@elvis1513/auto-coding-skill 4.2.7 → 4.2.8
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +39 -9
- package/cli/assets/agents/reviewer.toml +2 -2
- package/cli/assets/managed-install.json +1 -1
- package/cli/assets/skill/SKILL.md +6 -7
- package/cli/assets/skill/data/policies/workflow-migrations-v1.json +2 -2
- package/cli/assets/skill/data/templates/ENGINEERING.md +2 -2
- package/cli/assets/skill/data/templates/bridges/AGENTS.md +13 -16
- package/cli/assets/skill/scripts/ap.py +1274 -91
- package/cli/tests/ap_concurrency_test.py +321 -33
- package/cli/tests/ap_profiles_test.py +326 -36
- package/cli/tests/index.test.js +23 -22
- package/package.json +1 -1
package/README.md
CHANGED
|
@@ -9,6 +9,10 @@ The Skill is a selectable guardrail, not a command sequence that must run for
|
|
|
9
9
|
every task. The model skips machinery whose expected benefit does not exceed its
|
|
10
10
|
cost; read-only work and obvious small clean-checkout changes normally stay direct.
|
|
11
11
|
|
|
12
|
+
Version 4.2.8 makes the supervised Reviewer observable and recoverable without
|
|
13
|
+
weakening review semantics: it streams private event metadata, distinguishes
|
|
14
|
+
startup failure from analysis timeout, retries one no-event startup, and supports
|
|
15
|
+
an audited fingerprint-bound user override that is never reported as approval.
|
|
12
16
|
Version 4.2.7 freezes every pre-commit Reviewer diff as a private, SHA-256-bound
|
|
13
17
|
Git-local patch, including staged, unstaged, untracked, deleted, mode, symlink,
|
|
14
18
|
and binary changes. Version 4.2.6 makes the managed structure/optimization description defer to
|
|
@@ -236,16 +240,15 @@ never triggers a broader fallback.
|
|
|
236
240
|
- Parallel fixers require separate worktrees and non-overlapping paths.
|
|
237
241
|
- Reviewer fingerprint approval is required for high-risk, cross-module,
|
|
238
242
|
explicitly configured, or parallel integration work.
|
|
239
|
-
-
|
|
240
|
-
|
|
241
|
-
`blocked`; do not call `task-review` for that result.
|
|
243
|
+
- Focused review uses `high` effort and a 150-second hard wall over the supplied
|
|
244
|
+
stable diff and evidence. Deep review uses `xhigh` and 360 seconds.
|
|
242
245
|
- Cross-module or parallel work and API, auth, database, payment, file-transfer,
|
|
243
|
-
gateway, or production-configuration boundaries use
|
|
246
|
+
gateway, or production-configuration boundaries use deep review.
|
|
244
247
|
Fixing Reviewer JSON formatting reuses the same analysis and never starts
|
|
245
248
|
another substantive review.
|
|
246
249
|
- `review-assignment` writes the validated assignment under Git-local state with
|
|
247
250
|
the exact fingerprint, HEAD, scope revision, owning fixer, review depth, and
|
|
248
|
-
|
|
251
|
+
150/360-second deadline. Before that deadline begins, it freezes the exact
|
|
249
252
|
task-owned working-tree state into a mode-0600 binary patch and binds its
|
|
250
253
|
canonical path, format, and SHA-256 into the assignment. Reissuing the same
|
|
251
254
|
live assignment is idempotent. Snapshot inputs and emitted patches above
|
|
@@ -255,11 +258,20 @@ never triggers a broader fallback.
|
|
|
255
258
|
rejected by `task-review`.
|
|
256
259
|
- `review-run` is the default executable path: it creates/reuses that assignment,
|
|
257
260
|
starts a separate `codex exec` session in read-only mode, removes the parent
|
|
258
|
-
lifecycle identity,
|
|
259
|
-
|
|
260
|
-
|
|
261
|
+
lifecycle identity, consumes JSON events incrementally, and supervises the full
|
|
262
|
+
process group against the original remaining deadline. If no semantic event
|
|
263
|
+
arrives within 30 seconds it retries the same immutable assignment once. After
|
|
264
|
+
analysis starts it never retries. Private schema-2 receipts record only bounded,
|
|
265
|
+
allowlisted event/diagnostic categories, byte counts, and hashes. Plain `review-assignment` remains
|
|
261
266
|
available only when another host can enforce the deadline itself; it does not
|
|
262
|
-
stop an in-app subagent.
