@elvis1513/auto-coding-skill 4.1.1 → 4.1.2

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 CHANGED
@@ -9,7 +9,9 @@ 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 4.1.1 consolidates the 4.x delivery-first workflow into one canonical
12
+ Version 4.1.2 makes the workflow decision explicit: mechanisms are required,
13
+ model-selectable when beneficial, or forbidden for the current task. Version
14
+ 4.1.1 consolidated the 4.x delivery-first workflow into one canonical
13
15
  repository contract. Version 4.0 replaced the 3.x governance-first defaults with progressive
14
16
  guardrails. Clean single-writer work stays on the current branch. Worktrees,
15
17
  parallel fixers, fingerprint review, durable design records, and stronger
@@ -65,8 +67,11 @@ The result includes:
65
67
  - `execution_mode=isolated`: dirty checkout, explicit isolation, or multiple writers.
66
68
  - `execution_mode=none`: no task signal or change; create no branch/worktree.
67
69
  - `review_required` and `design_required`: risk-based escalation decisions.
70
+ - `task_kind`: `read_only`, `change`, `terminal_maintenance`, or `none`.
68
71
  - `mechanism_plan.required`: the complete minimum mechanism set for the task.
69
- - `mechanism_plan.not_required`: mechanisms that stay off unless the user asks.
72
+ - `mechanism_plan.optional_when_beneficial`: model-selectable mechanisms whose
73
+ expected benefit must exceed coordination cost.
74
+ - `mechanism_plan.forbidden`: mechanisms that stay off unless the user overrides.
70
75
  - an adaptive agent plan that does not force fan-out for ordinary work.
71
76
 
72
77
  Clean serial work proceeds directly on the current branch with normal Git; it does
@@ -75,7 +80,7 @@ requires isolation/review or the user explicitly requests lifecycle tracking:
75
80
 
76
81
  ```bash
77
82
  python3 docs/tools/autopipeline/ap.py task-start T0001 \
78
- --owned-path backend/internal/orders
83
+ --owned-path backend/internal/orders --force-lifecycle
79
84
  python3 docs/tools/autopipeline/ap.py commit-push T0001 \
80
85
  --msg "T0001: fix order retry"
81
86
  ```
@@ -90,8 +95,9 @@ Dirty or parallel work receives an isolated task branch/worktree:
90
95
  python3 docs/tools/autopipeline/ap.py task-start T0002 \
91
96
  --owned-path backend --review-required
92
97
  # implement in the printed worktree
98
+ python3 docs/tools/autopipeline/ap.py task-status T0002 --json
93
99
  python3 docs/tools/autopipeline/ap.py task-review T0002 \
94
- --verdict approved --diff-fingerprint "$SHA256"
100
+ --verdict approved --diff-fingerprint "$CURRENT_DIFF_FINGERPRINT"
95
101
  python3 docs/tools/autopipeline/ap.py commit-push T0002 \
96
102
  --msg "T0002: bounded change"
97
103
  python3 docs/tools/autopipeline/ap.py task-integrate T0002
@@ -122,6 +128,7 @@ validation:
122
128
  - name: backend
123
129
  paths: ["backend/**", "contracts/**"]
124
130
  commands: [backend_fast]
131
+ timeout_seconds: 90
125
132
  - name: frontend
126
133
  paths: ["frontend/**"]
127
134
  commands: [frontend_fast]
@@ -152,9 +159,10 @@ closure gate is limited to one stable-diff run. Full regression, Docker, builds,
152
159
  Jenkins, deployment, API verification, and target checks remain explicit
153
160
  diagnostics outside normal closure.
154
161
 
155
- The final closure gate is hard-bounded to at most 120 seconds per command and 180
156
- seconds total; projects may lower those values. A timeout requires a narrower
157
- route and never triggers a broader fallback.
162
+ The recommended final closure defaults are 120 seconds per command and 180
163
+ seconds total. Projects may raise them for measured affected-scope checks or set
164
+ a smaller `timeout_seconds` on a route. A timeout calls for a narrower route and
165
+ never triggers a broader fallback.
158
166
 
159
167
  ## Risk and subagent policy
160
168
 
@@ -192,11 +200,25 @@ project-specific facts outside the managed ENGINEERING block. Root `AGENTS.md` i
192
200
  replaced as a whole; its previous content is archived under
193
201
  `docs/archive/workflow/` as historical, non-authoritative context.
194
202
 
203
+ ## What changed in 4.1.2
204
+
205
+ - Added explicit read-only, change, and terminal-maintenance task kinds so
206
+ intent-only questions cannot accidentally trigger code-delivery machinery.
207
+ - Split mechanism planning into required, model-selectable, and forbidden sets;
208
+ read-only discovery agents remain available when they save time or add needed
209
+ expertise.
210
+ - Made 120/180-second gate budgets recommended defaults, added route-level
211
+ timeouts, and kept project-specific affected-scope overrides valid.
212
+ - Rejected unnecessary lifecycle creation before access checks, fetching, locks,
213
+ or branch work; corrected the review lifecycle examples.
214
+ - Added a release-package assertion and prepack cache cleanup so generated Python
215
+ caches cannot leak into the npm tarball.
216
+
195
217
  ## What changed in 4.1.1
196
218
 
197
219
  - Added a machine-readable minimum mechanism plan and rejected unnecessary clean
198
220
  serial task lifecycles unless explicitly requested.
199
- - Added hard per-command and total time ceilings for the final changed-scope gate.
221
+ - Added bounded per-command and total time budgets for the final changed-scope gate.
200
222
  - Made root `AGENTS.md` the sole behavioral protocol, reduced `SKILL.md` to
201
223
  invocation guidance, and reduced `ENGINEERING.md` to project configuration and
202
224
  facts.
@@ -1,16 +1,15 @@
1
1
  ---
2
2
  name: auto-coding-skill
3
- description: Delivery-first engineering workflow for repository changes. Use for analysis, task decomposition, necessary design, implementation, bounded changed-scope validation, safe commit/push, adaptive Git isolation, parallel writer coordination, and temporary branch cleanup. Keep clean serial work on the current branch and add lifecycle, worktrees, subagents, design, or review only when classify marks them required or the user explicitly requests them.
3
+ description: Optional delivery-first engineering workflow for repository changes. Use when its bounded validation, adaptive Git isolation, review, or coordination will improve delivery speed or defect prevention. Keep clean serial work on the current branch and let the model select beneficial design or subagents without turning them into mandatory ceremony.
4
4
  ---
5
5
 
6
6
  # Auto Coding Skill
7
7
 
8
8
  ## Use the two authorities
9
9
 
10
- Read root `AGENTS.md` for the shared behavioral protocol. Read
11
- `docs/ENGINEERING.md` frontmatter for project facts, access values, risk rules,
12
- validation routes, and time budgets. Do not reconstruct or duplicate those rules
13
- from historical task documents.
10
+ Read root `AGENTS.md` for the shared behavioral protocol and
11
+ `docs/ENGINEERING.md` for project facts, access, risk, validation, and budgets.
12
+ Do not reconstruct or duplicate those rules from historical task documents.
14
13
 
15
14
  Normal delivery is:
16
15
 
@@ -19,24 +18,32 @@ changed-scope gate → commit/push`.
19
18
 
20
19
  ## Select the minimum mechanism set
21
20
 
22
- Skip classification for read-only work, obvious clean serial edits, and terminal
23
- ledger maintenance. When impact, risk, or concurrency is unclear, run:
21
+ Read-only questions need no workflow command. Obvious clean serial edits and
22
+ terminal ledger maintenance may follow root `AGENTS.md` directly. When task kind,
23
+ impact, risk, or concurrency is unclear, run:
24
24
 
25
25
  ```bash
26
26
  python3 docs/tools/autopipeline/ap.py classify --scope auto \
27
- --planned-path <PATH> --intent "<intent>" [--writers <N>]
27
+ --planned-path <PATH> --intent "<intent>" [--writers <N>] \
28
+ [--task-kind read_only|change|terminal_maintenance]
28
29
  ```
29
30
 
30
- Follow `mechanism_plan.required`. Do not add anything listed under
31
- `mechanism_plan.not_required` unless the user explicitly requests it. In
32
- particular, ordinary direct work does not create a task lifecycle.
31
+ Run `mechanism_plan.required`. The model may select
32
+ `optional_when_beneficial` only when expected value exceeds coordination cost.
33
+ Do not run `forbidden` mechanisms unless the user explicitly overrides the plan.
34
+ Reclassify only when task kind, scope, risk, or writer count materially changes.
35
+ Ordinary direct work does not create a task lifecycle.
33
36
 
34
37
  Use the registered lifecycle only when classification requires isolation or
35
38
  fingerprinted review, or when the user explicitly requests it:
36
39
 
37
40
  ```bash
38
41
  python3 docs/tools/autopipeline/ap.py task-start T0001 \
39
- --owned-path src [--isolated] [--review-required]
42
+ --owned-path src [--isolated] [--review-required] [--force-lifecycle]
43
+ # For review-required work, obtain the current fingerprint and approve it:
44
+ python3 docs/tools/autopipeline/ap.py task-status T0001 --json
45
+ python3 docs/tools/autopipeline/ap.py task-review T0001 \
46
+ --verdict approved --diff-fingerprint <CURRENT_DIFF_FINGERPRINT>
40
47
  python3 docs/tools/autopipeline/ap.py commit-push T0001 --msg "T0001: summary"
41
48
  # Isolated tasks only:
42
49
  python3 docs/tools/autopipeline/ap.py task-integrate T0001
@@ -54,21 +61,20 @@ python3 docs/tools/autopipeline/ap.py light-gate --scope changed --explain
54
61
  ```
55
62
 
