docs/templates/harness.schema.json": "copied",
@@ -127,6 +172,7 @@
     "docs/templates/product-line/task.yaml": "copied",
     "docs/templates/product-line/closeout.yaml": "copied",
     "docs/templates/product-line/links.md": "copied",
+    "scripts/check-skill-entry.sh": "generated",
     "scripts/check-product-line-docs.sh": "copied",
     "scripts/check-repo.sh": "generated",
     "scripts/ci-governance.sh": "generated",
@@ -136,8 +182,11 @@
     ".harness/tasks": "directory",
     ".harness/dag-runs/active": "directory",
     ".harness/init-surface.json": "state",
+    "skills/agent-worker/SKILL.md": "copied",
+    "skills/agent-worker/references/agent-worker-operator.md": "copied",
     "skills/loop-agent/SKILL.md": "copied",
     "skills/loop-agent/references/command-reference.md": "copied",
+    "skills/loop-agent/references/docs-converge.md": "copied",
     "skills/ai-engineering-context/SKILL.md": "copied",
     "skills/verification-before-completion/SKILL.md": "copied",
     "skills/systematic-debugging/SKILL.md": "copied",
@@ -147,8 +196,21 @@
     "skills/code-review-core/SKILL.md": "copied",
     "skills/init-capability-evolution/SKILL.md": "copied",
     "skills/webapp-testing/SKILL.md": "copied",
+    "skills/frontend-implementation/SKILL.md": "copied",
+    "skills/frontend-implementation/references/node-contracts.md": "copied",
+    "skills/frontend-implementation/references/design-spec.md": "copied",
+    "skills/frontend-implementation/references/code-standards.md": "copied",
+    "skills/frontend-design-review/SKILL.md": "copied",
+    "skills/frontend-design-review/references/review-checklist.md": "copied",
+    "skills/frontend-review/SKILL.md": "copied",
+    "skills/frontend-review/references/review-findings.md": "copied",
+    "skills/frontend-verification/SKILL.md": "copied",
+    "skills/frontend-verification/references/verification-checklist.md": "copied",
+    ".agents/skills/agent-worker/SKILL.md": "copied",
+    ".agents/skills/agent-worker/references/agent-worker-operator.md": "copied",
     ".agents/skills/loop-agent/SKILL.md": "copied",
     ".agents/skills/loop-agent/references/command-reference.md": "copied",
+    ".agents/skills/loop-agent/references/docs-converge.md": "copied",
     ".agents/skills/ai-engineering-context/SKILL.md": "copied",
     ".agents/skills/verification-before-completion/SKILL.md": "copied",
     ".agents/skills/systematic-debugging/SKILL.md": "copied",

package/docs/loop-agent-harness.md

@@ -10,21 +10,25 @@ loop-agent 提供结构化 agent 工作的本地 harness。
 - `.harness/cache/` — 本地 runtime 缓存
 - `.harness/live/` — 瞬态 live-session 文件
 - `.harness/task-pool/` — `agent-worker` 的 Task Pool state、batch artifacts、failure handoffs 与 Observe events；它是 `.harness/` 内的独立 Worker runtime root，默认不提交
+- `.harness/dag-runs/<lifecycle>/<run-id>/.runtime/skill-snapshot.json` — 该 run 实际使用的 resolved skill profile snapshot；由 state 中相对 ref 与原始 bytes SHA-256 锚定，随 lifecycle 目录整体迁移
 
 ## Skill 指令
 
 - `skills/loop-agent/` — loop-agent 仓库自有的 skill 指令与参考资料
+- `skills/agent-worker/` — Feature Packet、TaskSpec、Task Pool、controller pinning、自举 candidate canary 与 Worker recovery 的可选 outer-loop operator skill；不是 DAG node skill
 - `skills/<skill-name>/` — DAG 模板引用的 skill 本地副本，含 supervised implementation 角色
 - 历史 `skill/` 不再是必需仓库路径；runtime 查找仅为 `loop-agent` 保留兼容 fallback
 
 ## Package 资产
 
-npm 包包含运行与说明 harness 所需的静态能力资产：`bin/`、`dist/`、`skills/`、顶层 `docs/*.md`、`docs/templates/`、`examples/`、`harness.json`、`AGENTS.md`、`README.md`、`CHANGELOG.md`。
+npm 包包含运行与说明 harness 所需的静态能力资产：`bin/`、`dist/`、`skills/`（包括 `loop-agent` 与 `agent-worker`）、顶层 `docs/*.md`、`docs/templates/`、`examples/`、`harness.json`、`AGENTS.md`、`README.md`、`CHANGELOG.md`。
 
 生成或历史的 task facts 不是 package 资产。`docs/progress/`、`docs/reports/`、`docs/exec-plans/`、`docs/decisions/` 下现有文件属于目标仓库历史；包只携带这些目录的 README。
 
 从 npm 安装 `loop-agent` 并在其他项目使用时，DAG skill 指令按配置、用户或目标本地 skill 目录解析（若存在），再回退到包内 `skills/`。目标项目可添加本地 skills，但不必复制 loop-agent 源码历史或 `skills/` 目录即可使用默认 DAG 工作流。
 