|
|
267
|
+
stop an in-app subagent. A descendant that deliberately creates a new POSIX
|
|
268
|
+
session is outside portable process-group termination, but cannot keep the
|
|
269
|
+
non-blocking event readers or Reviewer deadline open.
|
|
270
|
+
- Exhausted `runtime-unavailable` or `analysis-timed-out` states remain blocked by
|
|
271
|
+
default. `review-runtime-override` can record explicit user authorization bound
|
|
272
|
+
to the exact fingerprint, assignment, artifact, and runtime receipt. Its verdict
|
|
273
|
+
is `runtime-bypassed`, never `approved`; real blocked/changes-requested results
|
|
274
|
+
cannot use this path.
|
|
263
275
|
- DD/ADR is created only for lasting cross-module, API, data, security,
|
|
264
276
|
deployment, or key user-flow decisions.
|
|
265
277
|
- Historical debt does not block product work unless the current change worsens it.
|
|
@@ -308,6 +320,24 @@ model overrides, complete project `risk.rules`, supported project/access/
|
|
|
308
320
|
concurrency/route/structure values, and an existing project-owned structure
|
|
309
321
|
standard byte-for-byte. Removed content is archived outside active docs.
|
|
310
322
|
|
|
323
|
+
## What changed in 4.2.8
|
|
324
|
+
|
|
325
|
+
- Replaced end-only Reviewer output collection with bounded JSONL event streaming.
|
|
326
|
+
Schema-2 Git-local receipts now record CLI/model/effort identity, attempt phases,
|
|
327
|
+
first/last event times, event-log hashes, diagnostic categories, and byte counts
|
|
328
|
+
without persisting prompt, patch, stderr, tool output, or model content.
|
|
329
|
+
- Split runtime startup failure from analysis timeout. A no-semantic-event startup
|
|
330
|
+
is capped at 30 seconds and retried once with the same assignment, fingerprint,
|
|
331
|
+
artifact, command, and absolute deadline; an analysis that started is never
|
|
332
|
+
repeated. Focused review now uses `high/150s`, while deep uses `xhigh/360s`.
|
|
333
|
+
- Added `review-runtime-override` for explicit user-authorized delivery after an
|
|
334
|
+
exhausted runtime-unavailable or analysis-timeout failure. The private override
|
|
335
|
+
binds all immutable review evidence and records `runtime-bypassed`, never
|
|
336
|
+
`approved`; failure-time hashes reject pre-authorization tampering, and a
|
|
337
|
+
substantive Reviewer result observed before timeout remains non-bypassable.
|
|
338
|
+
- Restricted the hidden custom runner to explicit test harnesses and closed spawn
|
|
339
|
+
failure, unbounded-output, silent-startup, and lost-diagnostics paths.
|
|
340
|
+
|
|
311
341
|
## What changed in 4.2.7
|
|
312
342
|
|
|
313
343
|
- Replaced live pre-commit diff discovery with an immutable Git-local patch
|
|
@@ -8,8 +8,8 @@ developer_instructions = """
|
|
|
8
8
|
|
|
9
9
|
工作要求:
|
|
10
10
|
1. 只审查主代理给定的 task_id、base_sha、diff_fingerprint、diff_artifact_path、diff_artifact_sha256 和 owning_fixer;开始前必须运行 `python3 docs/tools/autopipeline/ap.py review-artifact --file <assignment.json>`,以验证 0600 权限和 SHA-256 并读取冻结补丁。不得用实时 `git diff` 或 `diff_base..diff_head` 替代该 artifact。
|
|
11
|
-
2. 若主代理指定 focused review,只对给定的稳定 diff 和相关 evidence
|
|
12
|
-
3. 跨模块、并行 writer,以及 API、鉴权、数据库、支付、文件传输、网关、生产配置边界使用 deep review
|
|
11
|
+
2. 若主代理指定 focused review,只对给定的稳定 diff 和相关 evidence 做一次实质分析,不扩展为全库扫描。focused 运行使用 high effort 和 150 秒硬截止;若你在终止前已发现无法继续,则返回 `status=blocked`、`verdict=blocked`。
|
|
12
|
+
3. 跨模块、并行 writer,以及 API、鉴权、数据库、支付、文件传输、网关、生产配置边界使用 deep review,采用 xhigh effort 和 360 秒硬截止;仍然只审查主代理交付的范围。运行器仅可在 30 秒内没有语义事件时复用同一 assignment 自动重启一次;一旦分析已开始不得重试实质审查。
|
|
13
13
|
4. 优先检查:正确性、安全性、边界条件、异常路径、并发风险、事务一致性、缓存一致性、兼容性。
|
|
14
14
|
5. 结构/优化 review 仅在已有 baseline/backlog 时读取;可选文档缺失不构成 blocker,不重复报告 accepted debt。
|
|
15
15
|
6. 检查已验证的冻结 diff artifact 和已执行验证是否一致;该 artifact 统一包含 staged、unstaged、untracked、删除、模式变化和二进制变更。普通任务不要求 taskbook、closure 或 evidence 文档。
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"schema_version":1,"skill_version":"4.2.
