@loomhelm/multi-task-flow 0.0.0-stage → 0.2.1

package/toolkit/skills/dev-plan/references/replay-cases/positive-legacy-direct-with-topology.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"恢复已锁定的旧会话，flow_type=fast-track，delivery_topology=direct。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Historical Direct Topology\n\n## Summary\n- 结论：只读恢复旧会话中的 direct topology。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- `delivery_topology`: `direct`\n- 开发基线：`25a137023908e3b08cc9b9b121395f2c56dcea90`\n\n## Goals\n- 目标 1: 恢复旧 direct 会话\n  保持冻结的 fast-track 路由并继续 PR 收敛。\n\n## Non-goals\n- 不热更新成 Initiative。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 按旧 direct 合同恢复 (skill: $fast-flow)\n\n## Acceptance & Validation\n- recovery checker 接受该字段，新输出 checker 拒绝同形状。\n\n## Assumptions/Defaults\n- 该会话已在旧合同下锁定：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-legacy-fast-aggregate.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"走 fast-track，规划 legacy aggregate-stack。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Legacy Aggregate\n\n## Summary\n- 结论：按冻结旧合同恢复。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- `delivery_topology`: `aggregate-stack`\n- `integration_branch`: `prd/legacy-demo`\n- `child_merge_policy`: `risk-gated`\n- `final_merge_policy`: `owner-explicit`\n- `ticket_plan_policy`: `inherit-approved-plan`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 恢复旧 Initiative\n  只恢复已经锁定的跨 Ticket frontier。\n\n## Non-goals\n- 不生成新的 legacy 计划。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 按冻结合同恢复旧 frontier (skill: $initiative-flow + $fast-flow)\n\n## Acceptance & Validation\n- 旧会话保持原 flow 与 exact-SHA 门禁。\n\n## Assumptions/Defaults\n- 这是 recovery-only fixture：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-legacy-unprofiled-subagent.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"恢复历史 fast-track session。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Historical Sub Agent\n\n## Summary\n- 结论：恢复旧 session 中未带 profile 的只读 sub-agent，并停在 merge-ready。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `default_fast_track`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n- 收口条件：`merge-ready / Step 5C Ready`\n\n## Goals\n- 目标 1: 读取旧 executor\n  recovery 允许历史 sub-agent 继续被解析。\n\n## Non-goals\n- 不让新计划继续生成无 profile executor。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:legacy-review => 只读检查历史 session; model=gpt-5.6-luna; reasoning_effort=high (skill: $fast-flow)\n- wave: 2\n  - main-agent => 继续执行批准范围 (skill: $fast-flow)\n\n## Acceptance & Validation\n- recovery 模式接受旧 sub-agent，普通新输出仍拒绝同样格式。\n\n## Assumptions/Defaults\n- 这是历史 recovery fixture：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-multi-plan-baseline-reset-explicit-dev-plan-with-continue-wording.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Explicit Branch Baseline Before Continue-wording Re-ask\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 release branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  第一轮计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不验证默认主干解析。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 第一轮请求已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 继续给我一份新的可执行计划，做导入页面。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Explicit Dev-plan Continue Wording Still Opens A Fresh Plan Session\n\n## Summary\n- 结论：新的显式 `$dev-plan` 请求即使同句带有 `继续` 词，也必须重置上一轮 owner baseline 并回到目标仓库主干 concrete ref。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `default_fast_track`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证显式 `$dev-plan` 新请求优先于 continuation 词\n  第二轮请求要求的是一份新的导入页面计划；它没有显式 branch/commit，也没有其他强 Git 事实，因此 baseline 应回到默认主干 concrete ref，而不是沿用上一轮 `upstream/release-2026q3`。\n\n## Non-goals\n- 不把显式 `$dev-plan` 新请求中的 `继续` 一词当成沿用上一轮 planning session 的证据。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划导入页面的实现、验证与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受显式 `$dev-plan` 新请求即使带 `继续` wording 也会正确重置 baseline 到 `origin/main`。\n\n## Assumptions/Defaults\n- 第二轮请求没有显式指定其他 base：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-multi-plan-baseline-reset-generic-reask.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Explicit Branch Baseline Generic Re-ask\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 release branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  第一轮计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不验证默认主干解析。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 第一轮请求已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请再给我一份新的可执行计划。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Second Plan Resets Previous Baseline On Generic Re-ask\n\n## Summary\n- 结论：新的泛化重提计划请求也应重置上一轮显式 baseline，并回到目标仓库主干的具体 ref。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `default_fast_track`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证 generic re-ask 不会继承上一轮显式 baseline\n  第二轮请求要求的是一份新的可执行计划，但没有显式 branch/commit，也没有其他强 Git 事实，因此 baseline 应回到默认主干 concrete ref。\n\n## Non-goals\n- 不延续第一轮 release baseline。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划实现、验证与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受 generic re-ask 场景下 baseline 被正确重置为 `origin/main`。\n\n## Assumptions/Defaults\n- 第二轮请求没有显式指定其他 base：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-multi-plan-baseline-reset-merge-queue-api.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Explicit Branch Baseline Before Merge Queue API Re-ask\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 release branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  第一轮计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不验证默认主干解析。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 第一轮请求已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 给我一份 merge queue API 的计划。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Merge Queue API Request Opens A Fresh Plan Session\n\n## Summary\n- 结论：显式 `$dev-plan` 的新任务请求即使包含 `merge` 词根，也必须重置上一轮 owner baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `default_fast_track`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证 `merge queue API` 这类任务名不会被误判成 continuation\n  第二轮请求要求的是一份新的计划，但没有显式 branch/commit，也没有其他强 Git 事实，因此 baseline 应回到默认主干 concrete ref。\n\n## Non-goals\n- 不把任务名里的 `merge` 词根当成 continuation 信号。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划 merge queue API 的实现、验证与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受 `merge queue API` 这类显式新计划请求会正确重置 baseline 到 `origin/main`。\n\n## Assumptions/Defaults\n- 第二轮请求没有显式指定其他 base：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-multi-plan-baseline-reset-merge-ready.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Explicit Branch Baseline Before Merge-ready Re-ask\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 release branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  第一轮计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不验证默认主干解析。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 第一轮请求已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 给我一份 merge-ready 的 `flow_type: fast-track` 可执行计划。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Second Plan Resets Previous Baseline Despite Merge-ready Wording\n\n## Summary\n- 结论：新的显式 `$dev-plan` 请求即使包含 `merge-ready` 这类 stop-condition 词，也必须重置上一轮 owner baseline 并回到目标仓库主干 concrete ref。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证包含 merge-ready wording 的新计划请求不会继承上一轮显式 baseline\n  第二轮请求要求的是一份新的 `fast-track` merge-ready 计划，但没有显式 branch/commit，也没有其他强 Git 事实，因此 baseline 应回到默认主干 concrete ref。\n\n## Non-goals\n- 不延续第一轮 release baseline。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划实现、验证与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受包含 `merge-ready` wording 的新 `$dev-plan` 请求正确重置 baseline 到 `origin/main`。\n\n## Assumptions/Defaults\n- 第二轮请求没有显式指定其他 base：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-multi-plan-baseline-reset-no-dev-plan-token.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Explicit Branch Baseline Without Later Token\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 release branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  第一轮计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不验证默认主干解析。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 第一轮请求已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"再按普通流程计划。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Second Plan Resets Previous Explicit Baseline Without Repeating Token\n\n## Summary\n- 结论：后一轮新计划请求即使没重复 `$dev-plan` token，也不应沿用上一轮 owner baseline。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证新一轮普通流程计划不会继承上一轮显式 baseline\n  第二轮请求显式锁了 `normal`，但没有显式 branch/commit，也没有其他强 Git 事实，因此 baseline 应回到默认主干 concrete ref。\n\n## Non-goals\n- 不延续第一轮 release baseline。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划普通流程下的实现、验证与本地提交收口 (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须接受后一轮未重复 `$dev-plan` token 时仍能正确重置 baseline 的计划输出。\n\n## Assumptions/Defaults\n- 第二轮请求没有显式指定其他 base：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-multi-plan-baseline-reset-plan-task-wording.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Explicit Branch Baseline Before Planning-Task Re-ask\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 release branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  第一轮计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不验证默认主干解析。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 第一轮请求已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请帮我规划这个任务：做一个导入页面。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Planning-Task Wording Opens A Fresh Plan Session\n\n## Summary\n- 结论：新的“规划这个任务”请求必须打开新的 planning session，而不是沿用上一轮 owner baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `default_fast_track`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证“规划这个任务” wording 会重置上一轮显式 baseline\n  第二轮请求要求的是一份新的任务计划，但没有显式 branch/commit，也没有其他强 Git 事实，因此 baseline 应回到默认主干 concrete ref。\n\n## Non-goals\n- 不延续第一轮 release baseline。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划导入页面的实现、验证与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受“请帮我规划这个任务”场景下 baseline 被正确重置为 `origin/main`。\n\n## Assumptions/Defaults\n- 第二轮请求没有显式指定其他 base：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-multi-plan-baseline-reset-replan-merge-ready-clause.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Explicit Branch Baseline Before Replan Merge-ready Clause\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 release branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  第一轮计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不验证默认主干解析。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 第一轮请求已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"重新规划这个任务，收口改为 merge-ready。