最新代码验证：host tests.test_evidence tests.test_task_recovery 80/80 通过，bash -n scripts/task.sh && make test-governance 491 项通过、18 项 skip，测试对象为 105901595805cc5fa5ce19f1a6d48317b950d6af。此前 CI green head 尚早于这些离线恢复修复，须以最终 PR head 重跑 GATE。
最新验证覆盖 80 个 evidence/recovery 测试与 491 项治理测试（18 项 skip），以上命令都成功；测试对象为当前实现代码 105901595805cc5fa5ce19f1a6d48317b950d6af。若后续只改任务卡，代码测试可复用但 CI head 会变，须在最终提交 head 重新等待 GATE。
Why: establish the authorized task workspace and its initial baseline.
Evidence: preserved precheck receipt in .task/feature-norm-remediation-u2.md; R0=b5e23daed2f0ea46261979de1d62e4de0cfa88f9.
Impact: task metadata only; the feature workspace starts at 01d9f50ee8.
Refs: .task/feature-norm-remediation-u2.md
Why: Specify the authorized continuation boundary before implementation begins.
Evidence: Task card records U2 files, hard stops, and acceptance criteria from the approved plan.
Impact: New-task remote triage and formal delivery gates remain unchanged.
Refs: Development Norm Remediation U2.
Why: Preserve the exact failure provenance and prepare a clean recovery record.
Evidence: The original agent-exec receipt shows overlapping edits in one Cursor session, terminal blocked, no transport, and no host fallback.
Impact: The next U2 work will use new run IDs and single-writer, short task packets.
Refs: Development Norm Remediation U2 recovery.
Why: include the current parent initiative baseline before verification.
Evidence: resolved feature/norm-remediation=74bd05a876136ed8506332c76d535a6eb6356433.
Impact: integrate parent initiative changes without rewriting task history.
Refs: task-sync
Why: Preserve the validated LOCAL_READY resume receipt before starting a worker.
Evidence: make precheck returned LOCAL_READY; make task-resume imported the receipt.
Impact: SELF-only task metadata; delivery still requires authoritative sync and checks.
Refs: feature/norm-remediation-u2
Why: Preserve worker transport outcomes and host verification evidence for U2.
Evidence: 66 focused evidence and task-recovery tests passed on the recorded HEAD.
Impact: Records SELF verification and the observed worker edit-serialization deviation; formal review and CI remain NOT_RUN.
Refs: feature/norm-remediation-u2
Why: Preserve the final synchronized candidate's local verification receipt before closing its task card.
Evidence: python3 -m unittest tests.test_evidence tests.test_task_recovery passed 66 tests on the synchronized candidate.
Impact: SELF evidence only; independent review and CI remain required for formal delivery.
Refs: feature/norm-remediation-u2
Why: close the task card inside its own PR so landing it needs no post-merge backfill or separate close PR.
Evidence: .task/feature-norm-remediation-u2.md -> .task/archive/feature-norm-remediation-u2.md; status 已完成（合入前闭合）; PR #618https://git-home.u1680832.nyat.app:30319/weichen/Agent-System-Dev/pulls/618.
Impact: task metadata only; the card is archived within PR #618 and lands with its merge commit.
Refs: task.sh close --self
Why: Capture the final offline recovery cases and host verification on the PR candidate.
Evidence: 68 focused evidence and task-recovery tests pass; two broader task-start failures reproduce on the parent baseline.
Impact: Local evidence remains SELF; exact-head CI and independent review are still required.
Refs: U2 / PR #618
Why: Preserve the closed-task and remote-target rewind corrections with their regression and governance test evidence.
Evidence: 71 focused tests and all 482 governance tests passed; the previous exact-head CI failed in an unrelated backend runtime test and requires a fresh run.
Impact: Review findings are fixed in scope; independent final review and current-head CI remain required.
Refs: U2 / PR #618
Why: Preserve the final test matrix and review corrections for the U2 candidate.
Evidence: All 484 governance tests pass (18 skipped); the targeted evidence and recovery cases pass on the same implementation.
Impact: SELF evidence only; final-head CI and independent review remain required.
Refs: U2 / PR #618
Why: Preserve the final guard for a reachable advertised target whose commit object cannot be compared locally.
Evidence: The regression test, 74 focused evidence/recovery tests, and all 485 governance tests pass.
Impact: Network outages retain LOCAL_READY; known but unverified target history returns REMOTE_UNVERIFIED. Exact-head CI remains pending.
Refs: U2 / PR #618
Why: Preserve host verification and worker transport records for offline root and child task resume.
Evidence: 79 evidence/recovery tests passed; governance validation is being rerun on the final implementation candidate.