+`scripts/self-host-canary.mjs` 是源码仓库维护脚本，不进入 npm package `files`；安装后的 package 不把它作为用户命令 surface。
+
 ## harness.json Schema
 
 `docs/templates/harness.schema.json` 是面向 IDE 的 `harness.json` JSON Schema。它用于补全字段、提示枚举值、解释 executor/model routing 配置，并尽早暴露拼写错误。
@@ -33,15 +37,73 @@ npm 包包含运行与说明 harness 所需的静态能力资产：`bin/`、`dis
 
 `loop-agent init --profile full --merge` 会把 schema 投影到目标项目，并在生成的 `harness.json` 写入与 `governanceRoot` 匹配的 `$schema`。旧初始化项目可通过 `loop-agent init check-update` 查看缺失项，并用 `loop-agent init update --apply-safe` 安全补齐 schema 文件和 `$schema` 引用；该安全更新只允许修改 `$schema`，不会重写用户已有 harness 配置。
 
+DAG run state 会记录 runner PID、hostname 和定期 heartbeat，并为节点记录最近活动时间。`loop-agent dag status` 和 `loop-agent dag doctor --run-id <run-id> --markdown` 会将同主机且 PID 已消失的运行识别为 `orphaned`，将进程仍在但 heartbeat 过期的运行识别为 `stale`，将 runner 存活但当前节点超过 5 分钟没有状态活动标记为 `node-quiet`；跨主机运行只标记为 `unknown-host`，不会据此判断远端进程死亡。Observe 使用同一组字段过滤活跃 DAG，并在详情中显示 lifecycle 与 liveness；`node-quiet` 仍视为活跃，只提示操作者检查 session events 和 executor logs。
+
+Observe 和 `dag status` 还会给出 `effectiveStatus`、`stateConsistent` 与 `recoveryEligibility`。主状态以 lifecycle + canonical status + liveness 的综合判定为准，原始 `state.status` 只作为诊断事实展示。遇到历史 `paused/` 目录仍记录 `status=running`、或 active runner 已明确 `orphaned` 时，先运行只读检查：
+
+```bash
+loop-agent dag reconcile-run --run-id <run-id>
+```
+
+确认不应继续后，必须显式选择收口含义并填写原因：
+
+```bash
+loop-agent dag reconcile-run --run-id <run-id> --action supersede --reason "implementation completed elsewhere"
+loop-agent dag reconcile-run --run-id <run-id> --action abandon --reason "operator cancelled obsolete work"
+```
+
+收口会保存 `reconciliation.json` 原始快照，将仍为 RUNNING 的节点标为 ERROR，并把 run 归档到 `completed/`；不会把未执行节点伪造成成功。同机 runner 仍存活、仅 `node-quiet`、heartbeat stale 但 PID 仍存在、跨主机或无法证明停止时都会拒绝收口。
+
+DAG 节点的模型由 `executorModels[executor][complexity]` 路由。新 run 会将解析后的 Pi/Cursor 模型写入节点 state；Observe 读取历史 run 时会从 `run.json` 补算缺失模型。Shell 和 Static 节点不使用模型，因此模型列保持 `-`。
+
+DAG agent 节点的输出语言由 `harness.json.workflowPolicy.dag.outputLanguage` 控制，支持 `zh-CN` 和 `en`。未配置时默认 `zh-CN`；`dag run-task` 会把配置写入生成的 DagSpec，手写 DagSpec 缺少该字段时 prompt envelope 同样默认中文。语言规则只约束说明、分析、报告、总结和文档正文，不翻译代码、命令、路径、标识符、JSON 字段、VERDICT 行及其他精确协议字面量。
+
+新 DAG run（包括 dry-run 与 init-only）会在首次 state write 前冻结所有初始 Pi task、dynamic map child template 与 loop body template 的 resolved skill profiles。profile id 包含 ordered skills、learned-pattern mode 和两个 instruction budgets；后续节点只从 snapshot 建 prompt。snapshot/hash/binding/profile 不一致时节点进入 `ERROR` / `skill-snapshot-integrity`，executor 不调用，也不回退实时 skill。旧 run 只有在 ref 与 artifact 都不存在时，才可在 resume 前创建 `legacy-resume-backfill`，且该 mode 只证明剩余节点被固定。
+
 ## Command Surface
 