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Replan Merge-ready Clause Resets Previous Baseline\n\n## Summary\n- 结论：新的重新规划请求即使同句包含 `改为 merge-ready`，也应回到目标仓库主干的具体 ref，而不是沿用上一轮 owner baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `default_fast_track`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证“重新规划这个任务，收口改为 merge-ready”会重置上一轮显式 baseline\n  第二轮请求是在重新规划当前任务，只是顺手把 stop condition 改成 `merge-ready`；它没有显式 branch/commit，也没有其他强 Git 事实，因此 baseline 应回到默认主干 concrete ref。\n\n## Non-goals\n- 不延续第一轮 release baseline。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划实现、验证与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受这类带 `改为 merge-ready` clause 的 replan 请求正确重置 baseline 到 `origin/main`。\n\n## Assumptions/Defaults\n- 第二轮请求没有显式指定其他 base：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-multi-plan-baseline-reset-replan-task.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Explicit Branch Baseline Before Replan Task\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 release branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  第一轮计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不验证默认主干解析。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 第一轮请求已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请重新规划这个任务，flow_type: normal。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Replan Task Resets Previous Explicit Baseline\n\n## Summary\n- 结论：新的重新规划请求未指定 base 时，应回到目标仓库主干的具体 ref，而不是沿用上一轮 owner baseline。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证“重新规划这个任务”会重置上一轮显式 baseline\n  第二轮请求显式锁了 `normal`，但没有显式 branch/commit，也没有其他强 Git 事实，因此 baseline 应回到默认主干 concrete ref。\n\n## Non-goals\n- 不延续第一轮 release baseline。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划普通流程下的实现、验证与本地提交收口 (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须接受“重新规划这个任务”场景下 baseline 被正确重置为 `origin/main`。\n\n## Assumptions/Defaults\n- 第二轮请求没有显式指定其他 base：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-multi-plan-baseline-reset.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Explicit Branch Baseline\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 release branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  第一轮计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不验证默认主干解析。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 第一轮请求已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 再给我一份新的 `flow_type: fast-track` 可执行计划。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Second Plan Resets Previous Explicit Baseline\n\n## Summary\n- 结论：新的显式 `$dev-plan` 请求未指定 base 时，应回到目标仓库主干的具体 ref，而不是沿用上一轮 owner baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证新一轮计划不会继承上一轮显式 baseline\n  第二轮请求显式锁了 `fast-track`，但没有显式 branch/commit，也没有其他强 Git 事实，因此 baseline 应回到默认主干 concrete ref。\n\n## Non-goals\n- 不延续第一轮 release baseline。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划实现、验证与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受多轮 `$dev-plan` 中后一轮回到 `origin/main` 的计划输出。\n\n## Assumptions/Defaults\n- 第二轮请求没有显式指定其他 base：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-no-delivery-topology.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 走 fast-track。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Fast Track Without Legacy Topology\n\n## Summary\n- 结论：新计划只声明一个顶层 flow，不再输出 delivery topology。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 验证单一顶层 flow 合同\n  fast-track 直接完成实现并收敛到 merge-ready。\n\n## Non-goals\n- 不进入 Initiative 生命周期。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 实现并收敛 direct 交付 (skill: $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受不含 delivery_topology 的新计划。\n\n## Assumptions/Defaults\n- 这是新产生的计划：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-agent-decision-log-api.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Agent Decision Log API\n\n## Summary\n- 结论：普通流程下实现 agent decision log export API，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: add the agent decision log export API\n  Implement a product-facing export API for agent decision logs without creating a persisted planning decision log.\n\n## Non-goals\n- Do not create any planning decision-log artifact.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => add agent decision log export API for workspace admins (skill: $normal-flow)\n\n## Acceptance & Validation\n- Workspace admins can export agent decision logs through the product API.\n\n## Assumptions/Defaults\n- This replay validates agent-facing decision-log product wording: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-agent-handoff-log-exporter.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Agent Handoff Log Exporter\n\n## Summary\n- 结论：普通流程下实现 agent handoff log exporter，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: add the agent handoff log exporter\n  Create an agent handoff log exporter and storage pipeline for customer records without introducing any planning handoff artifact for future sessions.\n\n## Non-goals\n- Do not create a planning handoff log for future sessions or turns.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => create agent handoff log exporter and storage pipeline (skill: $normal-flow)\n\n## Acceptance & Validation\n- The agent handoff log exporter stores customer records correctly.\n\n## Assumptions/Defaults\n- This replay validates agent-facing product wording around handoff logs: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-decision-log-entries.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Workflow Decision Log Entries\n\n## Summary\n- 结论：普通流程下实现 workflow decision log entries，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: implement workflow decision log entries\n  Record workflow decision log entries when users approve requests, while keeping the surrounding workflow unchanged.\n\n## Non-goals\n- Do not create any planning handoff artifact.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => record workflow decision log entries when users approve requests (skill: $normal-flow)\n\n## Acceptance & Validation\n- Product workflow decision log entries are recorded for approvals.\n\n## Assumptions/Defaults\n- This replay validates non-planning workflow decision log wording: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-decision-log-ui.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Decision Log UI\n\n## Summary\n- 结论：普通流程下实现 decision log UI 能力，不落盘计划载体。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: implement the decision log UI\n  Add a user-facing decision log viewer with filters and pagination while keeping the existing API surface unchanged.\n\n## Non-goals\n- Do not create any persisted plan artifact.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => update the decision log UI to show filters and pagination (skill: $normal-flow)\n\n## Acceptance & Validation\n- The UI renders decision log entries with filters and pagination.\n\n## Assumptions/Defaults\n- This replay validates non-plan decision log wording: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-future-handoff-log-timeline.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Future Handoff Log Timeline\n\n## Summary\n- 结论：普通流程下实现 future handoff log timeline 页面，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: add the future handoff log timeline page\n  Render a product-facing future handoff log timeline page for operators without introducing any persisted planning handoff artifact.\n\n## Non-goals\n- Do not create a planning handoff log document.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => add the future handoff log timeline page for operators (skill: $normal-flow)\n\n## Acceptance & Validation\n- Operators can open the future handoff log timeline page in the product.\n\n## Assumptions/Defaults\n- This replay validates temporal product wording around handoff-log timelines: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-handoff-log-ui.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Handoff Log UI\n\n## Summary\n- 结论：普通流程下实现 handoff log UI 能力，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: implement the handoff log UI\n  Add a user-facing handoff log viewer and API for transfer history while keeping the surrounding product behavior unchanged.\n\n## Non-goals\n- Do not create any planning handoff artifact.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => update the handoff log UI and API for transfer history (skill: $normal-flow)\n\n## Acceptance & Validation\n- The handoff log UI renders transfer history correctly.\n\n## Assumptions/Defaults\n- This replay validates non-planning handoff log wording: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-next-session-decision-log-api.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Next Session Decision Log API\n\n## Summary\n- 结论：普通流程下实现 next-session decision log entries API，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: add the next-session decision log entries API\n  Expose product-facing decision log entries for the next-session experience without introducing any persisted planning document.\n\n## Non-goals\n- Do not create a planning decision log artifact.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => add the next-session decision log entries API for product usage (skill: $normal-flow)\n\n## Acceptance & Validation\n- The next-session decision log entries API is available for product consumers.\n\n## Assumptions/Defaults\n- This replay validates temporal product wording around decision-log APIs: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-next-session-decision-log-exporter.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Next Session Decision Log Exporter\n\n## Summary\n- 结论：普通流程下实现 next-session decision log exporter，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: add the next-session decision log exporter\n  Create a CSV exporter for next-session decision log entries so operators can retrieve product data without creating any planning decision-log artifact.\n\n## Non-goals\n- Do not create a planning decision log artifact for future work or handoffs.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => create the CSV exporter for next-session decision log entries (skill: $normal-flow)\n\n## Acceptance & Validation\n- Operators can export next-session decision log entries as CSV.\n\n## Assumptions/Defaults\n- This replay validates exporter wording around next-session decision log entries: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-pending-state-acceptance.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Pending State Acceptance\n\n## Summary\n- 结论：普通流程下具体验收可以验证 pending-state 行为，不应被当成占位文本。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: preserve pending-state visibility\n  Keep pending jobs visible in the queue after refresh without changing the rest of the workflow.