|
|
1
|
+
{"schema_version":1,"skill_version":"4.2.8","manifest_path":".agents/managed-install.json","entries":[{"path":".agents/agents/browser-debugger.toml","source":"agents/browser-debugger.toml","ownership":"normalized-agent","sha256":"40727952944917eb0b573bb96dca2906c27bc794737f3b6b392b56df7b4208aa","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/agents/docs-researcher.toml","source":"agents/docs-researcher.toml","ownership":"normalized-agent","sha256":"aa1fed6a6a1cde282fc152c7a88b8e097a6c53bcf447d35978607d594e3cb3e4","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/agents/explorer.toml","source":"agents/explorer.toml","ownership":"normalized-agent","sha256":"4cd89888acf7c0c04fbbcd0b9a0f346c6efed42fb723ec2037af9331a84a1cac","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/agents/fixer.toml","source":"agents/fixer.toml","ownership":"normalized-agent","sha256":"7bc00a45b9d1bcc90b8b0f06d773556eba1d52effff664472c3cb81073f6abfb","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/agents/reviewer.toml","source":"agents/reviewer.toml","ownership":"normalized-agent","sha256":"e1ad676e608ffee04d782571fb9bfbc787f69c100941b1811d955ef80439db36","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/SKILL.md","source":"skill/SKILL.md","ownership":"exact","sha256":"b573dcb66a1d3d084a413c8cec32d701f3dc07aec956336315a400e54b54b342","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/data/contracts/orchestration-v1.schema.json","source":"skill/data/contracts/orchestration-v1.schema.json","ownership":"exact","sha256":"65a8c0d222b26f224794ec106101f903f0e3a357b249c2df22c5908a476a6835","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/data/policies/workflow-migrations-v1.json","source":"skill/data/policies/workflow-migrations-v1.json","ownership":"exact","sha256":"5912f6c1807f90fa2f6e8984eb0dd2465c4c55667d0bd7f915dab2fa795ce3b2","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/data/templates/ENGINEERING.md","source":"skill/data/templates/ENGINEERING.md","ownership":"exact","sha256":"750407c5ed571d799e759389dc7c954319e93b07529e4f0fe97b4113a6b285a4","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/data/templates/bridges/AGENTS.md","source":"skill/data/templates/bridges/AGENTS.md","ownership":"exact","sha256":"84ffa96cbafe5c70f984d65af0e292249893bde5fb0faec0fec606282bed6a79","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/data/templates/tools/ap.py","source":"skill/data/templates/tools/ap.py","ownership":"exact","sha256":"c352fbf2cbcb9692a34fd952d8541b1cbe6d475c6e11e64500a1f6b5283a00be","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/requirements.txt","source":"skill/requirements.txt","ownership":"exact","sha256":"0969da99a0bc2a1b71ed50584560f4588a37567ac63af3ddbaf3c4617ca5621a","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/scripts/ap.py","source":"skill/scripts/ap.py","ownership":"exact","sha256":"e5456e17ceaf51e47b90203a7b4331dd7a8eef21126627b385228897b34d03ae","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/scripts/core.py","source":"skill/scripts/core.py","ownership":"exact","sha256":"640c37e8b274d90d6b7765d4243e054360d5a5a85b3150fbfe6449527c1cd28f","executable":true,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/scripts/http_checks.py","source":"skill/scripts/http_checks.py","ownership":"exact","sha256":"5620e187a0129f135621425a84b3a4e8f31137263c7e549a36a801fc36ff54ea","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/scripts/install_integrity.py","source":"skill/scripts/install_integrity.py","ownership":"exact","sha256":"a9ca7bd67893bbd2d7b35ece30a40457f0cbcb9ae01580cfc5daca2cd96fd676","executable":false,"scope":"shared","version":