56
63
  Every changed code/config path must map to a real project command. The runtime
57
- de-duplicates matched commands and enforces both per-command and total closure
58
- budgets. A timeout means the route must be narrowed; it never falls back to a full
59
- build. Focused test/fix/retest loops remain allowed before the final gate.
64
+ de-duplicates commands and enforces route, command, and total budgets. Defaults
65
+ are 120 seconds per command and 180 total; projects may raise them for a measured
66
+ affected-scope check. Timeouts should narrow the route, never trigger a full build.
67
+ Focused test/fix/retest loops remain allowed before the final gate.
60
68
 
61
69
  Run full regression, Docker, Jenkins, deployment, browser/device writes, or target
62
- acceptance only as an explicit diagnostic. If the user asks to diagnose the
63
- just-pushed failure, continue in the same conversation/task without inventing a
64
- new ledger lifecycle.
70
+ acceptance only as an explicit diagnostic. A requested just-pushed failure
71
+ diagnosis continues in the same task without inventing another lifecycle.
65
72
 
66
73
  ## Keep state and upgrades safe
67
74
 
68
- Machine task state, leases, fingerprints, and gate evidence stay in Git
69
- common/local state. Ordinary work creates no taskbook, closure Markdown, evidence
70
- JSONL, active-task, or design artifact. Terminal ledger reconciliation validates
71
- and commits once without creating another lifecycle.
75
+ Machine task state, leases, fingerprints, and gate evidence stay in Git local
76
+ state. Ordinary work creates no taskbook, closure, evidence JSONL, active-task,
77
+ or design artifact. Terminal ledger reconciliation validates and commits once.
72
78
 
73
79
  ```bash
74
80
  autocoding init
@@ -78,6 +84,5 @@ python3 docs/tools/autopipeline/ap.py upgrade --write
78
84
  python3 docs/tools/autopipeline/ap.py doctor
79
85
  ```
80
86
 
81
- Finish registered tasks before changing runtime versions. Sync replaces the
82
- canonical root `AGENTS.md`, archives its previous content, preserves project facts
83
- and access values, and converges all managed Skill/runtime/agent assets together.
87
+ Finish registered tasks before changing versions. Sync replaces root `AGENTS.md`,
88
+ archives it, preserves project facts/access, and converges all managed assets.
@@ -13,9 +13,11 @@
13
13
  },
14
14
  "mechanismPlan": {
15
15
  "type": "object",
16
- "required": ["required", "not_required", "lifecycle_required", "rule"],
16
+ "required": ["required", "optional_when_beneficial", "forbidden", "not_required", "lifecycle_required", "rule"],
17
17
  "properties": {
18
18
  "required": { "$ref": "#/$defs/stringArray" },
19
+ "optional_when_beneficial": { "$ref": "#/$defs/stringArray" },
20
+ "forbidden": { "$ref": "#/$defs/stringArray" },
19
21
  "not_required": { "$ref": "#/$defs/stringArray" },
20
22
  "lifecycle_required": { "type": "boolean" },
21
23
  "rule": { "type": "string", "minLength": 1 }
@@ -452,6 +454,7 @@
452
454
  "classification_files",
453
455
  "intent_provided",
454
456
  "intent_categories",
457
+ "task_kind",
455
458
  "profile",
456
459
  "categories",
457
460
  "recommended_agents",
@@ -480,6 +483,9 @@
480
483
  "intent_categories": {
481
484
  "$ref": "#/$defs/stringArray"
482
485
  },
486
+ "task_kind": {
487
+ "enum": ["none", "read_only", "change", "terminal_maintenance"]
488
+ },
483
489
  "profile": {
484
490
  "enum": [
485
491
  "micro",
@@ -1,8 +1,8 @@
1
1
  {
2
2
  "schema_version": 1,
3
3
  "managed_versions": {
4
- "agents": "4.1.1",
5
- "engineering": "4.1.1"
4
+ "agents": "4.1.2",
5
+ "engineering": "4.1.2"
6
6
  },
7
7
  "known_official_agents_sha256": [
8
8
  "ebaaadab60327b3a4852948443d7191a4d48cb00bf4910038ab78c9149089373"
@@ -1,6 +1,6 @@
1
1
  ---
2
2
  workflow:
3
- skill_version: "4.1.1"
3
+ skill_version: "4.1.2"
4
4
  mode: "dev"
5
5
  profile: "auto"
6
6
  completion: "push"
@@ -56,8 +56,8 @@ commands:
56
56
 
57
57
  validation:
58
58
  on_unmapped: "error"
59
- # Normal closure cannot exceed these hard ceilings. Values may be lowered but
60
- # not raised above 120 seconds per command or 180 seconds total.
59
+ # Recommended fast defaults. Raise only for a measured affected-scope check;
60
+ # an individual route may set a smaller timeout_seconds.
61
61
  max_command_seconds: 120
62
62
  max_total_seconds: 180
63
63
  # Initialization must replace this with real path-to-command mappings. Empty
@@ -94,7 +94,7 @@ docs:
94
94
 
95
95
  # Engineering Configuration and Project Facts
96
96
 
97
- <!-- auto-coding-skill:managed-workflow:start version=4.1.1 -->
97
+ <!-- auto-coding-skill:managed-workflow:start version=4.1.2 -->
98
98
 
99
99
  Root `AGENTS.md` is the single shared behavioral protocol. The installed
100
100
  `SKILL.md` contains invocation guidance. This file is the single source for the
@@ -107,8 +107,8 @@ The frontmatter contract is:
107
107
  - `concurrency`: target branch and adaptive isolation settings.
108
108
  - `commands`: project-native executable commands.
109
109
  - `validation`: explicit path-to-command routes plus the bounded final-gate time
110
- budget. Code/config paths must be mapped; slower checks belong to explicit
111
- diagnostics.
110
+ budget. Code/config paths must be mapped; route-level `timeout_seconds` can
111
+ bound a slower command without changing the project defaults.
112
112
  - `risk`: project-specific signals that increase design/review depth but never
113
113
  expand the automatic local gate.
114
114
  - `structure` and `optimization`: advisory architecture and no-new-debt policy.
@@ -1,4 +1,4 @@
1
- <!-- auto-coding-skill:managed-agents:start version=4.1.1 -->
1
+ <!-- auto-coding-skill:managed-agents:start version=4.1.2 -->
2
2
  # Shared Engineering Protocol
3
3
 
4
4
  This file is fully managed by `auto-coding-skill`. Keep project-specific facts,
@@ -26,7 +26,7 @@ The default budget is:
26
26
 
27
27
  | Situation | Required mechanisms |
28
28
  | --- | --- |
29
- | Read-only question | none |
29
+ | Read-only question | analysis only; no workflow command |
30
30
  | Obvious clean serial edit | analysis, affected gate, commit/push |
31
31
  | Terminal ledger/archive maintenance | targeted consistency check, one commit |
32
32
  | Dirty checkout or concurrent writer | registered lifecycle and isolated worktree |
@@ -34,9 +34,11 @@ The default budget is:
34
34
  | Multiple independent writers | one task/worktree/lease per writer, ordered integration |
35
35
 
36
36
  Do not add classify, task lifecycle, durable design, subagents, reviewer, or a
37
- worktree outside this budget merely because the tool exists. If impact is unclear,
38
- run `classify` and follow `mechanism_plan.required`; items in
39
- `mechanism_plan.not_required` stay off unless the user explicitly requests them.
37
+ worktree merely because the tool exists. If task kind or impact is unclear, run
38
+ `classify`: run `mechanism_plan.required`, let the model select
39
+ `optional_when_beneficial` only when expected value exceeds coordination cost,
40
+ and keep `forbidden` mechanisms off unless the user explicitly overrides the
41
+ plan. Reclassify only after a material task-kind, scope, risk, or writer change.
40
42
 
41
43
  Normal delivery is:
42
44
 
@@ -64,8 +66,9 @@ failure may continue in the same conversation/task without a second lifecycle.
64
66
 