\n\n## Non-goals\n- Do not replace the pending-state flow with a placeholder acceptance line.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => implement pending-state handling for queue refresh behavior (skill: $normal-flow)\n\n## Acceptance & Validation\n- Pending jobs remain visible in the queue after refresh.\n\n## Assumptions/Defaults\n- This replay validates concrete pending-state acceptance wording: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-plan-file-upload-api.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal Plan File Upload API\n\n## Summary\n- 结论：普通流程下实现 plan file upload API，不应被当成 persisted plan artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: add the plan file upload API\n  Implement a product-facing upload API and storage flow for user plan files without introducing any persisted execution-plan document.\n\n## Non-goals\n- Do not create or maintain a planning artifact.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => update the plan file upload API and related storage integration (skill: $normal-flow)\n\n## Acceptance & Validation\n- Uploading a plan file succeeds through the product API.\n\n## Assumptions/Defaults\n- This replay validates generic product-facing plan-file wording: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-plan-md-cleanup.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal PLAN.md Cleanup\n\n## Summary\n- 结论：普通流程下清理遗留 PLAN.md 引用，不应被当成 persisted plan artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: remove stale legacy plan references\n  Clean up stale `PLAN.md` references from docs and tests so the repo no longer points to deprecated persisted plan files.\n\n## Non-goals\n- Do not create or maintain any persisted plan file.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => remove stale references to `PLAN.md` from the docs and tests (skill: $normal-flow)\n\n## Acceptance & Validation\n- No stale `PLAN.md` references remain in the touched docs and tests.\n\n## Assumptions/Defaults\n- This replay validates cleanup wording for deprecated plan-file references: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-plan-md-import-review-page.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal PLAN.md Import Review Page\n\n## Summary\n- 结论：普通流程下支持 user-uploaded PLAN.md 的 import API 和 review page，不应被当成 persisted plan artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: support user-uploaded PLAN.md product review\n  Add support for user-uploaded PLAN.md documents in the import API and review page without persisting any execution plan.\n\n## Non-goals\n- Do not reuse PLAN.md as a planning artifact for future handoff.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => add support for user-uploaded PLAN.md documents in import API and review page (skill: $normal-flow)\n\n## Acceptance & Validation\n- The import API accepts uploaded PLAN.md documents and the review page renders them as product content.\n\n## Assumptions/Defaults\n- This replay validates product-safe PLAN.md wording with review-page UI: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-plan-md-importer.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal PLAN.md Importer\n\n## Summary\n- 结论：普通流程下支持 importer 接受 customer-provided `PLAN.md`，不应被当成 persisted plan artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: support customer-provided PLAN.md imports\n  Update the importer to accept customer-provided `PLAN.md` as product content without persisting any execution plan.\n\n## Non-goals\n- Do not create or maintain a planning artifact for future execution.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => update the importer to accept customer-provided PLAN.md (skill: $normal-flow)\n\n## Acceptance & Validation\n- Customer-provided `PLAN.md` imports successfully through the product importer.\n\n## Assumptions/Defaults\n- This replay validates importer wording around customer-provided `PLAN.md`: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-plan-md-planning-ui-import.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal PLAN.md Planning UI Import\n\n## Summary\n- 结论：普通流程下支持 planning UI 和 import API 处理 user-provided PLAN.md，不应被当成 persisted plan artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: support user-provided PLAN.md in planning UI\n  Update the planning UI and import API to accept user-provided PLAN.md documents as product content without persisting any execution plan.\n\n## Non-goals\n- Do not reuse PLAN.md for future handoff or internal planning state.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => update the planning UI and import API to accept user-provided PLAN.md documents (skill: $normal-flow)\n\n## Acceptance & Validation\n- The planning UI and import API accept uploaded PLAN.md documents as product content.\n\n## Assumptions/Defaults\n- This replay validates product-safe PLAN.md wording with planning UI surface: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-plan-md-user-upload.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按 normal 流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Positive Normal PLAN.md User Upload\n\n## Summary\n- 结论：普通流程下支持用户上传 `PLAN.md` 文件，不应被当成 persisted plan artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: support user-provided PLAN.md uploads\n  Update the upload API to accept `PLAN.md` files from users as product content without persisting an execution plan.\n\n## Non-goals\n- Do not create or maintain a planning artifact for future execution.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => update the upload API to accept `PLAN.md` files from users (skill: $normal-flow)\n\n## Acceptance & Validation\n- The upload API accepts user-provided `PLAN.md` files successfully.\n\n## Assumptions/Defaults\n- This replay validates product file-support wording around `PLAN.md`: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-plan-template-decision-log-history.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按 normal 流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Positive Normal Plan Template Decision Log History\n\n## Summary\n- 结论：普通流程下实现 plan template decision log history，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: record decision log history for plan template edits\n  Record product-facing workflow decision log history when users edit plan templates, without creating any persisted planning artifact.\n\n## Non-goals\n- Do not create a planning decision log artifact for future work or handoffs.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => record workflow decision log history when users edit plan templates (skill: $normal-flow)\n\n## Acceptance & Validation\n- Product decision log history is recorded for plan template edits.\n\n## Assumptions/Defaults\n- Plan templates remain part of the product surface: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-plan-template-handoff-log-history.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按 normal 流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Positive Normal Plan Template Handoff Log History\n\n## Summary\n- 结论：普通流程下实现 plan template handoff log history，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: record handoff log history for plan template transfers\n  Record product-facing handoff log history when users transfer plan templates, without creating any persisted planning artifact.\n\n## Non-goals\n- Do not create a planning handoff log artifact for future sessions or turns.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => create agent handoff log history when users transfer plan templates (skill: $normal-flow)\n\n## Acceptance & Validation\n- Product handoff log history is recorded for plan template transfers.\n\n## Assumptions/Defaults\n- Plan templates remain part of the product surface: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal-progress-log-ui.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按 normal 流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Positive Normal Progress Log UI\n\n## Summary\n- 结论：普通流程下实现 progress log UI，不应被当成 planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: implement the progress log UI\n  Add product-facing progress log API and UI views so users can inspect runtime progress entries.\n\n## Non-goals\n- Do not preserve any execution-plan handoff artifact.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => create progress log API and UI views for operators (skill: $normal-flow)\n\n## Acceptance & Validation\n- Operators can inspect progress log entries in the product UI.\n\n## Assumptions/Defaults\n- Existing runtime progress events remain available: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-normal.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按普通流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# 普通流程示例计划\n\n## Summary\n- 结论：普通流程编排包含分支与本地提交收口。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 锁定示例范围合同\n  在执行编排前明确结果边界、关键实现细节与成功信号。\n\n## Non-goals\n- 不扩展示例合同之外的范围。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 创建分支并实现变更 (skill: $normal-flow)\n  - main-agent => 在本地运行验证 (skill: $normal-flow)\n- wave: 2\n  - main-agent => 准备本地提交并等待下一步指示 (skill: $normal-flow)\n\n## Acceptance & Validation\n- 示例验收用于合同回放验证。\n\n## Assumptions/Defaults\n- 示例仅用于回放验证：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-runtime-formal-review.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请按 fast-track 规划 runtime formal review。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Runtime Formal Review\n\n## Summary\n- 结论：formal review 仅在 ready candidate 上运行时创建。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- `review_tier`: `Tier 3`\n- `review_rationale`: `本次合同变化需要 current-head formal evidence。`\n- `runtime_review_strategy`: `验证、必要视觉证据和设计 readiness 通过后，运行时创建 required lanes；Repair Batch 按 delta 刷新。`\n- `review_profile_recommendation`: `四至六个 formal lane 默认建议 gpt-5.6-luna + xhigh，实际 pair 由当前 host 解析并记录。`\n- 开发基线：`origin/main`\n- 收口条件：`merge-ready / Step 5C Ready`\n\n## Goals\n- G1: 固化运行时 formal review。\n  计划不包含 reviewer schedule。\n\n## Non-goals\n- N1: 不锁定具体 reviewer 或每批次固定数量。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 完成 readiness、验证和 Step 5C 收口 (skill: $fast-flow)\n\n## Acceptance & Validation\n- A1 [contract/direct] readiness | evidence=`test:review` | assertions=`not-ready blocks launch,ready Tier 3 covers four lanes,live PR exists,current head,Step 5C Ready`\n- A2 [mechanical/direct] plan guard | evidence=`command:plan_contract_check` | assertions=`runtime strategy passes`\n\n## Assumptions/Defaults\n- current host profile availability 已确定\n</proposed_plan>"}]} }