 - `new-task`、`status`、`inspect`、`doctor`
-- `dag run-task`、`dag validate`、`run-dag`、`dag report`、`dag resume`
+- `dag run-task`、`dag validate`、`run-dag`、`dag report`、`dag resume`、`dag reconcile-run`
 - `delegate`、`harvest`、`promote-run`、`closeout`
 - `loop init`、`loop run`、`loop status`、`loop closeout`
 - `pi-prompt`、`cursor-prompt`
 - `docs audit`、`handoff check`、`spine audit`、`knowledge curate`
-- `agent-worker task validate|explain-profile|retry`、`agent-worker batch run-ready`、`agent-worker report morning`、`agent-worker observe serve|snapshot`
+- `agent-worker feature review|run|verify-final|delivery|closeout|approve-followup`、`agent-worker task validate|explain-profile|retry|draft-followup`、`agent-worker batch run-ready`、`agent-worker report morning|metrics`、`agent-worker observe serve|snapshot`
+
+## Feature Review
+
+`agent-worker feature review --feature-dir <feature-dir> --repo <target-repo>` 是只读的 Feature 级决策入口。它复用 Feature Packet validation、task graph 和 `.harness/task-pool/` 事实，派生 `Draft`、`Ready`、`Running`、`NeedsAction`、`AwaitingQA`、`Deliverable`、`Closed`，同时展示 required AC 覆盖、阻塞、证据和唯一主行动。
+
+默认输出按结果、原因、下一步、证据组织；`--json` 输出 schemaVersion 1 JSON。缺失或损坏的 Packet、state、JSONL 或 closeout 会暴露 validation error / projection warning，不会被误判为无数据、可交付或已关闭。该命令不写 Feature Packet、Task Pool 或 Git。
+
+`agent-worker feature run --feature-dir <feature-dir> --repo <target-repo>` 是薄编排入口：先执行 Feature validation 和一次目标仓库 preflight，再复用 `runReadyTasks()` 推进 Ready Task，最后刷新 morning report、`.harness/task-pool/observability/snapshot.json` 与 Feature review。`--dry-run` 只读取 Packet/Task Pool 并列出 Ready IDs、步骤和预期产物。默认无 Git 授权时单次最多执行一个 Ready 写任务；checkpoint 模式可按 `--limit` 串行推进，并在每个 Task 后建立 clean boundary。
+
+写入型 Worker 命令可传 `--expected-controller-version` 与 `--expected-controller-fingerprint`。`LoopAgentClient` 先把请求的 bin/PATH/npm shim 解析成冻结的绝对 launch spec，再校验 package name/version、binary SHA-256 和 portable package fingerprint；expected gate 在 materialize、Task Pool `Running` 或目标仓库写入前完成。预检后的命令使用该 launch spec，不会因 PATH、symlink 或安装内容变化而静默切换 controller。
+
+`agent-worker feature verify-final` 在 clean Delivery HEAD 上以已完成的 `qa-execute` TaskSpec 运行独立最终 DAG，跳过 promotion/task closeout，避免 repo write 或 HEAD 移动，并原子生成 canonical QA aggregate 与 HEAD-bound final verification。`agent-worker feature delivery` 从同一 Git transaction 和 Task Pool facts 生成带 hash 的 manifest、Acceptance Coverage 与稳定 `PR.md`，拒绝 dirty/wrong branch、历史或 trailer 漂移、缺失 run/QA/final verification 和 required AC 缺口。`agent-worker feature closeout` 默认只读预览 gates；显式 `--apply --owner` 才原子写回 Closeout，并在 stale facts、写入或 post-validation 失败时恢复原状态。
+
+Morning report 顶部和 Observe 首页直接消费共享 Feature read model，固定从 Status、Next Action、Why、Evidence 开始。`agent-worker report metrics --month YYYY-MM` 写入 `.harness/task-pool/artifacts/metrics/monthly-YYYY-MM.json|md`，每项指标保留分母、样本量、窗口和 missingData。
+
+显式 `--git-mode checkpoint` 会在 clean preflight 后创建或复用受 transaction record 管理的 `agent/<feature-id>` 本地分支。成功任务只有在 changed/untracked/sensitive 与 TaskSpec allowed/forbidden audit 通过后才 commit；失败任务的 patch、inventory、untracked content 与 boundary audit 写入 `.harness/task-pool/artifacts/features/<feature-id>/git/failures/`，随后 reset/clean 到最近 checkpoint 并再次验证 clean。默认模式不触碰 Git；`--keep-failed-diff` 只保留现场并停止，不允许继续下一个写任务。任何模式都不授权 stash、push、remote PR 或 merge。
+
+业务 Task 失败不会跳过报告与最终 review，但命令返回 `needs-action` 和非零退出码；`noReadyReason` 用于区分 Closed、Deliverable、AwaitingQA、NeedsAction、dependency blocked 与 empty。
+
+## Versioned Self-hosting Canary
+
+源码仓库使用下面的 repo-maintainer 命令验证隔离 candidate package 的确定性接棒：
+
+```bash
+npm run self-host:canary -- --deterministic --output <evidence.json>
+npm run self-host:canary -- --deterministic --tarball <candidate.tgz> --output <evidence.json>
+```
+
+canary 会记录 tarball SHA-256、slot containment、两个 bin identity、由维护脚本独立计算并与候选实现交叉核对的 package fingerprint、每个 required check、skill mirror hashes、Feature dry-run、static/shell DAG run/report 以及 PATH trap。所有子进程都有硬超时；它删除 `HARNESS_ALLOW_ACTIVE_DAG_RUNS`，并要求 trap 中没有 `loop-agent`、`agent-worker` 或 `pi` 调用。`--live` 当前明确失败；deterministic canary 不包含 Pi/model/provider/cost 证明。
+
+## Follow-up 决策与批准
+
+`agent-worker task draft-followup <task-id> --worker-run-id <id> --feature-dir <feature-dir> --repo <target-repo>` 从失败 run、TaskSpec 和 task graph 生成 `FU-<parent>-NN` 决策产物。ProductBug/TestBug/FlakyTest/DependencyFailure 生成可批准 TaskDraft；连续两次 EnvFailure 生成 `ENV-CHECK-*`，单次只建议 retry；SpecUnclear/ContractMismatch/RiskyChange/NeedsHuman/Unknown 生成人工行动卡。产物与 evidence/index 位于 `.harness/task-pool/artifacts/features/<feature-id>/follow-ups/`；生成过程受 Feature lock 和 dedupe key 保护，不写 Feature Packet、Ready state 或原失败事实。新 run 会 supersede 同 parent 的旧 unresolved 产物，但不会删除历史。
+
+人工批准只接受 TaskDraft；行动卡不能批准或进入 Ready。批准会校验 draft/evidence hash、repo/path containment、Feature baseline 和冲突，在 staging validation 后写 TaskSpec、graph、Ready、approval/index 与 event。rename、state、approval、index、event 五个故障点均有注入回滚测试；重复批准幂等复用首次结果。`--dry-run` 零写入，`--json` 失败保持 stdout JSON、stderr 无噪声和非零退出码。
 