65
67
  ## Design, agents, and review
66
68
 
67
- - Micro and standard tasks stay main-agent-only unless independent delegation has
68
- a concrete latency or expertise benefit.
69
+ - Micro and standard tasks stay main-agent-only unless the model identifies an
70
+ independent question whose latency or expertise benefit exceeds coordination
71
+ cost.
69
72
  - Create DD/ADR only for lasting cross-module, API, data, security, deployment,
70
73
  or key user-flow decisions.
71
74
  - Use read-only explorer/docs/browser agents only for independent questions.
@@ -90,9 +93,10 @@ failure may continue in the same conversation/task without a second lifecycle.
90
93
  validation.
91
94
  - Focused tests may run and rerun during development. Run one final routed gate
92
95
  after the diff is stable.
93
- - The final gate has hard project-configurable ceilings of 120 seconds per command
94
- and 180 seconds total. A timeout must narrow the route; never expand or retry it
95
- as a full gate.
96
+ - The final gate defaults to 120 seconds per command and 180 seconds total. A
97
+ project may raise either budget for a measured affected-scope check, and may set
98
+ a smaller `timeout_seconds` on an individual route. A timeout should narrow the
99
+ route; never expand or retry it as a full gate.
96
100
  - Do not proactively install dependencies. Only after the selected route fails
97
101
  because a repository-locked dependency is absent may it be restored once and
98
102
  only that route retried.
@@ -46,12 +46,12 @@ _GENERATED_NOISE_PATTERNS = [
46
46
  ]
47
47
  _AGENT_CONTRACT_VERSION = 1
48
48
  _AGENT_CONTRACT_SCHEMA = "data/contracts/orchestration-v1.schema.json"
49
- _MAX_FINAL_COMMAND_SECONDS = 120.0
50
- _MAX_FINAL_TOTAL_SECONDS = 180.0
49
+ _RECOMMENDED_FINAL_COMMAND_SECONDS = 120.0
50
+ _RECOMMENDED_FINAL_TOTAL_SECONDS = 180.0
51
51
  _WORKFLOW_MIGRATION_POLICY = Path("data/policies/workflow-migrations-v1.json")
52
52
  _FALLBACK_WORKFLOW_MIGRATION_POLICY = {
53
53
  "schema_version": 1,
54
- "managed_versions": {"agents": "4.1.1", "engineering": "4.1.1"},
54
+ "managed_versions": {"agents": "4.1.2", "engineering": "4.1.2"},
55
55
  "known_official_engineering_body_sha256": [
56
56
  "d1306cc626e8baf8c83c953b760fd771066de2bf125168eca3a7b7d6ff2b87a2",
57
57
  "305931c6edef770033a4f1970b00e5fb1c1728351856a173dbfa497daf563021",
@@ -638,8 +638,8 @@ def _migrate_fast_development_defaults(cfg: dict, repo: Path) -> list[str]:
638
638
  if _text(workflow_cfg.get("completion")).lower() != "push":
639
639
  workflow_cfg["completion"] = "push"
640
640
  changed.append("workflow.completion")
641
- if _text(workflow_cfg.get("skill_version")) != "4.1.1":
642
- workflow_cfg["skill_version"] = "4.1.1"
641
+ if _text(workflow_cfg.get("skill_version")) != "4.1.2":
642
+ workflow_cfg["skill_version"] = "4.1.2"
643
643
  changed.append("workflow.skill_version")
644
644
 
645
645
  commands_cfg = cfg.setdefault("commands", {})
@@ -689,8 +689,8 @@ def _migrate_fast_development_defaults(cfg: dict, repo: Path) -> list[str]:
689
689
  validation_cfg["on_unmapped"] = "error"
690
690
  changed.append("validation.on_unmapped")
691
691
  for key, default_value in [
692
- ("max_command_seconds", int(_MAX_FINAL_COMMAND_SECONDS)),
693
- ("max_total_seconds", int(_MAX_FINAL_TOTAL_SECONDS)),
692
+ ("max_command_seconds", int(_RECOMMENDED_FINAL_COMMAND_SECONDS)),
693
+ ("max_total_seconds", int(_RECOMMENDED_FINAL_TOTAL_SECONDS)),
694
694
  ]:
695
695
  if key not in validation_cfg:
696
696
  validation_cfg[key] = default_value
@@ -952,16 +952,16 @@ def cmd_upgrade(args: argparse.Namespace) -> None:
952
952
  if current_agents:
953
953
  archive_dir = repo / "docs" / "archive" / "workflow"
954
954
  archive_header = (
955
- "# Archived AGENTS.md before auto-coding-skill 4.1.1\n\n"
955
+ "# Archived AGENTS.md before auto-coding-skill 4.1.2\n\n"
956
956
  "This file is historical and non-authoritative. The root AGENTS.md is fully managed.\n"
957
957
  "Move any still-current project facts into docs/ENGINEERING.md or docs/project/,\n"
958
958
  "without copying workflow rules back into the root AGENTS.md.\n\n---\n\n"
959
959
  )
960
960
  archive_content = archive_header + current_agents
961
- archive = archive_dir / "AGENTS.pre-4.1.1.md"
961
+ archive = archive_dir / "AGENTS.pre-4.1.2.md"
962
962
  if archive.exists() and archive.read_text(encoding="utf-8") != archive_content:
963
963
  digest = hashlib.sha256(current_agents.encode("utf-8")).hexdigest()[:12]
964
- archive = archive_dir / f"AGENTS.pre-4.1.1-{digest}.md"
964
+ archive = archive_dir / f"AGENTS.pre-4.1.2-{digest}.md"
965
965
  if not archive.exists():
966
966
  add_action("policy", archive, "archive", "historical and non-authoritative")
967
967
  if write:
@@ -1732,6 +1732,17 @@ def _task_registry_path(repo: Path, task_id: str) -> Path:
1732
1732
  return _task_state_root(repo) / "tasks" / f"{_validate_task_id(task_id)}.json"
1733
1733
 
1734
1734
 
1735
+ def _has_registered_active_task(repo: Path) -> bool:
1736
+ registry = _task_state_root(repo) / "tasks"
1737
+ if not registry.exists():
1738
+ return False
1739
+ for path in sorted(registry.glob("*.json")):
1740
+ payload = _read_json_object(path) or {}
1741
+ if _text(payload.get("state")) not in {"", "integrated", "cleanup-pending"}:
1742
+ return True
1743
+ return False
1744
+
1745
+
1735
1746
  def _worktree_manifest_path(repo: Path) -> Path:
1736
1747
  return _git_dir(repo) / "auto-coding-skill-task.json"
1737
1748
 
@@ -2466,6 +2477,7 @@ def _resolve_secret(section_name: str, section_cfg: dict, field: str) -> str:
2466
2477
 
2467
2478
  _GATE_SCOPES = {"auto", "changed", "standard", "full"}
2468
2479
  _WORKFLOW_PROFILES = {"auto", "micro", "standard", "high-risk"}
2480
+ _TASK_KINDS = {"auto", "none", "read_only", "change", "terminal_maintenance"}
2469
2481
  _PROFILE_GATE_SCOPE = {"micro": "changed", "standard": "changed", "high-risk": "changed"}
2470
2482
  _PROFILE_RANK = {"micro": 0, "standard": 1, "high-risk": 2}
2471
2483
  _SCOPE_RANK = {"changed": 0, "standard": 1, "full": 2}
@@ -2529,24 +2541,14 @@ def _final_gate_budget(cfg: dict) -> dict[str, float]:
2529
2541
  validation_cfg = _validation_cfg(cfg)
2530
2542
  command_seconds = _positive_seconds(
2531
2543
  validation_cfg.get("max_command_seconds"),
2532
- _MAX_FINAL_COMMAND_SECONDS,
2544
+ _RECOMMENDED_FINAL_COMMAND_SECONDS,
2533
2545
  "validation.max_command_seconds",
2534
2546
  )
2535
2547
  total_seconds = _positive_seconds(
2536
2548
  validation_cfg.get("max_total_seconds"),
2537
- _MAX_FINAL_TOTAL_SECONDS,
2549
+ _RECOMMENDED_FINAL_TOTAL_SECONDS,
2538
2550
  "validation.max_total_seconds",
2539
2551
  )
2540
- if command_seconds > _MAX_FINAL_COMMAND_SECONDS:
2541
- raise APError(
2542
- f"validation.max_command_seconds cannot exceed {_MAX_FINAL_COMMAND_SECONDS:.0f}; "
2543
- "move slower checks to an explicit diagnostic command"
2544
- )
2545
- if total_seconds > _MAX_FINAL_TOTAL_SECONDS:
2546
- raise APError(
2547
- f"validation.max_total_seconds cannot exceed {_MAX_FINAL_TOTAL_SECONDS:.0f}; "
2548
- "move slower checks to an explicit diagnostic command"
2549
- )
2550
2552
  if command_seconds > total_seconds:
2551
2553
  raise APError("validation.max_command_seconds cannot exceed validation.max_total_seconds")
2552
2554
  return {"command_seconds": command_seconds, "total_seconds": total_seconds}
@@ -2588,6 +2590,7 @@ def _run_configured_command_list(
2588
2590
  *,
2589
2591
  deadline: Optional[float] = None,
2590
2592
  command_timeout_s: Optional[float] = None,
2593
+ command_timeouts: Optional[dict[str, float]] = None,
2591
2594
  ) -> list[str]:
2592
2595
  executed: list[str] = []
2593
2596
  missing: list[str] = []
@@ -2597,6 +2600,9 @@ def _run_configured_command_list(
2597
2600
  continue
2598
2601
  if _configured_command(cfg, name):
2599
2602
  timeout_s = command_timeout_s
2603
+ route_timeout = (command_timeouts or {}).get(name)
2604
+ if route_timeout is not None:
2605
+ timeout_s = min(timeout_s, route_timeout) if timeout_s is not None else route_timeout
2600
2606
  if deadline is not None:
2601
2607
  remaining = deadline - time.monotonic()
2602
2608
  if remaining <= 0:
@@ -2784,6 +2790,7 @@ def _validation_plan(cfg: dict, paths: list[str]) -> dict:
2784
2790
  normalized = _unique_paths(paths)
2785
2791
  routes = _validation_routes(cfg)
2786
2792
  command_names: list[str] = []
2793
+ command_timeouts: dict[str, float] = {}
2787
2794
  matched_routes: list[str] = []
2788
2795
  coverage: dict[str, list[str]] = {}
2789
2796
  unmapped: list[str] = []
@@ -2794,6 +2801,13 @@ def _validation_plan(cfg: dict, paths: list[str]) -> dict:
2794
2801
  if not _validation_route_matches(path, route):
2795
2802
  continue
2796
2803
  route_name = _text(route.get("name")) or f"route-{index + 1}"
2804
+ route_timeout: Optional[float] = None
2805
+ if route.get("timeout_seconds") not in {None, ""}:
2806
+ route_timeout = _positive_seconds(
2807
+ route.get("timeout_seconds"),
2808
+ 1.0,
2809
+ f"validation.routes[{index}].timeout_seconds",
2810
+ )
2797
2811
  names.append(route_name)
2798
2812
  if route_name not in matched_routes:
2799
2813
  matched_routes.append(route_name)
@@ -2801,6 +2815,13 @@ def _validation_plan(cfg: dict, paths: list[str]) -> dict:
2801
2815
  command_name = _command_name(command_ref)
2802
2816
  if command_name and command_name not in command_names:
2803
2817
  command_names.append(command_name)
2818
+ if command_name and route_timeout is not None:
2819
+ current_timeout = command_timeouts.get(command_name)
2820
+ command_timeouts[command_name] = (
2821
+ min(current_timeout, route_timeout)
2822
+ if current_timeout is not None
2823
+ else route_timeout
2824
+ )
2804
2825
  coverage[path] = names
2805
2826
  if not names and not _path_matches(path, _DOC_PATH_PATTERNS):
2806
2827
  unmapped.append(path)
@@ -2829,6 +2850,7 @@ def _validation_plan(cfg: dict, paths: list[str]) -> dict:
2829
2850
  return {
2830
2851
  "paths": normalized,
2831
2852
  "commands": command_names,
2853
+ "command_timeouts": command_timeouts,
2832
2854
  "matched_routes": matched_routes,
2833
2855
  "coverage": coverage,
2834
2856
  "unmapped": unmapped,
@@ -2965,6 +2987,7 @@ def _run_changed_gate(
2965
2987
  plan["commands"],
2966
2988
  deadline=deadline,
2967
2989
  command_timeout_s=command_timeout_s,
2990
+ command_timeouts=plan["command_timeouts"],
2968
2991
  )
2969
2992
  print(
2970
2993
  "[validation] routes="
@@ -2990,6 +3013,7 @@ def cmd_validation_map_check(args: argparse.Namespace) -> None:
2990
3013
  result = {
2991
3014
  "paths": plan["paths"],
2992
3015
  "commands": plan["commands"],
3016
+ "command_timeouts": plan["command_timeouts"],
2993
3017
  "matched_routes": plan["matched_routes"],
2994
3018
  "coverage": plan["coverage"],
2995
3019
  "docs_only": plan["docs_only"],
@@ -4464,6 +4488,47 @@ def _classify_intent(intent: str) -> list[str]:
4464
4488
  )
4465
4489
 