package/toolkit/skills/dev-plan/references/replay-cases/positive-scope-clarification-import-page-no-reset.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Plan Before Import Scope Clarification\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  当前计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不把后续范围补充误判为新的独立计划请求。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 主人已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请把导入页面也规划进去，其他不变。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Import Scope Clarification Keeps Existing Planning Session\n\n## Summary\n- 结论：普通范围补充即使使用“规划进去”措辞，也不应被误当成新的独立计划请求。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 在范围补充后继续保留同一 planning session 的 flow 与 baseline\n  后续 scope clarification 只是把导入页面纳入既有计划，不应清空当前 flow 锁或 `upstream/release-2026q3` baseline。\n\n## Non-goals\n- 不把“请把 X 也规划进去，其他不变”误判成新的计划请求。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 在既有 merge-ready 计划中纳入导入页面范围补充 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受这类 scope clarification 不会重置当前 flow lock 与显式 baseline。\n\n## Assumptions/Defaults\n- scope clarification 明确要求其他不变：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-scope-clarification-new-plan-file-upload-api-no-reset.jsonl

@@ -0,0 +1,6 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 规划这项工作，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"function_call","name":"request_user_input","call_id":"call_flow_type_lock","arguments":"{\"questions\":[{\"header\":\"流程类型\",\"id\":\"flow_type\",\"question\":\"当前 flow 信号存在歧义，请明确本次流程类型\",\"options\":[{\"label\":\"快车道 (Recommended)\",\"description\":\"推进到 PR merge-ready / Step 5C Ready\"},{\"label\":\"仅实现\",\"description\":\"只产出实现规格\"},{\"label\":\"普通流程\",\"description\":\"验证后本地提交并停下\"},{\"label\":\"本地合主干\",\"description\":\"规划到本地 merge-ready 与本地合并收尾。\"}]}]}"}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"normal"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Scope Clarification With New Plan File Upload API\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 branch baseline。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  当前计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不把产品范围补充误判为新的独立计划请求。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划普通流程下的实现、验证与本地提交收口 (skill: $normal-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 主人已先给出 branch baseline，流程类型在后续单独回答：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"再补充一下范围：需要覆盖 new plan file upload API，但 flow 和开发基线都不变。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Scope Clarification Keeps Existing Planning Session\n\n## Summary\n- 结论：产品范围补充即使包含 `new plan file upload API` 这类词组，也不应被误当成新的独立计划请求。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 在产品范围补充后继续保留同一 planning session 的 flow 与 baseline\n  后续 scope clarification 只是在既有 `normal` 计划上补充 `new plan file upload API` 这个产品能力，不应清空当前 flow 锁或 `upstream/release-2026q3` baseline。\n\n## Non-goals\n- 不把产品词组 `new plan file upload API` 当成新的计划请求信号。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 在既有普通流程计划中纳入 plan file upload API 的产品范围补充 (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须接受这类 scope clarification 不会重置当前 flow lock 与显式 baseline。\n\n## Assumptions/Defaults\n- scope clarification 明确说明 flow 与开发基线都不变：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-scope-clarification-plan-md-no-reset.jsonl