 ## Worker Retry
 

package/docs/progress/README.md

@@ -1,3 +1,30 @@
 # 进度日志
 
 本目录存放应超越聊天上下文的简洁 handoff 日志。
+
+**索引职责**：本文件维护本目录**全量文件列表**。`docs/README.md` 只链到本 README。
+
+## 近期要点（导读）
+
+- [`2026-07-14-docs-living-docs-calibration.md`](2026-07-14-docs-living-docs-calibration.md) — 活文档校准（Dynamic Workflow / 六个月 / ADR）
+- [`2026-07-14-docs-governance-hygiene.md`](2026-07-14-docs-governance-hygiene.md) — 文档治理与双树/索引收敛
+- [`2026-07-14-website-docs-ia-and-converge.md`](2026-07-14-website-docs-ia-and-converge.md) — Website 文档 IA 交接
+- [`2026-07-13-dev-controller-001-implementation.md`](2026-07-13-dev-controller-001-implementation.md) — controller identity 实施
+- [`2026-07-13-dev-snapshot-001-implementation.md`](2026-07-13-dev-snapshot-001-implementation.md) — skill snapshot 实施
+- [`2026-07-13-dev-canary-001-implementation.md`](2026-07-13-dev-canary-001-implementation.md) — candidate canary 实施
+
+## 全量列表（新→旧）
+
+- [`2026-07-14-website-docs-ia-and-converge.md`](2026-07-14-website-docs-ia-and-converge.md)
+- [`2026-07-14-docs-living-docs-calibration.md`](2026-07-14-docs-living-docs-calibration.md)
+- [`2026-07-14-docs-governance-hygiene.md`](2026-07-14-docs-governance-hygiene.md)
+- [`2026-07-13-dev-snapshot-001-implementation.md`](2026-07-13-dev-snapshot-001-implementation.md)
+- [`2026-07-13-dev-controller-001-implementation.md`](2026-07-13-dev-controller-001-implementation.md)
+- [`2026-07-13-dev-canary-001-implementation.md`](2026-07-13-dev-canary-001-implementation.md)
+- [`2026-07-12-pi-only-agent-runtime.md`](2026-07-12-pi-only-agent-runtime.md)
+- [`2026-07-11-observe-dag-visualization.md`](2026-07-11-observe-dag-visualization.md)
+- [`2026-07-11-m2-feature-delivery-loop.md`](2026-07-11-m2-feature-delivery-loop.md)
+- [`2026-07-10-observe-ui-optimization.md`](2026-07-10-observe-ui-optimization.md)
+- [`2026-07-04-dynamic-workflow-dag-engine.md`](2026-07-04-dynamic-workflow-dag-engine.md)
+- [`2026-07-04-commander-cli-migration.md`](2026-07-04-commander-cli-migration.md)
+- [`2026-07-02-loop-agent-subject-restructure.md`](2026-07-02-loop-agent-subject-restructure.md)

package/docs/reports/README.md

@@ -1,12 +1,78 @@
 # 报告
 
-本目录存放验证、审计、benchmark 与 closeout 报告。
+本目录存放验证、审计、benchmark、init evolution 与 closeout 报告。
 