4466
4490
 
4491
+ def _resolve_task_kind(
4492
+ requested: str,
4493
+ paths: list[str],
4494
+ intent: str,
4495
+ ) -> str:
4496
+ value = _text(requested).lower().replace("-", "_") or "auto"
4497
+ if value not in _TASK_KINDS:
4498
+ raise APError("task kind must be one of: " + ", ".join(sorted(_TASK_KINDS)))
4499
+ if value != "auto":
4500
+ return value
4501
+ if _is_terminal_ledger_maintenance(paths):
4502
+ return "terminal_maintenance"
4503
+ if paths:
4504
+ return "change"
4505
+ normalized_intent = _text(intent).lower()
4506
+ if not normalized_intent:
4507
+ return "none"
4508
+ terminal_action = bool(
4509
+ re.search(r"\b(archive|close|reconcile)\b", normalized_intent)
4510
+ or any(keyword in normalized_intent for keyword in ["归档", "对账", "收口"])
4511
+ )
4512
+ terminal_subject = bool(
4513
+ re.search(r"\b(closure|ledger|task\s+records?|taskbook)\b", normalized_intent)
4514
+ or any(keyword in normalized_intent for keyword in ["台账", "任务记录", "关闭记录"])
4515
+ )
4516
+ if terminal_action and terminal_subject:
4517
+ return "terminal_maintenance"
4518
+ english_change = re.search(
4519
+ r"\b(add|build|change|create|delete|deploy|develop|fix|implement|migrate|optimize|publish|refactor|release|remove|rename|resolve|update|upgrade)\b",
4520
+ normalized_intent,
4521
+ )
4522
+ chinese_change = any(
4523
+ keyword in normalized_intent
4524
+ for keyword in [
4525
+ "修改", "新增", "删除", "修复", "解决", "整改", "实现", "迁移", "发布",
4526
+ "重构", "改名", "升级", "优化", "完善", "补充", "清理", "落地", "提交", "推送",
4527
+ ]
4528
+ )
4529
+ return "change" if english_change or chinese_change else "read_only"
4530
+
4531
+
4467
4532
  def _normalize_workflow_profile(value: object, default: str = "auto") -> str:
4468
4533
  profile = _text(value).lower() or default
4469
4534
  if profile not in _WORKFLOW_PROFILES:
@@ -4645,46 +4710,66 @@ def _agent_execution_plan(
4645
4710
 
4646
4711
  def _mechanism_plan(
4647
4712
  *,
4648
- has_task_signal: bool,
4649
- terminal_maintenance: bool,
4713
+ task_kind: str,
4650
4714
  execution_mode: str,
4651
4715
  review_required: bool,
4652
4716
  design_required: bool,
4653
4717
  parallel_writers: int,
4654
4718
  ) -> dict:
4655
- code_delivery = has_task_signal and not terminal_maintenance
4656
- required = ["analysis"] if has_task_signal else []
4657
- if design_required:
4658
- required.append("durable_design")
4659
- if execution_mode == "isolated":
4660
- required.extend(["task_lifecycle", "worktree"])
4661
- elif review_required:
4662
- required.append("task_lifecycle")
4663
- if parallel_writers > 1:
4664
- required.append("parallel_fixers")
4665
- if review_required:
4666
- required.append("independent_review")
4667
- if code_delivery:
4719
+ universe = [
4720
+ "analysis",
4721
+ "targeted_consistency_check",
4722
+ "durable_design",
4723
+ "task_lifecycle",
4724
+ "worktree",
4725
+ "read_only_subagents",
4726
+ "parallel_fixers",
4727
+ "independent_review",
4728
+ "final_changed_scope_gate",
4729
+ "commit_push",
4730
+ ]
4731
+ required: list[str] = []
4732
+ optional: list[str] = []
4733
+ if task_kind == "read_only":
4734
+ required = ["analysis"]
4735
+ optional = ["read_only_subagents"]
4736
+ elif task_kind == "terminal_maintenance":
4737
+ required = ["analysis", "targeted_consistency_check", "commit_push"]
4738
+ elif task_kind == "change":
4739
+ required = ["analysis"]
4740
+ if design_required:
4741
+ required.append("durable_design")
4742
+ else:
4743
+ optional.append("durable_design")
4744
+ if execution_mode == "isolated":
4745
+ required.extend(["task_lifecycle", "worktree"])
4746
+ elif review_required:
4747
+ required.append("task_lifecycle")
4748
+ if parallel_writers > 1:
4749
+ required.append("parallel_fixers")
4750
+ else:
4751
+ optional.append("parallel_fixers")
4752
+ if review_required:
4753
+ required.append("independent_review")
4754
+ else:
4755
+ optional.append("independent_review")
4756
+ optional.append("read_only_subagents")
4668
4757
  required.extend(["final_changed_scope_gate", "commit_push"])
4669
4758
  ordered = list(dict.fromkeys(required))
4670
- optional = [
4671
- mechanism
4672
- for mechanism in [
4673
- "classify",
4674
- "durable_design",
4675
- "task_lifecycle",
4676
- "worktree",
4677
- "read_only_subagents",
4678
- "parallel_fixers",
4679
- "independent_review",
4680
- ]
4681
- if mechanism not in ordered
4682
- ]
4759
+ optional = [item for item in dict.fromkeys(optional) if item not in ordered]
4760
+ forbidden = [item for item in universe if item not in ordered and item not in optional]
4683
4761
  return {
4684
4762
  "required": ordered,
4685
- "not_required": optional,
4763
+ "optional_when_beneficial": optional,
4764
+ "forbidden": forbidden,
4765
+ # 4.1 compatibility: old clients treat only truly forbidden mechanisms
4766
+ # as not required; model-selectable mechanisms remain available.
4767
+ "not_required": forbidden,
4686
4768
  "lifecycle_required": "task_lifecycle" in ordered,
4687
- "rule": "do not add a mechanism that is not required unless the user explicitly requests it",
4769
+ "rule": (
4770
+ "run required mechanisms; select optional mechanisms only when expected value exceeds "
4771
+ "coordination cost; do not run forbidden mechanisms unless the user explicitly overrides"
4772
+ ),
4688
4773
  }
4689
4774
 
4690
4775
 
@@ -4700,6 +4785,7 @@ def _resolve_execution_plan(
4700
4785
  planned_paths: Optional[list[str]] = None,
4701
4786
  intent: str = "",
4702
4787
  parallel_writers: int = 1,
4788
+ requested_task_kind: str = "",
4703
4789
  ) -> dict:
4704
4790
  impact = _impact_summary(
4705
4791
  cfg,
@@ -4713,6 +4799,7 @@ def _resolve_execution_plan(
4713
4799
  {_normalize_owned_path(item) for item in (planned_paths or []) if _text(item)}
4714
4800
  )
4715
4801
  classification_inputs = [*paths, *normalized_planned_paths]
4802
+ task_kind = _resolve_task_kind(requested_task_kind, classification_inputs, intent)
4716
4803
  classification = _classify_paths(classification_inputs)
4717
4804
  intent_categories = _classify_intent(intent)
4718
4805
  merged_categories = set(classification["categories"]) | set(intent_categories)
@@ -4733,8 +4820,11 @@ def _resolve_execution_plan(
4733
4820
  (docs_only_paths and str(impact["selected_scope"]) != "full")
4734
4821
  or _tests_only(classification_inputs)
4735
4822
  )
4736
- terminal_maintenance = _is_terminal_ledger_maintenance(classification_inputs)
4737
- if docs_or_tests_only and str(impact["selected_scope"]) != "full":
4823
+ terminal_maintenance = task_kind == "terminal_maintenance"
4824
+ if task_kind in {"read_only", "terminal_maintenance", "none"}:
4825
+ detected_profile = "micro"
4826
+ profile_reasons.append(f"task kind: {task_kind}")
4827
+ elif docs_or_tests_only and str(impact["selected_scope"]) != "full":
4738
4828
  detected_profile = "micro"
4739
4829
  profile_reasons.append("only docs/test files changed")
4740
4830
 