@@ -0,0 +1,4 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 直接给我 `flow_type: fast-track` 计划，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Plan Before PLAN.md Scope Clarification\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 branch baseline。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  当前计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不把后续范围补充误判为新的独立计划请求。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划基于 release 分支的实现与 merge-ready 收口 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 主人已显式给出 branch baseline：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请按 fast-track 把 plan.md cleanup 也规划进去"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo PLAN.md Scope Clarification Keeps Existing Planning Session\n\n## Summary\n- 结论：包含 `plan.md` 的范围补充即使同时重述已锁定 flow，也不应被误当成新的独立计划请求。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 在 `plan.md` 范围补充后继续保留同一 planning session 的 flow 与 baseline\n  后续 scope clarification 只是把 plan.md cleanup 纳入既有计划，不应清空当前 flow 锁或 `upstream/release-2026q3` baseline。\n\n## Non-goals\n- 不把 `plan.md` 文件名中的 `plan` 误判成新的计划请求信号。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 在既有 merge-ready 计划中纳入 plan.md cleanup 范围补充 (skill: $review-agent + $fast-flow)\n\n## Acceptance & Validation\n- checker 必须接受这类 `plan.md` scope clarification 不会重置当前 flow lock 与显式 baseline。\n\n## Assumptions/Defaults\n- scope clarification 只是扩大既有任务范围：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-scope-clarification-plan-page-upload-api-no-reset.jsonl

@@ -0,0 +1,6 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请用 $dev-plan 规划这项工作，开发基线使用 `upstream/release-2026q3`。"}]}}
+{"type":"response_item","payload":{"type":"function_call","name":"request_user_input","call_id":"call_flow_type_plan_page","arguments":"{\"questions\":[{\"header\":\"流程类型\",\"id\":\"flow_type\",\"question\":\"当前 flow 信号存在歧义，请明确本次流程类型\",\"options\":[{\"label\":\"普通流程 (Recommended)\",\"description\":\"验证后本地提交并停下\"},{\"label\":\"快车道\",\"description\":\"推进到 PR merge-ready / Step 5C Ready\"},{\"label\":\"仅实现\",\"description\":\"只产出实现规格\"},{\"label\":\"本地合主干\",\"description\":\"规划到本地 merge-ready 与本地合并收尾。\"}]}]}"}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"normal"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo First Plan Before Plan-page Upload Scope Clarification\n\n## Summary\n- 结论：第一轮计划显式保留 owner 指定的 branch baseline。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 保留第一轮显式 branch baseline\n  当前计划围绕 `upstream/release-2026q3` 编排，不回退到默认主干。\n\n## Non-goals\n- 不把后续范围补充误判为新的独立计划请求。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划普通流程下的实现、验证与本地提交收口 (skill: $normal-flow)\n\n## Acceptance & Validation\n- 第一轮计划必须原样保留显式 branch baseline。\n\n## Assumptions/Defaults\n- 主人已先给出 branch baseline，流程类型在后续单独回答：已确定\n</proposed_plan>"}]}}
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请按 normal 把计划页面上传 API 也规划进去，flow 和开发基线都不变。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Plan-page Upload Scope Clarification Keeps Existing Planning Session\n\n## Summary\n- 结论：`计划页面上传 API` 这类产品范围补充即使包含“计划”字样，也不应被误当成新的独立计划请求。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`upstream/release-2026q3`\n\n## Goals\n- 目标 1: 在计划页面上传 API 的 scope clarification 后继续保留同一 planning session 的 flow 与 baseline\n  后续 scope clarification 只是把 `计划页面上传 API` 纳入既有 `normal` 计划，不应清空当前 flow 锁或 `upstream/release-2026q3` baseline。\n\n## Non-goals\n- 不把产品词组 `计划页面上传 API` 当成新的计划请求信号。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 在既有普通流程计划中纳入计划页面上传 API 的范围补充 (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须接受这类 scope clarification 不会重置当前 flow lock 与显式 baseline。\n\n## Assumptions/Defaults\n- scope clarification 明确说明 flow 与开发基线都不变：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-skill-binding-fullwidth-comma.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"flow_type: normal"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Fullwidth Skill Delimiter\n\n## Summary\n- 结论：用于验证全角分隔符下的 skill refs 仍可解析。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- Goal 1: lock the demo scope contract\n  Capture the result boundary, key implementation detail, and success signal before execution arrangement.\n\n## Non-goals\n- Do not expand beyond the demo contract.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => execute local validation tasks (skill: $normal-flow，$topic-spec)\n\n## Acceptance & Validation\n- Demo acceptance for contract replay.\n\n## Assumptions/Defaults\n- 示例仅用于回放验证：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-specialist-review-profiles.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请按 normal 规划只读分析 sidecar profile。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Analysis Sidecar Profiles\n\n## Summary\n- 结论：非审查只读 sidecar 可以在计划中声明 profile。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 保留 sidecar profile\n  sidecar 只读分析并返回观察，不形成 formal review proof。\n\n## Non-goals\n- 不启动 formal lane。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:analysis [gpt-5.6-luna, high] => 运行普通只读分析 sidecar (skill: $normal-flow)\n- wave: 2\n  - sub-agent:proof-helper [gpt-5.6-sol, low] => 运行只读验证 sidecar (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 接受非审查 sidecar 的显式 profile。\n\n## Assumptions/Defaults\n- sidecar profile 合法：已确定\n</proposed_plan>"}]} }