+**索引职责**：本文件维护本目录**全量文件列表**。`docs/README.md` 只链到本 README，不再精选单篇 report。
+
+查找约定：
+
+- 按日期前缀浏览下方列表（新→旧）
+- init surface 审查：文件名含 `init-evolution-review`
+- 能力基线滚动分析：[`2026-07-02-repository-analysis.md`](2026-07-02-repository-analysis.md)（长文，非单次任务证据）
+
+## 近期要点（导读，非权威全集）
+
+- [`current-capability-summary.md`](current-capability-summary.md) — **活**能力摘要（新进展写这里 / CHANGELOG / 新 report）
+- [`2026-07-14-docs-living-docs-calibration.md`](2026-07-14-docs-living-docs-calibration.md) — 活文档校准（Dynamic Workflow / 六个月 / analysis 冻结 / ADR）
+- [`2026-07-14-docs-governance-hygiene.md`](2026-07-14-docs-governance-hygiene.md) — 文档治理扫描与索引/双树收敛
+- [`2026-07-14-website-docs-ia-and-converge.md`](2026-07-14-website-docs-ia-and-converge.md) — Website IA 与 docs-converge 收口
+- [`2026-07-13-versioned-self-hosting-bootstrap.md`](2026-07-13-versioned-self-hosting-bootstrap.md) — 版本化自举证据
+- [`2026-07-12-m2-completion-audit.md`](2026-07-12-m2-completion-audit.md) — 第二月完成审计
+- [`2026-07-02-repository-analysis.md`](2026-07-02-repository-analysis.md) — **已冻结**仓库能力基线快照（勿再当 CHANGELOG）
+
+## 全量列表（新→旧）
+
+- [`dogfood-preflight-static-20260709090443.md`](dogfood-preflight-static-20260709090443.md)
+- [`current-capability-summary.md`](current-capability-summary.md)
+- [`2026-07-14-website-docs-ia-init-evolution-review.md`](2026-07-14-website-docs-ia-init-evolution-review.md)
+- [`2026-07-14-website-docs-ia-and-converge.md`](2026-07-14-website-docs-ia-and-converge.md)
+- [`2026-07-14-docs-living-docs-calibration.md`](2026-07-14-docs-living-docs-calibration.md)
+- [`2026-07-14-docs-governance-hygiene.md`](2026-07-14-docs-governance-hygiene.md)
+- [`2026-07-13-versioned-self-hosting-init-evolution-review.md`](2026-07-13-versioned-self-hosting-init-evolution-review.md)
+- [`2026-07-13-versioned-self-hosting-bootstrap.md`](2026-07-13-versioned-self-hosting-bootstrap.md)
+- [`2026-07-13-frontend-verification-command-safety-init-evolution-review.md`](2026-07-13-frontend-verification-command-safety-init-evolution-review.md)
+- [`2026-07-13-frontend-skill-contracts-init-evolution-review.md`](2026-07-13-frontend-skill-contracts-init-evolution-review.md)
 - [`2026-07-12-production-readiness-hardening.md`](2026-07-12-production-readiness-hardening.md)
-- [`2026-07-10-next-stage-dogfood-runs.md`](2026-07-10-next-stage-dogfood-runs.md) — 发布控制器下的 BE/FE/QA Worker 样本、retry 与 Observe 证据
-- [`2026-07-10-init-evolution-review.md`](2026-07-10-init-evolution-review.md) — Worker 初始化投影演进审查
-- [`2026-07-10-session-events-dogfood.md`](2026-07-10-session-events-dogfood.md) — Pi SDK DAG 新 run 产生非空 `session-events.jsonl` 与 Observe 过程时间线 API 证据
-- [`2026-07-11-command-performance-followups.md`](2026-07-11-command-performance-followups.md) — Observe 快照复用、输出缓冲和 reference index 遍历优化的验证记录
-- [`2026-07-11-observe-terminal-dag-kpi.md`](2026-07-11-observe-terminal-dag-kpi.md) — 终态 DAG 被误计为活跃数的修复记录
-- [`2026-07-12-observe-warm-console-redesign.md`](2026-07-12-observe-warm-console-redesign.md) — 暖白运行控制台视觉重构与桌面验证记录
-- [`2026-07-02-repository-analysis.md`](2026-07-02-repository-analysis.md) — 仓库能力基线分析（2026-07-10 再采样：0.5.0 + Observe UI 优化 UI-1～UI-9 / Unreleased）
+- [`2026-07-12-pi-only-agent-runtime-init-evolution-review.md`](2026-07-12-pi-only-agent-runtime-init-evolution-review.md)
+- [`2026-07-12-observe-warm-console-redesign.md`](2026-07-12-observe-warm-console-redesign.md)
+- [`2026-07-12-observe-dag-node-model.md`](2026-07-12-observe-dag-node-model.md)
+- [`2026-07-12-observe-dag-node-model-init-evolution-review.md`](2026-07-12-observe-dag-node-model-init-evolution-review.md)
+- [`2026-07-12-m2-completion-audit.md`](2026-07-12-m2-completion-audit.md)
+- [`2026-07-12-m2-08-init-evolution-review.md`](2026-07-12-m2-08-init-evolution-review.md)
+- [`2026-07-12-m2-08-dogfood-release-evidence.md`](2026-07-12-m2-08-dogfood-release-evidence.md)
+- [`2026-07-12-m2-07-init-evolution-review.md`](2026-07-12-m2-07-init-evolution-review.md)
+- [`2026-07-12-m2-06-init-evolution-review.md`](2026-07-12-m2-06-init-evolution-review.md)
+- [`2026-07-12-m2-05-init-evolution-review.md`](2026-07-12-m2-05-init-evolution-review.md)
+- [`2026-07-12-m2-04-init-evolution-review.md`](2026-07-12-m2-04-init-evolution-review.md)
+- [`2026-07-12-m2-03-init-evolution-review.md`](2026-07-12-m2-03-init-evolution-review.md)
+- [`2026-07-12-m2-02-init-evolution-review.md`](2026-07-12-m2-02-init-evolution-review.md)
+- [`2026-07-12-init-evolution-strict-report-range.md`](2026-07-12-init-evolution-strict-report-range.md)
+- [`2026-07-12-dag-state-reconciliation.md`](2026-07-12-dag-state-reconciliation.md)
+- [`2026-07-12-dag-state-reconciliation-init-evolution-review.md`](2026-07-12-dag-state-reconciliation-init-evolution-review.md)
+- [`2026-07-12-dag-output-language.md`](2026-07-12-dag-output-language.md)
+- [`2026-07-12-dag-output-language-init-evolution-review.md`](2026-07-12-dag-output-language-init-evolution-review.md)
+- [`2026-07-12-dag-liveness-recovery.md`](2026-07-12-dag-liveness-recovery.md)
+- [`2026-07-12-dag-liveness-init-evolution-review.md`](2026-07-12-dag-liveness-init-evolution-review.md)
+- [`2026-07-12-0.10.0-init-evolution-review.md`](2026-07-12-0.10.0-init-evolution-review.md)
+- [`2026-07-11-task-pool-hard-cutover-init-evolution-review.md`](2026-07-11-task-pool-hard-cutover-init-evolution-review.md)
+- [`2026-07-11-round3-worker-expansion.md`](2026-07-11-round3-worker-expansion.md)
+- [`2026-07-11-round2-interactive-ui-experiment.md`](2026-07-11-round2-interactive-ui-experiment.md)
+- [`2026-07-11-round2-init-evolution-review.md`](2026-07-11-round2-init-evolution-review.md)
+- [`2026-07-11-observe-terminal-dag-kpi.md`](2026-07-11-observe-terminal-dag-kpi.md)
+- [`2026-07-11-observe-dag-visualization.md`](2026-07-11-observe-dag-visualization.md)
+- [`2026-07-11-m2-01-init-evolution-review.md`](2026-07-11-m2-01-init-evolution-review.md)
+- [`2026-07-11-init-evolution-review.md`](2026-07-11-init-evolution-review.md)
+- [`2026-07-11-commit-1d0f3b7-review.md`](2026-07-11-commit-1d0f3b7-review.md)
+- [`2026-07-11-command-performance-followups.md`](2026-07-11-command-performance-followups.md)
+- [`2026-07-11-command-performance-audit.md`](2026-07-11-command-performance-audit.md)
+- [`2026-07-11-0.7.5-init-evolution-review.md`](2026-07-11-0.7.5-init-evolution-review.md)
+- [`2026-07-10-session-events-dogfood.md`](2026-07-10-session-events-dogfood.md)
+- [`2026-07-10-nightly-drill-round-1.md`](2026-07-10-nightly-drill-round-1.md)
+- [`2026-07-10-nightly-drill-round-1-evidence.md`](2026-07-10-nightly-drill-round-1-evidence.md)
+- [`2026-07-10-next-stage-dogfood-runs.md`](2026-07-10-next-stage-dogfood-runs.md)
+- [`2026-07-10-init-evolution-review.md`](2026-07-10-init-evolution-review.md)
+- [`2026-07-09-be-001-pi-real-run.md`](2026-07-09-be-001-pi-real-run.md)
+- [`2026-07-05-website-docs-governance-multi-executor-dogfood.md`](2026-07-05-website-docs-governance-multi-executor-dogfood.md)
+- [`2026-07-04-website-docs-dynamic-workflow-dogfood.md`](2026-07-04-website-docs-dynamic-workflow-dogfood.md)
+- [`2026-07-02-repository-analysis.md`](2026-07-02-repository-analysis.md)
+- [`2026-06-06-pi-sdk-runtime-reuse-benchmark.md`](2026-06-06-pi-sdk-runtime-reuse-benchmark.md)