Impact: LOCAL_READY remains local-only; exact-head CI and final independent review remain required.
Refs: U2 / PR #618
Why: Preserve final root/child offline resume behavior and its reachable-parent-missing guard.
Evidence: 80 evidence/recovery tests and all 491 governance tests pass (18 expected skips) on the code candidate.
Impact: Local resume remains bounded; known missing remote parent refuses, network outage may use the registered local parent. Final-head CI remains required.
Refs: U2 / PR #618
Why: Preserve the final network-error classification, child-parent identity gate, and full governance test result.
Evidence: 83 focused tests and all 494 governance tests passed (18 skips) on the code candidate; child outage/missing-parent contrast tests passed.
Impact: New task and formal delivery gates remain strict; current PR head CI and final review are still required.
Refs: U2 / PR #618
Why: Record the cached-target outage fix, host review, and final local verification for U2.\nEvidence: SELF receipts cover 84 focused tests and 495 governance tests; 18 checks skipped.\nImpact: Preserve exact-head delivery traceability without changing implementation.\nRefs: PR #618
weichen
merged commit fe0f1cd464 into feature/norm-remediation2026-10-18 16:43:37 +00:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
feature/norm-remediation-u2 任务卡
任务主记录：PR #618 https://git-home.u1680832.nyat.app:30319/weichen/Agent-System-Dev/pulls/618；状态：已完成（合入前闭合）；owner：weichen
权威仓：ssh://git@git-ssh.u1680832.nyat.app:31411/weichen/Agent-System-Dev.git；目标分支：main；R0（initial_base_sha）=
b5e23daed2工作区：/home/weichen/projects/Agent-System-Dev-wt-feature-norm-remediation-u2（worktree）；分支：feature/norm-remediation-u2
分支类型：child；initiative_id：feature-norm-remediation；root_target：main；immediate_target：feature/norm-remediation；accepting_requirements：false；parent_branch：feature/norm-remediation；parent_head_sha：01d9f50ee8aba1c40cb9b1769598698c91c30c4b；parent_pr：无；parent_source：origin
需求分流：免（父 initiative 子任务登记已列出 feature/norm-remediation-u2；parent_head_sha=01d9f50ee8aba1c40cb9b1769598698c91c30c4b）
授权与交付终点：用户于 2026-10-18 授权按规范整改计划实施 U2。仅允许安全既有任务的本地续作，并保留新任务启动及正式交付门禁。未授权扩展：改平台保护、host fallback、worker 权限、凭据轮换或其他 U-ID。
交付模式：goal-first（目标、范围和交付终点一次确认；范围内连续推进）
实施执行：agent-per-unit（配置：AGENT-EXECUTION.md）
合入执行：agent（已授权交付范围内，Review 通过、冲突解决且当前 HEAD 的 CI 通过后合入）
流程依据：仓库根目录 policy/delivery.md（归档后仍按仓库根目录定位）
目标与非目标
目标：当远端暂时不可用或落后时，已有且仍获授权的任务只有在身份、记录基线、本地工作区和请求范围均可核验时，才能进入
LOCAL_READY继续有界本地工作；本地收据不得被解释为远端新鲜或正式交付证据。范围：
scripts/task.sh、scripts/evidence.py、policy/governance.md、tests/test_task.py、tests/test_task_recovery.py、tests/test_evidence.py。测试可按现有代码模式作必要的最小补充，但不得改动task-pr-check、task-sync、CI 或平台准入语义。硬停止：关闭/放弃任务、授权或身份缺失/冲突、请求超范围、R0 本地对象缺失、dirty/共享工作区状态不安全、活动 Git 操作、未知外部副作用或未对账状态均拒绝续作。
验收：有效既有任务可在远端失败时继续本地工作；新任务仍需新鲜远端 triage；
LOCAL_READY不能成为READY、REPLAY、GATE或 merge check 的依据；远端恢复后 Review/merge 仍需同步权威当前目标。Agent 实施收据
路由与 fail-stop 规则见根目录 AGENT-EXECUTION.md 与 policy/delivery.md；每次尝试一行，原始日志留在源码外收据路径。
30a582372bd6df427e39210bb031d4dcbd36656ba43aa171eaeb6be71cfcd2986f1f5f74e4b06e5ed5b8e74c4305a9a3bc892e0da0dadb95d17188589c0e765708fbb2aacb718d8585afecf81326bd081ff6444d335b5537aa0f10590159588d8585afecf81326bd08bc304111a4471948c59c1ff6444d33同步记录