@@ -4745,10 +4835,10 @@ def _resolve_execution_plan(
4745
4835
  "matched gate rule has invalid profile: " + ", ".join(sorted(invalid_rule_profiles))
4746
4836
  )
4747
4837
  rule_profiles = raw_rule_profiles
4748
- if "micro" in rule_profiles and detected_profile == "standard":
4838
+ if task_kind == "change" and "micro" in rule_profiles and detected_profile == "standard":
4749
4839
  detected_profile = "micro"
4750
4840
  profile_reasons.append("matched gate rule with profile=micro")
4751
- if "standard" in rule_profiles and detected_profile == "micro":
4841
+ if task_kind == "change" and "standard" in rule_profiles and detected_profile == "micro":
4752
4842
  detected_profile = "standard"
4753
4843
  profile_reasons.append("matched gate rule with profile=standard")
4754
4844
 
@@ -4762,9 +4852,9 @@ def _resolve_execution_plan(
4762
4852
  "release_or_tooling",
4763
4853
  }
4764
4854
  high_signals: list[str] = []
4765
- if categories & high_categories and not docs_or_tests_only:
4855
+ if task_kind == "change" and categories & high_categories and not docs_or_tests_only:
4766
4856
  high_signals.append("high-risk category: " + ", ".join(sorted(categories & high_categories)))
4767
- if "high-risk" in rule_profiles:
4857
+ if task_kind == "change" and "high-risk" in rule_profiles:
4768
4858
  high_signals.append("matched gate rule with profile=high-risk")
4769
4859
  if high_signals:
4770
4860
  detected_profile = "high-risk"
@@ -4817,19 +4907,19 @@ def _resolve_execution_plan(
4817
4907
  and categories & {"api", "db", "auth", "payment", "file_transfer", "gateway", "prod_config"}
4818
4908
  )
4819
4909
  review_required = bool(
4820
- not terminal_maintenance
4910
+ task_kind == "change"
4821
4911
  and (effective_profile == "high-risk" or writer_count > 1 or cross_module_contract_risk or rule_review)
4822
4912
  )
4823
4913
  design_required = bool(
4824
- not terminal_maintenance
4914
+ task_kind == "change"
4825
4915
  and (rule_design or (effective_profile == "high-risk" and cross_module))
4826
4916
  )
4827
- has_task_signal = bool(classification_inputs or _text(intent))
4828
4917
  workspace_dirty = bool(_working_tree_paths(repo))
4918
+ active_writer = _has_registered_active_task(repo)
4829
4919
  configured_isolation = _task_isolation(cfg)
4830
- if terminal_maintenance or not has_task_signal:
4920
+ if task_kind != "change":
4831
4921
  execution_mode = "none"
4832
- elif configured_isolation == "worktree" or writer_count > 1 or workspace_dirty:
4922
+ elif configured_isolation == "worktree" or writer_count > 1 or workspace_dirty or active_writer:
4833
4923
  execution_mode = "isolated"
4834
4924
  else:
4835
4925
  execution_mode = "direct"
@@ -4849,8 +4939,7 @@ def _resolve_execution_plan(
4849
4939
  review_required=review_required,
4850
4940
  )
4851
4941
  mechanism_plan = _mechanism_plan(
4852
- has_task_signal=has_task_signal,
4853
- terminal_maintenance=terminal_maintenance,
4942
+ task_kind=task_kind,
4854
4943
  execution_mode=execution_mode,
4855
4944
  review_required=review_required,
4856
4945
  design_required=design_required,
@@ -4863,6 +4952,7 @@ def _resolve_execution_plan(
4863
4952
  "planned_files": normalized_planned_paths,
4864
4953
  "classification_files": classification_inputs,
4865
4954
  "intent_provided": bool(_text(intent)),
4955
+ "task_kind": task_kind,
4866
4956
  "intent_categories": intent_categories,
4867
4957
  "reasons": execution_reasons,
4868
4958
  "configured_profile": configured_profile,
@@ -4875,6 +4965,7 @@ def _resolve_execution_plan(
4875
4965
  "execution_mode": execution_mode,
4876
4966
  "terminal_maintenance": terminal_maintenance,
4877
4967
  "workspace_dirty": workspace_dirty,
4968
+ "active_writer": active_writer,
4878
4969
  "parallel_writers": writer_count,
4879
4970
  "cross_module": cross_module,
4880
4971
  "review_required": review_required,
@@ -4923,6 +5014,7 @@ def cmd_classify(args: argparse.Namespace) -> None:
4923
5014
  planned_paths=list(getattr(args, "planned_path", []) or []),
4924
5015
  intent="\n".join(part for part in intent_parts if part),
4925
5016
  parallel_writers=int(getattr(args, "writers", 1) or 1),
5017
+ requested_task_kind=_text(getattr(args, "task_kind", "")),
4926
5018
  )
4927
5019
  risk = {"micro": "P3", "standard": "P2", "high-risk": "P1"}[plan["profile"]]
4928
5020
  commands: list[str] = []
@@ -4937,6 +5029,7 @@ def cmd_classify(args: argparse.Namespace) -> None:
4937
5029
  print(json.dumps(result, ensure_ascii=False, indent=2))
4938
5030
  else:
4939
5031
  print(f"[classify] risk={risk}")
5032
+ print(f"[classify] task_kind={result['task_kind']}")
4940
5033
  print(f"[classify] profile={result['profile']}")
4941
5034
  print(f"[classify] effective_mode={result['effective_mode']}")
4942
5035
  print(f"[classify] execution_mode={result['execution_mode']}")
@@ -4951,8 +5044,12 @@ def cmd_classify(args: argparse.Namespace) -> None:
4951
5044
  + (", ".join(result["mechanism_plan"]["required"]) or "(none)")
4952
5045
  )
4953
5046
  print(
4954
- "[classify] not_required="
4955
- + (", ".join(result["mechanism_plan"]["not_required"]) or "(none)")
5047
+ "[classify] optional_when_beneficial="
5048
+ + (", ".join(result["mechanism_plan"]["optional_when_beneficial"]) or "(none)")
5049
+ )
5050
+ print(
5051
+ "[classify] forbidden="
5052
+ + (", ".join(result["mechanism_plan"]["forbidden"]) or "(none)")
4956
5053
  )
4957
5054
  for stage in result["agent_plan"]["stages"]:
4958
5055
  print(
@@ -5217,6 +5314,7 @@ def cmd_doctor(args: argparse.Namespace) -> None:
5217
5314
 
5218
5315
  missing: List[str] = []
5219
5316
  validation_errors: List[str] = []
5317
+ advisories: List[str] = []
5220
5318
  validation_errors.extend(_workflow_policy_issues(repo, cfg))
5221
5319
 
5222
5320
  mode = str(workflow_cfg.get("mode") or "dev").strip().lower()
@@ -5303,9 +5401,23 @@ def cmd_doctor(args: argparse.Namespace) -> None:
5303
5401
  if on_unmapped not in {"error", "fallback"}:
5304
5402
  validation_errors.append("validation.on_unmapped must be error or fallback")
5305
5403
  try:
5306
- _final_gate_budget(cfg)
5404
+ final_budget = _final_gate_budget(cfg)
5405
+ if final_budget["command_seconds"] > _RECOMMENDED_FINAL_COMMAND_SECONDS:
5406
+ advisories.append(
5407
+ "validation.max_command_seconds exceeds the recommended 120s default; "
5408
+ "keep the affected-scope route targeted"
5409
+ )
5410
+ if final_budget["total_seconds"] > _RECOMMENDED_FINAL_TOTAL_SECONDS:
5411
+ advisories.append(
5412
+ "validation.max_total_seconds exceeds the recommended 180s default; "
5413
+ "keep the final gate bounded to affected scope"
5414
+ )
5307
5415
  except APError as exc:
5308
5416
  validation_errors.append(str(exc))
5417
+ final_budget = {
5418
+ "command_seconds": _RECOMMENDED_FINAL_COMMAND_SECONDS,
5419
+ "total_seconds": _RECOMMENDED_FINAL_TOTAL_SECONDS,
5420
+ }
5309
5421
  routes = validation_cfg.get("routes") or []
5310
5422
  if not isinstance(routes, list):
5311
5423
  validation_errors.append("validation.routes must be a list")
@@ -5324,6 +5436,20 @@ def cmd_doctor(args: argparse.Namespace) -> None:
5324
5436
  validation_errors.append(
5325
5437
  f"validation.routes[{index}] references missing commands.{command_name}"
5326
5438
  )
5439
+ if "timeout_seconds" in route:
5440
+ try:
5441
+ route_timeout = _positive_seconds(
5442
+ route.get("timeout_seconds"),
5443
+ final_budget["command_seconds"],
5444
+ f"validation.routes[{index}].timeout_seconds",
5445
+ )
5446
+ if route_timeout > final_budget["command_seconds"]:
5447
+ advisories.append(
5448
+ f"validation.routes[{index}].timeout_seconds exceeds max_command_seconds "
5449
+ "and will be capped"
5450
+ )
5451
+ except APError as exc:
5452
+ validation_errors.append(str(exc))
5327
5453
  fallback_name = _command_name(validation_cfg.get("fallback_command"))
5328
5454
  if on_unmapped == "fallback" and not fallback_name:
5329
5455
  validation_errors.append(
@@ -5390,6 +5516,8 @@ def cmd_doctor(args: argparse.Namespace) -> None:
5390
5516
  "pass",
5391
5517
  {"mode": mode, "completion": completion, "access_fields": len(_REQUIRED_ACCESS_FIELDS)},
5392
5518
  )
5519
+ for advisory in advisories:
5520
+ print(f"[doctor] WARN: {advisory}")
5393
5521
  print("[doctor] OK")
5394
5522
 