"4.2.8"},{"path":".agents/skills/auto-coding-skill/scripts/scaffold_templates.py","source":"skill/scripts/scaffold_templates.py","ownership":"exact","sha256":"6f0ee92eff2d5db9627a64b332738c9c917a93cc559d5bbacf771c5966215821","executable":false,"scope":"shared","version":"4.2.8"},{"path":"AGENTS.md","source":"skill/data/templates/bridges/AGENTS.md","ownership":"exact","sha256":"84ffa96cbafe5c70f984d65af0e292249893bde5fb0faec0fec606282bed6a79","executable":false,"scope":"project","version":"4.2.8"},{"path":"docs/ENGINEERING.md","source":"skill/data/templates/ENGINEERING.md#managed-workflow","ownership":"managed-workflow","sha256":"bac3eeed72e3654fe8ec3f2c79e5fcabb968ee6374fa4971bbf041969455bc4d","executable":false,"scope":"project","version":"4.2.8"},{"path":"docs/architecture/adr/_TEMPLATE-ADR.md","source":"generated:scaffold_templates:docs/architecture/adr/_TEMPLATE-ADR.md","ownership":"exact","sha256":"88c89065ae0f8555412dbe096c05342152efc7becae8d8981e6442cdfee8e672","executable":false,"scope":"project","version":"4.2.8"},{"path":"docs/deployment/deploy-records/_TEMPLATE-DEPLOY-RECORD.md","source":"generated:scaffold_templates:docs/deployment/deploy-records/_TEMPLATE-DEPLOY-RECORD.md","ownership":"exact","sha256":"f2c17c0febc322b0ad50fb9d6f7e64e6c0cfb15c8a3e445a99da40f0a0dd1901","executable":false,"scope":"project","version":"4.2.8"},{"path":"docs/design/_TEMPLATE-DD.md","source":"generated:scaffold_templates:docs/design/_TEMPLATE-DD.md","ownership":"exact","sha256":"6ca022d127bfff68521f466b60bace4f1206baf0913daa855b9f751e74198db3","executable":false,"scope":"project","version":"4.2.8"},{"path":"docs/reviews/_TEMPLATE-REVIEW.md","source":"generated:scaffold_templates:docs/reviews/_TEMPLATE-REVIEW.md","ownership":"exact","sha256":"134f0d1bd3c7a317ef3cf4c2aea9f0539f3d53dccd5bcdabf6915b3629ef8802","executable":false,"scope":"project","version":"4.2.8"},{"path":"docs/tools/autopipeline/ap.py","source":"skill/data/templates/tools/ap.py","ownership":"exact","sha256":"c352fbf2cbcb9692a34fd952d8541b1cbe6d475c6e11e64500a1f6b5283a00be","executable":false,"scope":"project","version":"4.2.8"}],"managed_namespaces":[{"path":".agents/skills/auto-coding-skill","scope":"shared"},{"path":"docs/tools/autopipeline","scope":"project"}],"preserved":[".agents/archive/**",".agents/agents/* (except paths listed in entries)","docs/project/**","docs/architecture/structure-standard.md","docs/architecture/adr/[0-9]*.md","docs/design/*.md","docs/interfaces/*.md","docs/deployment/deploy-records/*.md","docs/reviews/*.md"]}
|
|
@@ -52,13 +52,12 @@ Each parallel writer uses a task ID, worktree, lease, dependency SHAs, and disti
|
|
|
52
52
|
with `agent-contract-check`; parallel fixers require reclassification with
|
|
53
53
|
`--writers >1`.
|
|
54
54
|
|
|
55
|
-
|
|
56
|
-
|
|
57
|
-
|
|
58
|
-
|
|
59
|
-
`review-
|
|
60
|
-
|
|
61
|
-
normalizes presentation fields from the same analysis.
|
|
55
|
+
Focused review uses `high`/150 seconds; deep review uses `xhigh`/360 seconds. `review-run` launches a separate read-only Codex process, streams allowlisted event metadata to a private Git-local log, stops the process group at deadline, and binds the result to the exact HEAD/scope/fingerprint.
|
|
56
|
+
No semantic event within 30 seconds retries the same assignment once; started analysis is never retried. `review-assignment` freezes every working-tree change into a mode-0600, assignment-bound, SHA-256 patch.
|
|
57
|
+
The Reviewer reads it with `review-artifact --file <assignment.json>`, never a reconstructed live diff. `review-assignment` alone needs another deadline-capable host; `agent-result-template` supplies all 16 fields.