b5e23daed201d9f50ee8证据（Evidence）
python3 -m unittest tests.test_evidence tests.test_task_recovery210bb031d4python3 -m unittest tests.test_evidence tests.test_task_recoveryd2986f1f5fpython3 -m unittest tests.test_task tests.test_task_recovery tests.test_evidenced2986f1f5fpython3 -m unittest tests.test_evidence tests.test_task_recovery05a9a3bc89make test-governance05a9a3bc89make test-governance2e0da0dadbbash -n scripts/task.sh && make test-governance9c0e765708python3 -m unittest tests.test_evidence tests.test_task_recovery1ff6444d33bash -n scripts/task.sh && make test-governance1ff6444d33python3 -m unittest tests.test_evidence tests.test_task_recovery471948c59cbash -n scripts/task.sh && make test-governance471948c59c结果与验证
LOCAL_READY。新任务和正式交付仍要求严格 READY/权威同步。python3 -m unittest tests.test_evidence tests.test_task_recovery79/79 通过（SELF，HEAD1ff6444d3352c7e13c2b75cefe7d65c631181454），覆盖 root/child 离线 task-resume 与严格 delivery 拒绝；hostbash -n scripts/task.sh && make test-governance490 项通过、18 项 skip（SELF，同 HEAD）。不含后续主脚本改动的 transport。local precheck no-cache 的 2 个红灯回归已由紧随其后的1ff6444d修复，三条 no-cache 状态单测 3/3 通过；最终候选完整测试仍待重跑。TestToolCallDeltaBeforeExecution失败，后续b492bea...CI 通过；当前最终 head 的 GATE 与新候选完整 Review 待执行。tests/test_task_recovery.py的两个 edit 调用重叠约 449 ms，并被 fail-stop；新的 evidence 单元明确要求串行编辑，但 worker 仍在同一批次向scripts/evidence.py发出两次编辑调用（未重叠，最终 diff 已审阅且验证通过）。不能据此声称提示已强制串行；后续规范计划须区分提示、可观测证据与工具级并发防护。scripts/evidence.py和scripts/task.sh发出两次互不重叠编辑，违反了任务包的串行编辑要求；最终 diff 已逐项审阅，定向与完整治理测试均通过，但该结果不能证明提示具备工具级强制力。cmd_resume的 fetch-first 缺口，均已补 test-first 回归与修复。root 离线 resume 使用登记 R0；child resume 仅使用已登记、包含parent_head_sha的本地 parent 作为局部参考，delivery 仍 fetch-first。最终候选完整 Review 与 exact-head CI 待完成。parent_head_sha。tests.test_evidence tests.test_task_recovery80/80 通过，bash -n scripts/task.sh && make test-governance491 项通过、18 项 skip，测试对象为105901595805cc5fa5ce19f1a6d48317b950d6af。此前 CI green head 尚早于这些离线恢复修复，须以最终 PR head 重跑 GATE。105901595805cc5fa5ce19f1a6d48317b950d6af。若后续只改任务卡，代码测试可复用但 CI head 会变，须在最终提交 head 重新等待 GATE。parent_head_sha。delivery 仍保持 fetch-first。target_history仅将已识别的 DNS/连接类传输错误视为 outage；认证/权限/未知错误不再降级。无 tracking ref 时 precheck 以 active task R0 比较 advertised target；reachable target/parent 缺失或不可比均拒绝。child local resume 要求有效且本地可用的parent_head_sha。去掉了无 tracking-ref 时会触发 strict-shell 退出的 merged-branch 查询。471948c59cfabff36c018f219998139fea3757fe：host evidence+recovery 83/83 通过；bash -n scripts/task.sh && make test-governance494 项通过、18 项 skip。随后105901595805cc5fa5ce19f1a6d48317b950d6af又通过 targeted unreachable-vs-reachable missing-parent 对照测试；其后未再改生产逻辑。exact final-head CI 与当前候选 Review 尚待重跑。TASK_WORK_MODE=local make task-resume导入；child 仅使用本地已登记 parent 并核对parent_head_sha。同一环境下 delivery resume 仍以 REMOTE_UNVERIFIED 拒绝。--local-context与resume-local的有界验证。start与resume-delivery继续要求当前权威目标。/tmp/norm-remediation-u2-receipt.json，run_idaf20482c3e24cc32f3873d1451043481；其 worktree 保留为诊断证据。该 runblocked、transport_ref=null、host_fallback_authorizations=0。Development-Manager PR #26 已合入b178dec；Dev gitlink adoption 仍待 initiative 集成，不在本 U2 单元内。合入与清理
| 同步（记录前对象） |
74bd05a876|f65d0169f6| 2026-10-18T10:58:22Z |闭合记录（合入前折叠）