5395
5523
 
@@ -6600,6 +6728,41 @@ def cmd_task_start(args: argparse.Namespace) -> None:
6600
6728
  repo = Path(args.repo).resolve()
6601
6729
  ensure_git_repo(repo)
6602
6730
  cfg = _load_cfg(repo)
6731
+ _require_control_checkout(repo)
6732
+ task_id = _validate_task_id(args.task_id)
6733
+ owned_paths = sorted({_normalize_owned_path(item) for item in (args.owned_path or [])})
6734
+ if not owned_paths:
6735
+ raise APError("task-start requires at least one --owned-path.")
6736
+ if _is_terminal_ledger_maintenance(owned_paths):
6737
+ raise APError(
6738
+ "Pure taskbook/closure/archive reconciliation is terminal maintenance and must not "
6739
+ "create another task lifecycle. Run the targeted document check, commit once, and push."
6740
+ )
6741
+ writer_count = max(1, int(getattr(args, "writers", 1) or 1))
6742
+ if writer_count != 1:
6743
+ raise APError(
6744
+ "task-start creates exactly one writer lease and one worktree. Start each parallel "
6745
+ "writer as its own task ID; use --writers only with classify for planning."
6746
+ )
6747
+ explicit_lifecycle = bool(
6748
+ getattr(args, "isolated", False)
6749
+ or getattr(args, "review_required", False)
6750
+ or getattr(args, "force_lifecycle", False)
6751
+ )
6752
+ preflight_plan = _resolve_execution_plan(
6753
+ cfg,
6754
+ repo,
6755
+ planned_paths=owned_paths,
6756
+ parallel_writers=writer_count,
6757
+ requested_task_kind="change",
6758
+ )
6759
+ if not preflight_plan["mechanism_plan"]["lifecycle_required"] and not explicit_lifecycle:
6760
+ raise APError(
6761
+ "This clean serial task does not need a machine task lifecycle. Work directly on "
6762
+ "the current branch, run the routed final gate, commit, and push. Use task-start "
6763
+ "only when classify requires isolation/review or the user explicitly requests it."
6764
+ )
6765
+
6603
6766
  _require_workflow_policy_clean(repo, cfg)
6604
6767
  configured_isolation = _task_isolation(cfg)
6605
6768
  access_issues = _access_config_issues(cfg)
@@ -6608,8 +6771,6 @@ def cmd_task_start(args: argparse.Namespace) -> None:
6608
6771
  "Project initialization is incomplete; fill docs/ENGINEERING.md before starting work:\n- "
6609
6772
  + "\n- ".join(access_issues)
6610
6773
  )
6611
- task_id = _validate_task_id(args.task_id)
6612
- _require_control_checkout(repo)
6613
6774
  remote, target_branch, base_ref = _task_remote_and_target(cfg, args)
6614
6775
  concurrency_cfg = _concurrency_cfg(cfg)
6615
6776
  timeout_s = float(concurrency_cfg.get("lock_timeout_sec") or 30)
@@ -6626,20 +6787,6 @@ def cmd_task_start(args: argparse.Namespace) -> None:
6626
6787
  if not getattr(args, "no_fetch", False) and base_ref.startswith(f"{remote}/"):
6627
6788
  _fetch_target(repo, remote, target_branch)
6628
6789
  base_sha = _resolve_commit(repo, base_ref)
6629
- owned_paths = sorted({_normalize_owned_path(item) for item in (args.owned_path or [])})
6630
- if not owned_paths:
6631
- raise APError("task-start requires at least one --owned-path.")
6632
- if _is_terminal_ledger_maintenance(owned_paths):
6633
- raise APError(
6634
- "Pure taskbook/closure/archive reconciliation is terminal maintenance and must not "
6635
- "create another task lifecycle. Run the targeted document check, commit once, and push."
6636
- )
6637
- writer_count = max(1, int(getattr(args, "writers", 1) or 1))
6638
- if writer_count != 1:
6639
- raise APError(
6640
- "task-start creates exactly one writer lease and one worktree. Start each parallel "
6641
- "writer as its own task ID; use --writers only with classify for planning."
6642
- )
6643
6790
  registry_dir = _task_state_root(repo) / "tasks"
6644
6791
  for manifest_path in sorted(registry_dir.glob("*.json")) if registry_dir.exists() else []:
6645
6792
  active_manifest = _read_json_object(manifest_path) or {}
@@ -6679,11 +6826,7 @@ def cmd_task_start(args: argparse.Namespace) -> None:
6679
6826
  repo,
6680
6827
  planned_paths=owned_paths,
6681
6828
  parallel_writers=writer_count,
6682
- )
6683
- explicit_lifecycle = bool(
6684
- getattr(args, "isolated", False)
6685
- or getattr(args, "review_required", False)
6686
- or getattr(args, "force_lifecycle", False)
6829
+ requested_task_kind="change",
6687
6830
  )
6688
6831
  if not plan["mechanism_plan"]["lifecycle_required"] and not explicit_lifecycle:
6689
6832
  raise APError(
@@ -7925,6 +8068,7 @@ def main(argv: Optional[List[str]] = None) -> int:
7925
8068
  s.add_argument("--planned-path", action="append")
7926
8069
  s.add_argument("--intent")
7927
8070
  s.add_argument("--intent-file")
8071
+ s.add_argument("--task-kind", choices=sorted(_TASK_KINDS), default="auto")
7928
8072
  s.add_argument("--writers", type=int, default=1)
7929
8073
  s.add_argument("--json", action="store_true")
7930
8074
  s.set_defaults(func=cmd_classify)
@@ -0,0 +1,38 @@
1
+ #!/usr/bin/env node
2
+
3
+ import { spawnSync } from "node:child_process";
4
+ import path from "node:path";
5
+ import { fileURLToPath } from "node:url";
6
+
7
+ const here = path.dirname(fileURLToPath(import.meta.url));
8
+ const repoRoot = path.resolve(here, "..", "..");
9
+ const npmCommand = process.platform === "win32" ? "npm.cmd" : "npm";
10
+ const packed = spawnSync(
11
+ npmCommand,
12
+ ["pack", "--dry-run", "--json", "--ignore-scripts"],
13
+ { cwd: repoRoot, encoding: "utf8" },
14
+ );
15
+
16
+ if (packed.status !== 0) {
17
+ process.stderr.write(packed.stderr || packed.stdout || "npm pack --dry-run failed\n");
18
+ process.exit(packed.status || 1);
19
+ }
20
+
21
+ let payload;
22
+ try {
23
+ payload = JSON.parse(packed.stdout);
24
+ } catch (error) {
25
+ process.stderr.write(`Cannot parse npm pack output: ${error.message}\n`);
26
+ process.exit(1);
27
+ }
28
+
29
+ const files = payload?.[0]?.files?.map(item => String(item.path || "")) || [];
30
+ const forbidden = files.filter(file =>
31
+ /(^|\/)__pycache__(\/|$)/.test(file) || /\.py[cod]$/i.test(file),
32
+ );
33
+ if (forbidden.length > 0) {
34
+ process.stderr.write(`Generated Python cache leaked into package:\n- ${forbidden.join("\n- ")}\n`);
35
+ process.exit(1);
36
+ }
37
+
38
+ process.stdout.write(`[package-check] OK files=${files.length}\n`);
package/cli/src/index.js CHANGED
@@ -1125,15 +1125,15 @@ function projectStatus(project, assetSkill, assetAgents){
1125
1125
  if (isolationState.present && !["adaptive", "worktree"].includes(frontmatterScalarValue(isolationState.value).toLowerCase())) {
1126
1126
  invalidConfigTokens.push("concurrency.isolation (must be adaptive or worktree)");
1127
1127
  }
1128
- for (const [pathParts, maximum] of [
1129
- [["validation", "max_command_seconds"], 120],
1130
- [["validation", "max_total_seconds"], 180],
1128
+ for (const pathParts of [
1129
+ ["validation", "max_command_seconds"],
1130
+ ["validation", "max_total_seconds"],
1131
1131
  ]) {
1132
1132
  const state = frontmatterPathState(text, pathParts);
1133
1133
  if (!state.present) continue;
1134
1134
  const value = Number(frontmatterScalarValue(state.value));
1135
- if (!Number.isFinite(value) || value <= 0 || value > maximum) {
1136
- invalidConfigTokens.push(`${pathParts.join(".")} (must be > 0 and <= ${maximum})`);
1135
+ if (!Number.isFinite(value) || value <= 0) {
1136
+ invalidConfigTokens.push(`${pathParts.join(".")} (must be > 0)`);
1137
1137
  }
1138
1138
  }
1139
1139
  const commandState = frontmatterPathState(text, ["validation", "max_command_seconds"]);
@@ -420,6 +420,30 @@ class AutoCodingConcurrencyTests(unittest.TestCase):
420
420
  self.assertEqual(before, git_output(repo, "rev-parse", "HEAD"))
421
421
  self.assertFalse(self.registry_manifest_path(repo, "DIRECT-NOOP").exists())
422
422
 