package/toolkit/skills/dev-plan/references/replay-cases/positive-subagent-commit-history-review.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Read Only Commit History Review\n\n## Summary\n- 结论：审阅 commit history 属于只读检查。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 检查提交历史\n  sub-agent 只检查 commit history in the repository 并回传 findings。\n\n## Non-goals\n- 不允许 sub-agent 写入代码。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:review [gpt-5.6-luna, xhigh] => inspect commit history in the repository and report findings (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须接受只读历史审阅。\n\n## Assumptions/Defaults\n- 用于正例：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-subagent-descriptive-implementation-review.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Descriptive Implementation Review\n\n## Summary\n- 结论：描述代码行为的 review 任务保持只读。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 检查实现行为\n  The sub-agent verifies whether the source code implements the required behavior and reports findings.\n\n## Non-goals\n- 不允许 sub-agent 写入代码。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:review [gpt-5.6-luna, xhigh] => verify whether the source code implements the required behavior and report findings (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 接受描述性 review 任务。\n\n## Assumptions/Defaults\n- 用于正例：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-subagent-git-config-worktree-readonly.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Read Only Git Config And Worktree Inspection\n\n## Summary\n- 结论：只读 sub-agent 可以读取 Git 配置和列出 worktree。\n\n## Flow Decision\n- flow_type: normal\n- flow_source: user_explicit\n- lock_status: locked\n- 开发基线：origin/main\n\n## Goals\n- 目标 1: 保留 Git 状态检查\n  checker 应接受 git config 查询和 git worktree list 检查。\n\n## Non-goals\n- 不开放 sub-agent 修改 Git 配置或 worktree。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:config [gpt-5.6-luna, xhigh] => git config --get user.name (skill: $normal-flow)\n  - sub-agent:config-list [gpt-5.6-luna, xhigh] => git config --list (skill: $normal-flow)\n  - sub-agent:worktree-list [gpt-5.6-luna, xhigh] => git worktree list --porcelain (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须接受这些只读 Git 检查命令。\n\n## Assumptions/Defaults\n- 用于正例：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-subagent-git-readonly-commands.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Read Only Git Validation\n\n## Summary\n- 结论：只读 sub-agent 可以使用 Git patch 检查命令，不应被误判为写入。\n\n## Flow Decision\n- flow_type: normal\n- flow_source: user_explicit\n- lock_status: locked\n- 开发基线：origin/main\n\n## Goals\n- 目标 1: 保留 Git 只读检查\n  checker 应接受 git apply --check、git apply --stat、git apply --numstat 与 git apply --summary。\n\n## Non-goals\n- 不开放 sub-agent 修改 index 或工作树。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:check [gpt-5.6-luna, xhigh] => git apply --check patch.diff (skill: $normal-flow)\n  - sub-agent:stat [gpt-5.6-luna, xhigh] => git apply --stat patch.diff (skill: $normal-flow)\n  - sub-agent:numstat [gpt-5.6-luna, xhigh] => git apply --numstat patch.diff (skill: $normal-flow)\n  - sub-agent:summary [gpt-5.6-luna, xhigh] => git apply --summary patch.diff (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须接受这些只读 patch 验证命令。\n\n## Assumptions/Defaults\n- 用于正例：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-subagent-git-stash-readonly.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Read Only Git Stash Inspection\n\n## Summary\n- 结论：只读 sub-agent 可以查看 stash，不应被误判为写入。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 保留 stash 检查\n  checker 应接受 git stash list 与 git stash show --stat。\n\n## Non-goals\n- 不允许 sub-agent 创建、删除或应用 stash。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:stash-list [gpt-5.6-luna, xhigh] => git stash list (skill: $normal-flow)\n  - sub-agent:stash-show [gpt-5.6-luna, xhigh] => git stash show --stat stash@{0} (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须接受这两个只读 Git 检查命令。\n\n## Assumptions/Defaults\n- 用于正例：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-subagent-indirect-readonly-negation.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Indirect Read Only Negation\n\n## Summary\n- 结论：只读 sub-agent 可以用间接禁止句表达权限边界。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 识别间接只读约束\n  checker 应接受 not allowed、forbidden from 和中文不允许写入表达。\n\n## Non-goals\n- 不开放 sub-agent 的任何写入能力。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:review [gpt-5.6-luna, xhigh] => the sub-agent is not allowed to modify the shared workspace (skill: $normal-flow)\n  - sub-agent:test [gpt-5.6-luna, xhigh] => the sub-agent is forbidden from writing files or changing the current worktree (skill: $normal-flow)\n  - sub-agent:docs [gpt-5.6-luna, xhigh] => 不允许修改当前工作树，只回传检查结果 (skill: $normal-flow)\n  - sub-agent:git [gpt-5.6-luna, xhigh] => the sub-agent is not allowed to git push (skill: $normal-flow)\n  - sub-agent:shell [gpt-5.6-luna, xhigh] => the sub-agent is forbidden from using sed -i on the current file (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须通过所有间接只读禁止声明。\n\n## Assumptions/Defaults\n- 用于正例：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-subagent-readonly-negation.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Read Only Negation\n\n## Summary\n- 结论：只读 sub-agent 可以明确声明禁止写入工作区，而不会被误判为写入。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 保留只读边界\n  checker 应区分禁止写入和实际声明写入。\n\n## Non-goals\n- 不开放 sub-agent 的代码写入能力。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:review [gpt-5.6-luna, xhigh] => verify that the sub-agent must not write to the shared workspace (skill: $normal-flow)\n  - sub-agent:test [gpt-5.6-luna, xhigh] => 明确不得写入当前工作树，只回传测试结果 (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须通过两条只读禁止声明。\n\n## Assumptions/Defaults\n- 用于正例：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-subagent-review-implementation.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Read Only Review Wording\n\n## Summary\n- 结论：常见 review 描述不能被误判为写入。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- 开发基线：`origin/main`\n\n## Goals\n- 目标 1: 保留 review 只读语义\n  sub-agent 只检查当前功能实现并回传 commit history。\n\n## Non-goals\n- 不开放 sub-agent 的代码写入能力。\n\n## Task Orchestration\n- wave: 1\n  - sub-agent:review [gpt-5.6-luna, xhigh] => 检查当前功能实现并回传 commit history (skill: $normal-flow)\n\n## Acceptance & Validation\n- checker 必须接受普通 review 描述。\n\n## Assumptions/Defaults\n- 用于正例：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/positive-wave-thread-v2-legacy.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"恢复旧会话，flow_type=wave-thread。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n## Summary\n- 兼容历史双段合同并在 Step 5C 停止。\n## Flow Decision\n- `flow_type`: `wave-thread`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n- `empirical_acceptance`: `not_required`\n- `empirical_acceptance_rationale`: `历史输入由回放证明。`\n- `review_tier`: `Tier 1`\n- `review_rationale`: `只做合同检查。`\n- `runtime_review_strategy`: `不创建 formal lane。`\n- `review_profile_recommendation`: `无需 profile。`\n- 开发基线：`origin/main`\n- 收口条件：`merge-ready / Step 5C Ready`\n## Goals\n- G1: 保持历史双段合同可读。\n  兼容路径只读历史 JSONL。\n## Non-goals\n- N1: 不创建任务或 PR。\n## Task Orchestration\n- batch:legacy [parent-child] => target=integration/main;depends_on=none;condition=true;acceptance=A1;observation_window=2026-12-31T00:00:00Z\n  - wave: 1\n    - task:legacy-task => owned_surfaces=src/legacy;depends_on=none;condition=true;acceptance=A2;skill=$wave-thread-flow\n## Wave Thread Contracts\n- batch:legacy => mode=parent-child;target=integration/main;depends_on=none;condition=true;acceptance=A1;observation_window=2026-12-31T00:00:00Z\n- task:legacy-task => owned_surfaces=src/legacy;depends_on=none;condition=true;acceptance=A2;skill=$wave-thread-flow\n## Acceptance & Validation\n- A1 [contract/direct] batch | scope=`batch:legacy` | phase=`pre_merge` | evidence=`skills/bin/plan_contract_check.sh` | assertions=`legacy readable; Agent default release decision card publish applicability components channel bump rule predicted version target repository policy source unknown/blocked version expectation`\n- A2 [contract/direct] task | scope=`task:legacy-task` | phase=`pre_merge` | evidence=`skills/bin/wave_thread_plan_contract.py --legacy` | assertions=`legacy readable`\n## Assumptions/Defaults\n- Historical JSONL is read-only compatibility input：已确定。\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/prepare-handoff-log-next-session-artifact.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Prepare Handoff Log for Next Session\n\n## Summary\n- 结论：用于验证 prepare/setup 形式的 planning artifact 仍会被拦截。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n\n## Goals\n- Goal 1: stage a future handoff artifact\n  Prepare a handoff log for the next session review.\n\n## Non-goals\n- Do not change runtime behavior.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => prepare handoff log for next session review (skill: $normal-flow)\n\n## Acceptance & Validation\n- The handoff artifact is ready for the next session.\n\n## Assumptions/Defaults\n- Review handoff format is unchanged: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/read-plan-md-handoff-context-artifact.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按普通流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Read PLAN.md for Handoff Context\n\n## Summary\n- 结论：用于验证 generic PLAN.md read/open wording 也会被当成 persisted planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n\n## Goals\n- Goal 1: reuse a persisted plan file\n  Read PLAN.md for the next handoff context before continuing implementation.\n\n## Non-goals\n- Do not introduce any product-facing PLAN.md file support.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => read PLAN.md for next handoff context (skill: $normal-flow)\n\n## Acceptance & Validation\n- The checker rejects persisted plan reads in Task Orchestration.\n\n## Assumptions/Defaults\n- 该样例专门用于验证 generic PLAN.md read wording：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/redundant-execution-steps.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按普通流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Demo Redundant Execution Steps\n\n## Summary\n- 结论：用于验证 Execution Steps 冗余章节。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n\n## Goals\n- Goal 1: lock the demo scope contract\n  Capture the result boundary, key implementation detail, and success signal before execution arrangement.\n\n## Non-goals\n- Do not expand beyond the demo contract.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => do work (skill: $normal-flow)\n\n## Execution Steps\n1. This section should not exist. (skill: $normal-flow)\n\n## Acceptance & Validation\n- Demo acceptance for contract replay.\n\n## Assumptions/Defaults\n- 示例仅用于回放验证：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/skill-binding-mixed-dollar-and-plain.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Mixed Skill Binding\n\n## Summary\n- 结论：用于验证混合 skill refs（部分缺少 `$` 前缀）会被拦截。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n\n## Goals\n- Goal 1: lock the demo scope contract\n  Capture the result boundary, key implementation detail, and success signal before execution arrangement.\n\n## Non-goals\n- Do not expand beyond the demo contract.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => execute local validation tasks (skill: $normal-flow, topic-spec)\n\n## Acceptance & Validation\n- Demo acceptance for contract replay.\n\n## Assumptions/Defaults\n- 示例仅用于回放验证：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/skill-binding-no-dollar-prefix.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Skill Binding Without Dollar Prefix\n\n## Summary\n- 结论：用于验证 skill 绑定未使用 `$` 前缀会被拦截。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n\n## Goals\n- Goal 1: lock the demo scope contract\n  Capture the result boundary, key implementation detail, and success signal before execution arrangement.\n\n## Non-goals\n- Do not expand beyond the demo contract.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => execute local validation tasks (skill: normal-flow)\n\n## Acceptance & Validation\n- Demo acceptance for contract replay.\n\n## Assumptions/Defaults\n- 示例仅用于回放验证：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/summary-blank-bullet.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按普通流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Summary Blank Bullet\n\n## Summary\n-\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n\n## Goals\n- Goal 1: lock the demo scope contract\n  Capture the result boundary and success signal before execution arrangement.\n\n## Non-goals\n- Do not expand beyond the demo contract.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => run local validation only (skill: $normal-flow)\n\n## Acceptance & Validation\n- Replay fixture must fail before execution starts.\n\n## Assumptions/Defaults\n- 该样例专门用于验证空 summary bullet：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/unrelated-flow-wording-default-fast-track.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"请先解释一下 flow_type 锁定策略，再直接给我一个计划。"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Unrelated Flow-type Wording Wrong Default\n\n## Summary\n- 结论：用于验证与 flow-type / routing 相关但不构成锁定的 wording 不能静默落到 default_fast_track。\n\n## Flow Decision\n- `flow_type`: `fast-track`\n- `flow_source`: `default_fast_track`\n- `lock_status`: `locked`\n\n## Goals\n- 目标 1: 验证无关 flow-type wording 的默认阻断\n  当前请求提到的是 `flow_type` 锁定策略，而不是要把本次 run 锁到某个具体 flow，因此 assistant 应继续追问，而不是直接按默认快车道规划。\n\n## Non-goals\n- 不把与 `flow_type` / routing 相关但不构成锁定的 wording 当成 happy path。\n\n## Task Orchestration\n- wave: 1\n  - main-agent => 规划快车道执行路径 (skill: $fast-flow)\n\n## Acceptance & Validation\n- checker 必须拒绝在相关 wording 存在但未真正锁 flow 时直接使用 `default_fast_track`。\n\n## Assumptions/Defaults\n- 该样例专门用于验证 unrelated flow-type wording guard：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/unrelated-flow-wording.jsonl