|
|
58
|
+
|
|
59
|
+
After exhausted startup attempts or an analysis deadline, explicit user authorization may record a one-fingerprint `review-runtime-override` with identity, reference, reason, evidence, and `--confirm-runtime-bypass`.
|
|
60
|
+
It records `runtime-bypassed`, never `approved`, and cannot replace a substantive `blocked` or `changes-requested` result.
|
|
62
61
|
|
|
63
62
|
## Close with the bounded routed gate
|
|
64
63
|
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
---
|
|
2
2
|
workflow:
|
|
3
|
-
skill_version: "4.2.
|
|
3
|
+
skill_version: "4.2.8"
|
|
4
4
|
mode: "dev"
|
|
5
5
|
profile: "auto"
|
|
6
6
|
completion: "push"
|
|
@@ -83,7 +83,7 @@ docs:
|
|
|
83
83
|
api_change_log: "docs/interfaces/api-change-log.md"
|
|
84
84
|
---
|
|
85
85
|
# Engineering Configuration and Project Facts
|
|
86
|
-
<!-- auto-coding-skill:managed-workflow:start version=4.2.
|
|
86
|
+
<!-- auto-coding-skill:managed-workflow:start version=4.2.8 -->
|
|
87
87
|
Root `AGENTS.md` is the single shared behavioral protocol. The installed
|
|
88
88
|
`SKILL.md` contains invocation guidance. This file is the exact-schema source for
|
|
89
89
|
project workflow configuration and access values that neither file can infer.
|
|
@@ -1,4 +1,4 @@
|
|
|
1
|
-
<!-- auto-coding-skill:managed-agents:start version=4.2.
|
|
1
|
+
<!-- auto-coding-skill:managed-agents:start version=4.2.8 -->
|
|
2
2
|
# Shared Engineering Protocol
|
|
3
3
|
|
|
4
4
|
This file is fully managed by `auto-coding-skill`. Keep project-specific facts,
|
|
@@ -79,22 +79,19 @@ failure may continue in the same conversation/task without a second lifecycle.
|
|
|
79
79
|
- Validate delegated assignment/result JSON with `agent-contract-check` before use.
|
|
80
80
|
- Require fingerprinted review only for path/rule-confirmed high-risk,
|
|
81
81
|
contract-crossing, parallel, or configured work; intent words are candidates only.
|
|
82
|
-
-
|
|
83
|
-
|
|
84
|
-
|
|
82
|
+
- Use `high`/150 seconds for focused review and `xhigh`/360 seconds for deep parallel,
|
|
83
|
+
cross-module, or sensitive review. Retry once only if no semantic event arrives
|
|
84
|
+
within 30 seconds; never retry started analysis.
|
|
85
85
|
- Run `python3 docs/tools/autopipeline/ap.py review-run <TASK> --reviewer <ID> --json`.
|
|
86
|
-
It
|
|
87
|
-
|
|
88
|
-
|
|
89
|
-
deadline-capable host and cannot stop an in-app subagent.
|
|
90
|
-
- `agent-result-template` supplies all 16 fields
|
|
91
|
-
|
|
92
|
-
-
|
|
93
|
-
|
|
94
|
-
|
|
95
|
-
may receive a targeted recheck by the same reviewer.
|
|
96
|
-
- Historical debt does not block ordinary delivery; block only new or worsened
|
|
97
|
-
P0/P1 issues.
|
|
86
|
+
It runs a separate read-only identity, streams allowlisted metadata privately,
|
|
87
|
+
stops the process group at deadline, and binds a mode-0600 SHA-256 patch to the
|
|
88
|
+
assigned HEAD, scope, identity, and fingerprint. Reviewers read the frozen patch
|
|
89
|
+
with `review-artifact`; `review-assignment` alone needs a deadline-capable host and cannot stop an in-app subagent.
|
|
90
|
+
- `agent-result-template` supplies all 16 fields and `review-run` normalizes presentation.
|
|
91
|
+
- Runtime exhaustion needs explicit `review-runtime-override` authorization; it records
|
|
92
|
+
`runtime-bypassed`, never approval, and cannot replace a substantive blocked result.
|
|
93
|
+
- Block only new or worsened in-scope P0/P1 defects. Adjacent debt is a follow-up;
|
|
94
|
+
semantic changes invalidate approval, while mechanical docs may get a targeted recheck.
|
|
98
95
|
|
|
99
96
|
## Bounded real validation
|
|
100
97
|
|