423
+ def test_unnecessary_lifecycle_rejects_before_access_or_fetch_checks(self) -> None:
424
+ _, repo, _ = self.make_repo("adaptive", require_review=False)
425
+ engineering = repo / "docs" / "ENGINEERING.md"
426
+ text = engineering.read_text(encoding="utf-8")
427
+ engineering.write_text(text.replace("front-pass", "", 1), encoding="utf-8")
428
+ git(repo, "add", "docs/ENGINEERING.md")
429
+ git(repo, "commit", "-qm", "invalidate access for preflight test")
430
+ git(repo, "push", "-q", "origin", "dev")
431
+
432
+ result = self.ap(
433
+ repo,
434
+ "task-start",
435
+ "DIRECT-PREFLIGHT",
436
+ "--base",
437
+ "origin/branch-that-must-not-be-fetched",
438
+ "--owned-path",
439
+ "shared.txt",
440
+ check=False,
441
+ )
442
+ output = result.stdout + result.stderr
443
+ self.assertNotEqual(0, result.returncode)
444
+ self.assertIn("does not need a machine task lifecycle", output)
445
+ self.assertNotIn("initialization is incomplete", output)
446
+
423
447
  def test_schema_one_task_status_fails_closed_with_migration_recovery(self) -> None:
424
448
  _, repo, _ = self.make_repo()
425
449
  self.start_task(repo, "LEGACY-SCHEMA")
@@ -243,6 +243,7 @@ class AutoCodingProfileTests(unittest.TestCase):
243
243
  repo, cfg = self.make_repo()
244
244
  plan = self.plan(repo, cfg, planned_paths=["src/widget.py"])
245
245
  self.assertEqual("standard", plan["profile"])
246
+ self.assertEqual("change", plan["task_kind"])
246
247
  self.assertEqual("changed", plan["selected_scope"])
247
248
  self.assertEqual("dev", plan["effective_mode"])
248
249
  self.assertEqual([], plan["recommended_agents"])
@@ -269,7 +270,48 @@ class AutoCodingProfileTests(unittest.TestCase):
269
270
  plan["mechanism_plan"]["required"],
270
271
  )
271
272
  self.assertFalse(plan["mechanism_plan"]["lifecycle_required"])
272
- self.assertIn("task_lifecycle", plan["mechanism_plan"]["not_required"])
273
+ self.assertIn("read_only_subagents", plan["mechanism_plan"]["optional_when_beneficial"])
274
+ self.assertIn("task_lifecycle", plan["mechanism_plan"]["forbidden"])
275
+
276
+ def test_intent_only_classifies_read_only_change_and_terminal_work(self) -> None:
277
+ repo, cfg = self.make_repo()
278
+ read_only = self.plan(repo, cfg, intent="Explain how authentication currently works")
279
+ self.assertEqual("read_only", read_only["task_kind"])
280
+ self.assertEqual("none", read_only["execution_mode"])
281
+ self.assertEqual(["analysis"], read_only["mechanism_plan"]["required"])
282
+ self.assertIn(
283
+ "read_only_subagents",
284
+ read_only["mechanism_plan"]["optional_when_beneficial"],
285
+ )
286
+ self.assertIn("commit_push", read_only["mechanism_plan"]["forbidden"])
287
+
288
+ change = self.plan(repo, cfg, intent="Fix the login redirect bug")
289
+ self.assertEqual("change", change["task_kind"])
290
+ self.assertEqual("direct", change["execution_mode"])
291
+ self.assertIn("commit_push", change["mechanism_plan"]["required"])
292
+
293
+ chinese_change = self.plan(repo, cfg, intent="请解决并优化登录跳转问题")
294
+ self.assertEqual("change", chinese_change["task_kind"])
295
+
296
+ forced_read_only = self.plan(
297
+ repo,
298
+ cfg,
299
+ intent="Fix wording in the explanation",
300
+ requested_task_kind="read_only",
301
+ )
302
+ self.assertEqual("read_only", forced_read_only["task_kind"])
303
+
304
+ terminal = self.plan(
305
+ repo,
306
+ cfg,
307
+ intent="reconcile the completed task ledger",
308
+ )
309
+ self.assertEqual("terminal_maintenance", terminal["task_kind"])
310
+ self.assertEqual(
311
+ ["analysis", "targeted_consistency_check", "commit_push"],
312
+ terminal["mechanism_plan"]["required"],
313
+ )
314
+ self.assertIn("task_lifecycle", terminal["mechanism_plan"]["forbidden"])
273
315
 
274
316
  def test_execution_mode_is_none_direct_isolated_or_parallel(self) -> None:
275
317
  repo, cfg = self.make_repo()
@@ -312,7 +354,7 @@ class AutoCodingProfileTests(unittest.TestCase):
312
354
  ap._validate_validation_plan(cfg, ap._validation_plan(cfg, ["contracts/api.yaml"]))
313
355
  self.assertIn("no validation route", str(context.exception))
314
356
 
315
- def test_final_gate_budget_is_hard_bounded_and_times_out(self) -> None:
357
+ def test_final_gate_budget_is_configurable_and_times_out(self) -> None:
316
358
  cfg = base_config()
317
359
  cfg["validation"] = {
318
360
  "on_unmapped": "error",
@@ -335,10 +377,42 @@ class AutoCodingProfileTests(unittest.TestCase):
335
377
  )
336
378
  self.assertIn("timed out", str(context.exception))
337
379
 
338
- cfg["validation"]["max_command_seconds"] = 121
380
+ cfg["validation"]["max_command_seconds"] = 240
381
+ cfg["validation"]["max_total_seconds"] = 300
382
+ self.assertEqual(
383
+ {"command_seconds": 240.0, "total_seconds": 300.0},
384
+ ap._final_gate_budget(cfg),
385
+ )
386
+
387
+ cfg["validation"]["max_total_seconds"] = 200
339
388
  with self.assertRaises(APError) as context:
340
389
  ap._final_gate_budget(cfg)
341
- self.assertIn("cannot exceed 120", str(context.exception))
390
+ self.assertIn("cannot exceed validation.max_total_seconds", str(context.exception))
391
+
392
+ def test_route_timeout_uses_smallest_matching_budget(self) -> None:
393
+ cfg = base_config()
394
+ cfg["commands"] = {"one": "true", "two": "true"}
395
+ cfg["validation"] = {
396
+ "on_unmapped": "error",
397
+ "max_command_seconds": 120,
398
+ "max_total_seconds": 180,
399
+ "routes": [
400
+ {
401
+ "name": "all-src",
402
+ "paths": ["src/**"],
403
+ "commands": ["one", "two"],
404
+ "timeout_seconds": 90,
405
+ },
406
+ {
407
+ "name": "python",
408
+ "paths": ["**/*.py"],
409
+ "commands": ["two"],
410
+ "timeout_seconds": 30,
411
+ },
412
+ ],
413
+ }
414
+ plan = ap._validation_plan(cfg, ["src/widget.py"])
415
+ self.assertEqual({"one": 90.0, "two": 30.0}, plan["command_timeouts"])
342
416
 
343
417
  def test_test_only_change_resolves_micro(self) -> None:
344
418
  repo, cfg = self.make_repo("tests/widget_test.py")
@@ -733,7 +733,7 @@ function testManagedEngineeringSyncIsControlledAndIdempotent() {
733
733
 
734
734
  const engineering = path.join(repo, "docs", "ENGINEERING.md");
735
735
  const initial = fs.readFileSync(engineering, "utf8");
736
- const startPattern = /<!-- auto-coding-skill:managed-workflow:start version=4\.1\.1 -->/;
736
+ const startPattern = /<!-- auto-coding-skill:managed-workflow:start version=4\.1\.2 -->/;
737
737
  const endMarker = "<!-- auto-coding-skill:managed-workflow:end -->";
738
738
  assert(startPattern.test(initial), "new projects should include a versioned managed workflow marker");
739
739
  assert(initial.includes(endMarker), "new projects should include the managed workflow end marker");
@@ -752,7 +752,7 @@ function testManagedEngineeringSyncIsControlledAndIdempotent() {
752
752
  const dryRun = run("node", [cli, "sync", "--projects", repo, "--dry-run", "--json"]);
753
753
  const dryResult = JSON.parse(dryRun.stdout).results[0];
754
754
  assert(dryResult.managedWorkflow.state === "stale", `dry-run should expose stale workflow state: ${dryRun.stdout}`);
755
- assert(dryResult.managedWorkflow.version === "4.1.1", "dry-run should expose the target workflow version");
755
+ assert(dryResult.managedWorkflow.version === "4.1.2", "dry-run should expose the target workflow version");
756
756
  assert(dryResult.actions.some(item => item.action === "would-update" && item.path === "docs/ENGINEERING.md"), "dry-run should plan the managed body update");
757
757
  assert(fs.readFileSync(engineering, "utf8") === customized, "dry-run must not write ENGINEERING.md");
758
758
 
@@ -762,14 +762,14 @@ function testManagedEngineeringSyncIsControlledAndIdempotent() {
762
762
  const updatedEnd = updated.indexOf(endMarker) + endMarker.length;
763
763
  assert(updated.slice(0, updatedStart) === outsideBefore, "sync must preserve frontmatter and content before the managed block byte-for-byte");
764
764
  assert(updated.slice(updatedEnd) === outsideAfter, "sync must preserve content after the managed block byte-for-byte");
765
- assert(updated.includes("version=4.1.1"), "sync should install the current managed workflow version");
765
+ assert(updated.includes("version=4.1.2"), "sync should install the current managed workflow version");
766
766
  assert(updated.includes("The frontmatter contract is:"), "sync should refresh stale managed workflow content");
767
767
 