@@ -0,0 +1,3 @@
+{"type":"response_item","payload":{"type":"function_call","name":"request_user_input","call_id":"call_detached_only","arguments":"{\"questions\":[{\"header\":\"Detached策略\",\"id\":\"detached_policy\",\"question\":\"当 normal-flow 检测到 detached HEAD 时选哪种？\",\"options\":[{\"label\":\"阻断并询问 (Recommended)\",\"description\":\"提交前阻断并询问\"},{\"label\":\"自动建分支\",\"description\":\"自动创建分支并继续\"},{\"label\":\"仅告警不阻断\",\"description\":\"只提示风险\"}]}]}"}}
+{"type":"response_item","payload":{"type":"function_call_output","call_id":"call_detached_only","output":"{\"answers\":{\"detached_policy\":{\"answers\":[\"自动建分支\"]}}}"}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Demo Unrelated Flow Wording\n\n## Summary\n- 结论：问了流程相关问题，但不是 flow_type 锁定问题。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `fallback_default`\n- `lock_status`: `pending_confirmation`\n\n## Goals\n- Goal 1: lock the demo scope contract\n  Capture the result boundary, key implementation detail, and success signal before execution arrangement.\n\n## Non-goals\n- Do not expand beyond the demo contract.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => continue with default normal path (skill: $normal-flow)\n\n## Acceptance & Validation\n- Demo acceptance for contract replay.\n\n## Assumptions/Defaults\n- 主人尚未明确流程类型：待确认\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/update-plan-md-implementation-steps-artifact.jsonl