package/docs/skills/README.md

@@ -3,4 +3,5 @@
 This directory records repo-local skill wrappers and vetting notes used by Agent DAG role mapping.
 
 - `vetted-skill-registry.md` — supported roles, source inspiration, risk notes, and default/optional usage.
-
+- `../../skills/agent-worker/SKILL.md` — optional outer-loop operator skill for Feature Packet, TaskSpec, Task Pool, controller pinning, self-hosting canaries, and Worker recovery. Single DAG implementation or runtime repair stays with `loop-agent`; this skill is not a default DAG role skill.
+- `../../scripts/check-skill-entry.sh` — validates both public skill entries, their required references, line budgets, and the `agent-worker` trigger vocabulary.

package/docs/skills/vetted-skill-registry.md

@@ -8,6 +8,7 @@ The entries below are local wrappers or existing local skills. They are not whol
 |---|---|---|---|---|---|
 | `ai-engineering-context` | local existing | `skills/ai-engineering-context/SKILL.md` | scout, default context | default/scout | Read-only engineering context; not a private platform memory skill. |
 | `loop-agent` | local existing | `skills/loop-agent/SKILL.md` | planner, supervisor, closeout | planner/closeout | Long references may be resolved by strict audit with expanded budget; executor behavior unchanged. |
+| `agent-worker` | local operator skill | `skills/agent-worker/SKILL.md` | outer-loop operator only | never a default DAG role | Routes Feature Packet, TaskSpec, Task Pool, controller pinning, self-hosting canary, and failure recovery. Must not launch recursively from DAG leaves or duplicate executor/kernel behavior. |
 | `verification-before-completion` | local wrapper inspired by verification discipline | `skills/verification-before-completion/SKILL.md` | implementer, verifier, closeout | implementer/verifier/closeout | Requires shell evidence before completion claims. |
 | `systematic-debugging` | local wrapper inspired by systematic debugging discipline | `skills/systematic-debugging/SKILL.md` | implementer, verifier | verifier | Advisory prompt guidance only; does not run tools by itself. |
 | `requesting-code-review` | local existing | `skills/requesting-code-review/SKILL.md` | reviewer | reviewer | Review prompt guidance only. |
@@ -20,7 +21,7 @@ The entries below are local wrappers or existing local skills. They are not whol
 ## Vetting Rules
 