768
768
  fillRequiredAccess(repo);
769
769
  const status = run("node", [cli, "status", "--projects", repo, "--json"]);
770
770
  const statusResult = JSON.parse(status.stdout).results[0];
771
771
  assert(statusResult.managedWorkflow.state === "current", `status should expose current managed workflow state: ${status.stdout}`);
772
- assert(statusResult.managedWorkflow.version === "4.1.1", "status should expose the installed managed workflow version");
772
+ assert(statusResult.managedWorkflow.version === "4.1.2", "status should expose the installed managed workflow version");
773
773
 
774
774
  const beforeSecondSync = fs.readFileSync(engineering, "utf8");
775
775
  const second = run("node", [cli, "sync", "--projects", repo, "--json"]);
@@ -797,7 +797,7 @@ function testLegacyEngineeringMigrationPreservesExistingBody() {
797
797
  assert(dryResult.actions.find(item => item.path === "docs/ENGINEERING.md")?.detail.includes("preserved-custom"), "custom legacy action should be labeled preserved-custom");
798
798
  run("node", [cli, "sync", "--projects", repo]);
799
799
  const migrated = fs.readFileSync(engineering, "utf8");
800
- assert(migrated.includes("version=4.1.1"), "legacy migration should insert the current managed workflow");
800
+ assert(migrated.includes("version=4.1.2"), "legacy migration should insert the current managed workflow");
801
801
  assert(migrated.includes(legacyNote), "legacy migration must preserve the complete existing body");
802
802
  const migratedStart = migrated.indexOf("<!-- auto-coding-skill:managed-workflow:start");
803
803
  const migratedEnd = migrated.indexOf(endMarker) + endMarker.length;
@@ -901,7 +901,7 @@ function testManagedAgentsMigrationReplacesWholeFileAndArchivesPreviousRules() {
901
901
  run("node", [cli, "sync", "--projects", repo]);
902
902
  const agents = path.join(repo, "AGENTS.md");
903
903
  const initial = fs.readFileSync(agents, "utf8");
904
- assert(initial.includes("managed-agents:start version=4.1.1"), "new projects should receive the versioned root AGENTS block");
904
+ assert(initial.includes("managed-agents:start version=4.1.2"), "new projects should receive the versioned root AGENTS block");
905
905
 
906
906
  const custom = [
907
907
  "# Project rules",
@@ -920,12 +920,12 @@ function testManagedAgentsMigrationReplacesWholeFileAndArchivesPreviousRules() {
920
920
 
921
921
  run("node", [cli, "sync", "--projects", repo]);
922
922
  const migrated = fs.readFileSync(agents, "utf8");
923
- assert(migrated.includes("managed-agents:start version=4.1.1"), "AGENTS migration should install the current managed block");
923
+ assert(migrated.includes("managed-agents:start version=4.1.2"), "AGENTS migration should install the current managed block");
924
924
  assert(!migrated.includes("Preserve this repository-specific rule exactly."), "root AGENTS must contain no project-specific tail");
925
925
  assert(!migrated.includes("must execute `commands.gate_full`"), "known official conflicting rule should be removed");
926
926
  const canonical = fs.readFileSync(path.join(repoRoot, "cli", "assets", "skill", "data", "templates", "bridges", "AGENTS.md"), "utf8");
927
927
  assert(migrated === canonical, "root AGENTS must be byte-identical to the packaged canonical file");
928
- const archive = path.join(repo, "docs", "archive", "workflow", "AGENTS.pre-4.1.1.md");
928
+ const archive = path.join(repo, "docs", "archive", "workflow", "AGENTS.pre-4.1.2.md");
929
929
  assert(fs.readFileSync(archive, "utf8").includes("Preserve this repository-specific rule exactly."), "previous AGENTS content must be archived once");
930
930
  const stable = fs.readFileSync(agents);
931
931
  run("node", [cli, "sync", "--projects", repo]);
@@ -984,7 +984,7 @@ function testEngineeringFrameworkAllowsProjectFactsButRejectsDuplicateWorkflow()
984
984
  const migrationResult = JSON.parse(migration.stdout).results[0];
985
985
  assert(migrationResult.managedWorkflow.state === "stale", "known duplicate sections should use controlled migration");
986
986
  assert(migrationResult.managedWorkflow.migrations.includes("engineering-section-delivery-flow"), "known section migration should be explicit");
987
- assert(migrationResult.actions.some(item => item.action === "would-archive" && item.path.includes("ENGINEERING.pre-4.1.1")), "controlled section cleanup should archive the previous ENGINEERING file");
987
+ assert(migrationResult.actions.some(item => item.action === "would-archive" && item.path.includes("ENGINEERING.pre-4.1.2")), "controlled section cleanup should archive the previous ENGINEERING file");
988
988
  run("node", [cli, "sync", "--projects", repo]);
989
989
  assert(!fs.readFileSync(engineering, "utf8").includes("Known obsolete workflow"), "known duplicate workflow section should be removed");
990
990
 
@@ -996,14 +996,14 @@ function testEngineeringFrameworkAllowsProjectFactsButRejectsDuplicateWorkflow()
996
996
  }
997
997
 
998
998
  function testReleaseVersionMarkersStayInSync() {
999
- const expected = "4.1.1";
999
+ const expected = "4.1.2";
1000
1000
  const pkg = JSON.parse(fs.readFileSync(path.join(repoRoot, "package.json"), "utf8"));
1001
1001
  const lock = JSON.parse(fs.readFileSync(path.join(repoRoot, "package-lock.json"), "utf8"));
1002
1002
  const policy = JSON.parse(fs.readFileSync(
1003
1003
  path.join(repoRoot, "src", "auto-coding-skill", "data", "policies", "workflow-migrations-v1.json"),
1004
1004
  "utf8",
1005
1005
  ));
1006
- assert(pkg.version === expected, "package version must match the 4.1.1 release");
1006
+ assert(pkg.version === expected, "package version must match the 4.1.2 release");
1007
1007
  assert(lock.version === expected && lock.packages[""].version === expected, "package-lock versions must match");
1008
1008
  assert(policy.managed_versions.engineering === expected && policy.managed_versions.agents === expected, "managed workflow versions must match");
1009
1009
  for (const rel of [
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@elvis1513/auto-coding-skill",
3
- "version": "4.1.1",
3
+ "version": "4.1.2",
4
4
  "description": "CLI installer for a delivery-first .agents workflow with adaptive Git isolation, routed changed-scope validation, risk-based review, push completion, and safe cleanup.",
5
5
  "type": "module",
6
6
  "bin": {
@@ -15,11 +15,14 @@
15
15
  "LICENSE"
16
16
  ],
17
17
  "scripts": {
18
+ "clean:cache": "find cli src -type d -name '__pycache__' -prune -exec rm -rf {} +",
19
+ "prepack": "npm run clean:cache",
20
+ "check:package": "npm run clean:cache && node cli/src/check-package.js",
18
21
  "test": "node -c cli/src/index.js && node -c cli/src/sync-assets.js && node -c cli/tests/index.test.js && node cli/tests/index.test.js && python3 -c \"import ast, pathlib; files = ['cli/assets/skill/scripts/ap.py', 'cli/assets/skill/scripts/core.py', 'cli/assets/skill/scripts/http_checks.py', 'cli/assets/skill/scripts/scaffold_templates.py']; [ast.parse(pathlib.Path(rel).read_text(encoding='utf-8'), filename=rel, feature_version=(3, 11)) for rel in files]; print('python311-asset-grammar-ok')\" && python3 -m py_compile cli/assets/skill/scripts/ap.py cli/assets/skill/scripts/core.py cli/assets/skill/scripts/http_checks.py cli/assets/skill/scripts/scaffold_templates.py && python3 cli/tests/ap_profiles_test.py && python3 cli/tests/ap_concurrency_test.py && python3 -c \"import pathlib, tomllib; files = sorted(pathlib.Path('cli/assets/agents').glob('*.toml')); [tomllib.loads(path.read_text(encoding='utf-8')) for path in files]; print('asset-agent-toml-ok')\" && find cli -type d -name '__pycache__' -prune -exec rm -rf {} +",
19
22
  "test:src": "python3 -c \"import ast, pathlib; files = ['src/auto-coding-skill/scripts/ap.py', 'src/auto-coding-skill/scripts/core.py', 'src/auto-coding-skill/scripts/http_checks.py', 'src/auto-coding-skill/scripts/scaffold_templates.py']; [ast.parse(pathlib.Path(rel).read_text(encoding='utf-8'), filename=rel, feature_version=(3, 11)) for rel in files]; print('python311-src-grammar-ok')\" && python3 -m py_compile src/auto-coding-skill/scripts/ap.py src/auto-coding-skill/scripts/core.py src/auto-coding-skill/scripts/http_checks.py src/auto-coding-skill/scripts/scaffold_templates.py && python3 -c \"import pathlib, tomllib; files = sorted(pathlib.Path('src/agents').glob('*.toml')); [tomllib.loads(path.read_text(encoding='utf-8')) for path in files]; print('src-agent-toml-ok')\" && find src -type d -name '__pycache__' -prune -exec rm -rf {} +",
20
23
  "check:assets": "node cli/src/sync-assets.js --check",
21
24
  "sync-assets": "node cli/src/sync-assets.js",
22
- "release:check": "npm run check:assets && npm run test:src && npm run test && npm pack --dry-run"
25
+ "release:check": "npm run check:assets && npm run test:src && npm run test && npm run check:package"
23
26
  },
24
27
  "license": "MIT",
25
28
  "publishConfig": {