@@ -0,0 +1,2 @@
+{"type":"response_item","payload":{"type":"message","role":"user","content":[{"type":"input_text","text":"按普通流程计划"}]}}
+{"type":"response_item","payload":{"type":"message","role":"assistant","content":[{"type":"output_text","text":"<proposed_plan>\n# Update PLAN.md with Implementation Steps\n\n## Summary\n- 结论：用于验证 generic PLAN.md update wording 仍会被当成 persisted planning artifact。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n\n## Goals\n- Goal 1: refresh a persisted plan file\n  Update PLAN.md with implementation steps for the next execution pass.\n\n## Non-goals\n- Do not change runtime behavior.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => update PLAN.md with implementation steps (skill: $normal-flow)\n\n## Acceptance & Validation\n- The persisted plan file reflects the latest implementation steps.\n\n## Assumptions/Defaults\n- PLAN.md stays as the execution source: 已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/wave-sequence-jump.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按普通流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Demo Wave Jump\n\n## Summary\n- 结论：用于验证 wave 跳号。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n\n## Goals\n- Goal 1: lock the demo scope contract\n  Capture the result boundary, key implementation detail, and success signal before execution arrangement.\n\n## Non-goals\n- Do not expand beyond the demo contract.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => first task (skill: $normal-flow)\n- wave: 3\n  - main-agent => second task (skill: $normal-flow)\n\n## Acceptance & Validation\n- Demo acceptance for contract replay.\n\n## Assumptions/Defaults\n- 示例仅用于回放验证：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/references/replay-cases/wave-sequence-repeat.jsonl

@@ -0,0 +1,2 @@
+{"type": "response_item", "payload": {"type": "message", "role": "user", "content": [{"type": "input_text", "text": "按普通流程计划"}]}}
+{"type": "response_item", "payload": {"type": "message", "role": "assistant", "content": [{"type": "output_text", "text": "<proposed_plan>\n# Demo Wave Repeat\n\n## Summary\n- 结论：用于验证 wave 重号。\n\n## Flow Decision\n- `flow_type`: `normal`\n- `flow_source`: `user_explicit`\n- `lock_status`: `locked`\n\n## Goals\n- Goal 1: lock the demo scope contract\n  Capture the result boundary, key implementation detail, and success signal before execution arrangement.\n\n## Non-goals\n- Do not expand beyond the demo contract.\n\n## Task Orchestration\n- wave: 1\n  - main-agent => first task (skill: $normal-flow)\n- wave: 1\n  - main-agent => second task (skill: $normal-flow)\n\n## Acceptance & Validation\n- Demo acceptance for contract replay.\n\n## Assumptions/Defaults\n- 示例仅用于回放验证：已确定\n</proposed_plan>"}]}}

package/toolkit/skills/dev-plan/scripts/test_migrate_task_orchestration.ts

@@ -0,0 +1 @@
+import "../../bin/tests/plan/task_migration.test.ts";

package/toolkit/skills/docs/vendor/dot-codex/README.md

@@ -0,0 +1,36 @@
+# Docs 总览
+
+`docs/` 用来承载**项目文档**。这里的主视角是“项目当前是什么、怎么理解、怎么维护”，默认面向人类阅读；具体路径、层级与拆分方式应尊重项目自身情况，而不是强推一套全局目录骨架。
+
+## 阅读入口
+
+- [工程流程指南](engineering-workflow.md)：执行流程、开发验证、review、合并发布和流程恢复。
+- [主题规格](specs/README.md)：按主题查阅长期需求、实现状态与演进原因。
+- [工程记忆](solutions/README.md)：查阅可复用的问题诊断和解决经验。
+- [多任务流程设计](design/multi-task-flow/orchestration.md)：`multi-task-flow` 顶层流程的执行、数据和页面设计。
+
+## 文档角色
+
+- `docs/specs/<topic>/`：面向工程流程的 slug-only topic-level specification；`SPEC.md` 回答“这个稳定主题长期应该保持什么为真”，同目录的 `IMPLEMENTATION.md` / `HISTORY.md` 保存实现状态与主题局部演进。历史 ID-prefixed 目录继续可读。
+- `docs/adr/<NNNN>-<slug>.md`：面向难逆转、非显然且存在真实取舍的架构或边界决定，回答“为什么这样选”；它可被 topic Spec 引用，但不替代当前规范。
+- `docs/solutions/**`：面向 Agent 的可复用工程记忆，回答“下次遇到类似问题时不要再从零想一遍”。
+- `docs/**` 里的其他项目文档：面向人类的项目真相，回答“这个项目现在是什么、如何使用/维护”。
+
+## 组织原则
+
+- 优先复用项目已有的人类文档结构与命名。
+- 需要新文档时，优先放到最贴近现有主题的位置；不要为了对齐全局约定硬造目录。
+- 若项目还没有清晰的人类文档入口，可以补一个最小索引，但入口文件名也应服从项目现有习惯。
+
+## Promotion 规则
+
+- `specs` 与 `solutions` 可以更细碎，因为它们承担工程记忆与流程对齐职责。
+- 当一次变更形成了**稳定的项目级当前真相**，或后续人类维护者需要在不读 spec/solution 的情况下直接理解它，就应把结论整理到合适的项目文档中。
+- 这类整理默认通过 `$project-doc-promote` 进行；历史文档或批量收敛场景则走 `$docs-migrate MODE=project-docs`。
+
+## 当前仓库
+
+- 本仓当前显式维护的工程记忆入口为 `docs/specs/`、`docs/adr/` 与 `docs/solutions/`。
+- `docs/specs/` 现在是主题级规格目录；它不再把整份 `SPEC.md` 写成任务单。持久决策的完整取舍留在 ADR，topic 局部生命周期仍留在 companion docs。
+- Initiative v1 的 tracker 入口是 `docs/agents/issue-tracker.md`，只支持 GitHub + `gh`。
+- 后续若新增人类项目文档，应继续遵守“路径尊重项目自身情况、内容表达当前真相”的原则。

package/toolkit/skills/docs/vendor/dot-codex/adr/0001-compose-adrs-with-topic-specs.md

@@ -0,0 +1,7 @@
+---
+status: accepted
+---
+
+# ADR 与 Topic Spec 的职责分离
+
+工程工作流把 ADR 与 Topic Spec 视为可组合、但不互相替代的决策记录：ADR 仅用于难以逆转、有真实取舍且未来读者可能不理解的架构或边界决策，记录为何选择；`SPEC.md` 记录当前持久行为、接口和验收，`IMPLEMENTATION.md` 记录实现覆盖，`HISTORY.md` 仅记录该 Spec 的局部演进、替换、兼容性与必要背景。这样让少量稳定的“为何”与持续演进的“当前合同”分别拥有唯一事实源，避免把 ADR 写成规格或把 History 写成决策档案。

package/toolkit/skills/docs/vendor/dot-codex/adr/0002-xp-infra-public-hostname-convention.md

@@ -0,0 +1,11 @@
+---
+status: accepted
+---
+
+# XP Infra 公网 hostname 采用成对严格命名
+
+标准公网 XP 节点的控制面和托管 endpoint 分别使用 `<node>-xp.707979.xyz` 与 `<node>-ep.707979.xyz`。其中 `<node>` 是节点运行时声明的原始小写 DNS label，最多 60 个字符，以便角色后缀仍留在同一个有效 DNS label 内；不能为了 DNS 或 Tunnel 冲突添加随机或数字后缀。
+
+选择严格审计而不是把现有 hostname 视为节点名称的一部分，是为了让节点身份、控制面入口和 endpoint 入口保持可预测的一对一关系。冲突应停下来由操作员解决；静默生成后缀或覆写现有记录会把错误身份写入集群元数据。
+
+已有节点若不符合该规则，`xp-infra audit` 必须报告 drift，但不得自动变更 DNS、 Tunnel 或节点配置。其迁移需要单独的主人授权。