 - Default role mappings may reference only repo-local skills that resolve cleanly under `dag validate --strict-skills`.
+- `agent-worker` is explicitly outside default role mappings. Its trigger description must cover `agent-worker`, Feature Packet, TaskSpec, Task Pool, self-host/candidate and the `loop-agent` routing boundary; `scripts/check-skill-entry.sh` enforces this public entry contract.
 - Optional/security/web skills remain task- or profile-specific until their tool, network, credential, and write behavior is reviewed.
 - This registry records source inspiration, not license clearance for vendored third-party content. Vendoring requires a separate license/security review.
 - `SKILL.md` is the entry point. References must be declared in frontmatter and stay within the skill directory.
-

package/docs/templates/agent-dag-report.schema.json

@@ -328,7 +328,7 @@
         "order": { "type": "integer", "minimum": 0 },
         "executor": {
           "type": "string",
-          "enum": ["cursor", "pi", "shell", "static"]
+          "enum": ["pi", "shell", "static"]
         },
         "model": { "type": "string" },
         "status": {
@@ -402,7 +402,9 @@
             "finished",
             "partial_failed",
             "failed",
-            "paused"
+            "paused",
+            "superseded",
+            "abandoned"
           ]
         },
         "runDir": { "type": "string", "minLength": 1 },

package/docs/templates/agent-dag.base.json

@@ -57,11 +57,6 @@
     ]
   },
   "executorModels": {
-    "cursor": {
-      "LOW": "composer-2.5",
-      "MED": "composer-2.5",
-      "HIGH": "gpt-5.5"
-    },
     "pi": {
       "LOW": "gpt-5.3-codex-spark",
       "MED": "glm-5.2",

package/docs/templates/agent-dag.final-verification.json

@@ -25,11 +25,6 @@
     ]
   },
   "executorModels": {
-    "cursor": {
-      "LOW": "composer-2.5",
-      "MED": "composer-2.5",
-      "HIGH": "gpt-5.5"
-    },
     "pi": {
       "LOW": "gpt-5.3-codex-spark",
       "MED": "glm-5.2",

package/docs/templates/agent-dag.schema.json

@@ -59,7 +59,6 @@
       "description": "Executor-owned model routing template keyed by executor then complexity. Do not use legacy top-level models or defaults.model.",
       "additionalProperties": false,
       "properties": {
-        "cursor": { "$ref": "#/$defs/complexityModelMap" },
         "pi": { "$ref": "#/$defs/complexityModelMap" }
       }
     },
@@ -77,7 +76,7 @@
       "enum": ["LOW", "MED", "HIGH"]
     },
     "executor": {
-      "enum": ["cursor", "pi", "shell", "static"]
+      "enum": ["pi", "shell", "static"]
     },
     "toolProfile": {
       "enum": ["read-only", "write"]

package/docs/templates/agent-dag.supervised-implementation.json

@@ -22,7 +22,7 @@
   "globalConstraints": [
     "Replace every REPLACE/WITH/... placeholder with concrete repo paths before execution; do not leave template placeholders in production DAG JSON.",
     "Do not commit from DAG nodes; main session inspects and commits after Decision Gate approval.",
-    "Use only existing executors: pi, shell, static; cursor is optional when explicitly enabled. Pi writer nodes must set toolProfile=write.",
+    "Use only existing executors: pi, shell, static. Cursor is available only through the explicit cursor-prompt sidecar, never as a DAG executor. Pi writer nodes must set toolProfile=write.",
     "Read-only nodes must not write repository files, including root artifacts/**.",
     "Exclusive writer nodes must stay within declared writeSet.",
     "Do not implement automatic retry/resume or mutate historical completed facts.",
@@ -73,11 +73,6 @@
     ]
   },
   "executorModels": {
-    "cursor": {
-      "LOW": "composer-2.5",
-      "MED": "composer-2.5",
-      "HIGH": "gpt-5.5"
-    },
     "pi": {
       "LOW": "gpt-5.3-codex-spark",
       "MED": "glm-5.2",

package/docs/templates/backend-test-dag.generate-pytest.prompt.md

@@ -0,0 +1,131 @@
+# Backend Test DAG Generate Pytest Prompt Template
+
+## Purpose
+
+Use this prompt for a **pytest code generation** node: `executor: "pi"`, `role: "implementer"`, `toolProfile: "write"`, `writePolicy: "exclusive"`. The implementer converts reviewed backend functional test cases into pytest automation code with 1:1 traceability.
+
+Do **not** create a new executor type. This is a standard `executor: pi` writer node.
+
+## Recommended DAG Node Shape
+
+```json
+{
+  "id": "generate-backend-pytest-pi",
+  "depends_on": ["review-backend-cases-pi"],
+  "complexity": "HIGH",
+  "executor": "pi",
+  "role": "implementer",
+  "toolProfile": "write",
+  "writePolicy": "exclusive",
+  "writeSet": ["testcase/**"],
+  "allowedPaths": ["testcase/**"],
+  "forbiddenPaths": [".harness/**", "artifacts/**"],
+  "outputContract": "Pytest test files under tests/backend/ with 1:1 mapping to functional test case IDs. Summary lists generated files, test function count, and any skipped cases with reasons.",
+  "subtask_prompt_markdown": "./backend-test-dag.generate-pytest.prompt.md"
+}
+```
+
+## Prompt Body
+
+You are the Backend Test DAG **pytest code generator**.
+
+Your job is to convert reviewed backend functional test cases into pytest automation code. You are an implementer — write test files under `tests/backend/` only. Stay within `writeSet`. Do not write root `artifacts/**`.
+
+### Mandatory First Line
+
+The **first non-empty line** of your response must be a summary line:
+
+```
+Generated N pytest functions across M files under tests/backend/.
+```
+
+### Inputs
+
+1. **Reviewed test cases** — files under `testcase/md/` (approved by `review-backend-cases-pi` with `VERDICT: pass`).
+2. **Source references** — `source/references/` for API specs, data models, database schemas.
+3. **Target project conventions** — inspect `conftest.py`, `pytest.ini` / `pyproject.toml` pytest config, fixture patterns, existing test structure.
+
+### Conversion Rules
+
+#### File Naming
+
+- Every test file must start with `test_` prefix (e.g. `test_order.py`, `test_user_api.py`)
+- pytest collects tests from files matching `test_*.py` or `*_test.py` — use `test_` prefix exclusively
+- Never create test files without the `test_` prefix
+
+#### Write Boundary
+
+- Only **create new** test script files under `testcase/`
+- Do NOT modify existing files: `conftest.py`, `pytest.ini`, `pyproject.toml`, `setup.cfg`, `__init__.py`, or any other framework/config file
+- If the project needs new fixtures, create them in new files (e.g. `testcase/conftest.py` if it does not exist), never edit existing ones
+- Read existing framework files to understand conventions, but treat them as immutable
+
+#### Naming Conflict Resolution
+
+- If a file with the target name already exists under `testcase/`, add a numeric suffix: `test_order.py` → `test_order_01.py` → `test_order_02.py`
+- Never overwrite or append to existing files — each test script must be a standalone file
+- Check for existing files before writing; if `test_<module>.py` exists, use `test_<module>_01.py`
+
+#### 1:1 Traceability
+
+Every functional test case ID (`BE-<MODULE>-<NNN>`) must map to exactly one pytest function:
+
+```python
+# test-cases/backend/BE-ORDER-001 → tests/backend/test_order.py
+def test_BE_ORDER_001_create_order_with_valid_data():
+    """BE-ORDER-001: Create order with valid request body."""
+    ...
+```
+
+- Function name: `test_<CASE_ID_with_underscores>` (e.g. `test_BE_ORDER_001_...`)
+- Docstring first line: `<CASE_ID>: <Case Title>`
+
+#### File Organization
+
+- All test files go under `testcase/` directory in the host project root
+- Group test files by MODULE segment: `BE-ORDER-*` → `testcase/test_order.py`, `BE-USER-*` → `testcase/test_user.py`
+- Follow existing project conventions for import style, fixture scope
+
+#### Fixture Strategy
+
+- Reuse existing project fixtures from `conftest.py` when available
+- Create new fixtures only when test data setup is complex or shared across cases
+- Prefer `@pytest.fixture(scope="function")` for test isolation
+- Use `@pytest.mark.parametrize` for boundary condition cases with multiple inputs
+
+#### Test Integrity
+
+- Tests verify implementation correctness — if a test fails, the implementation likely has a bug, not the test
+- Do NOT weaken assertions, remove test cases, or modify test logic to make tests pass
+- Do NOT add workarounds, skips, or try/except blocks to hide failures without explicit justification
+- Report all failures honestly in the output; the downstream `execute-backend-pytest-shell` node captures exit codes and stdout/stderr as-is
+
+#### Assertions
+
+- Use `assert` statements, not `unittest` assertions
+- Assert specific values, not just "no exception"
+- For API tests: assert status code, response body keys, and specific field values
+- For database tests: assert record state after operation
+
+#### Markers
+
+- `@pytest.mark.positive` — happy path cases
+- `@pytest.mark.negative` — error/exception cases
+- `@pytest.mark.boundary` — edge cases
+- `@pytest.mark.<MODULE>` — module-specific marker (e.g. `@pytest.mark.order`)
+
+#### Skip Policy
+
+If a test case cannot be automated (requires external service not mockable, requires manual verification), add it with `@pytest.mark.skip(reason="...")` and document the reason. Do not omit the function — traceability requires it exists.
+
+### Output Shape (after summary line)
+
+After the mandatory summary line, provide:
+
+1. **Generated Files** — list of files written under `tests/backend/`.
+2. **Function Mapping Table** — `| Test Case ID | Pytest Function | File | Marker |`.
+3. **Skipped Cases** — if any, list with reason.
+4. **Conventions Observed** — note project fixtures/config discovered and followed.
+5. **Residual Risks** — cases that may need manual verification or environment setup.
+
+Do not include chain-of-thought. Do not write root `artifacts/**`.