@elvis1513/auto-coding-skill 4.1.1 → 4.1.3

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 +9,10 @@ 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 4.1.1 consolidates the 4.x delivery-first workflow into one canonical
12
+ Version 4.1.3 closes classification bypasses and keeps current-task changes direct
13
+ without weakening dirty-workspace isolation. Version 4.1.2 made mechanisms required,
14
+ model-selectable when beneficial, or forbidden for the current task. Version
15
+ 4.1.1 consolidated the 4.x delivery-first workflow into one canonical
13
16
  repository contract. Version 4.0 replaced the 3.x governance-first defaults with progressive
14
17
  guardrails. Clean single-writer work stays on the current branch. Worktrees,
15
18
  parallel fixers, fingerprint review, durable design records, and stronger
@@ -65,17 +68,27 @@ The result includes:
65
68
  - `execution_mode=isolated`: dirty checkout, explicit isolation, or multiple writers.
66
69
  - `execution_mode=none`: no task signal or change; create no branch/worktree.
67
70
  - `review_required` and `design_required`: risk-based escalation decisions.
71
+ - `task_kind`: `read_only`, `change`, `terminal_maintenance`, or internal `none`.
68
72
  - `mechanism_plan.required`: the complete minimum mechanism set for the task.
69
- - `mechanism_plan.not_required`: mechanisms that stay off unless the user asks.
73
+ - `mechanism_plan.optional_when_beneficial`: model-selectable mechanisms whose
74
+ expected benefit must exceed coordination cost.
75
+ - `mechanism_plan.forbidden`: mechanisms that stay off unless the user overrides.
76
+ - `optional_agents`: model-selectable explorer/docs/browser candidates; they are
77
+ never automatic stages.
70
78
  - an adaptive agent plan that does not force fan-out for ordinary work.
71
79
 
72
80
  Clean serial work proceeds directly on the current branch with normal Git; it does
73
81
  not create machine lifecycle state. Use `task-start` only when classification
74
82
  requires isolation/review or the user explicitly requests lifecycle tracking:
75
83
 
84
+ When a direct task materially changes scope after writing, reclassify with
85
+ `--continue-direct` and repeat every current task `--planned-path`. Undeclared
86
+ dirty paths, another writer, or mandatory isolation still selects a worktree. If
87
+ the new plan requires review lifecycle, reuse `--continue-direct` on `task-start`.
88
+
76
89
  ```bash
77
90
  python3 docs/tools/autopipeline/ap.py task-start T0001 \
78
- --owned-path backend/internal/orders
91
+ --owned-path backend/internal/orders --force-lifecycle
79
92
  python3 docs/tools/autopipeline/ap.py commit-push T0001 \
80
93
  --msg "T0001: fix order retry"
81
94
  ```
@@ -90,8 +103,10 @@ Dirty or parallel work receives an isolated task branch/worktree:
90
103
  python3 docs/tools/autopipeline/ap.py task-start T0002 \
91
104
  --owned-path backend --review-required
92
105
  # implement in the printed worktree
106
+ python3 docs/tools/autopipeline/ap.py task-status T0002 --json
93
107
  python3 docs/tools/autopipeline/ap.py task-review T0002 \
94
- --verdict approved --diff-fingerprint "$SHA256"
108
+ --verdict approved --diff-fingerprint "$CURRENT_DIFF_FINGERPRINT" \
109
+ --reviewer "$REVIEWER_ID"
95
110
  python3 docs/tools/autopipeline/ap.py commit-push T0002 \
96
111
  --msg "T0002: bounded change"
97
112
  python3 docs/tools/autopipeline/ap.py task-integrate T0002
@@ -122,6 +137,7 @@ validation:
122
137
  - name: backend
123
138
  paths: ["backend/**", "contracts/**"]
124
139
  commands: [backend_fast]
140
+ timeout_seconds: 90
125
141
  - name: frontend
126
142
  paths: ["frontend/**"]
127
143
  commands: [frontend_fast]
@@ -152,9 +168,10 @@ closure gate is limited to one stable-diff run. Full regression, Docker, builds,
152
168
  Jenkins, deployment, API verification, and target checks remain explicit
153
169
  diagnostics outside normal closure.
154
170
 
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.
171
+ The recommended final closure defaults are 120 seconds per command and 180
172
+ seconds total. Projects may raise them for measured affected-scope checks or set
173
+ a smaller `timeout_seconds` on a route. A timeout calls for a narrower route and
174
+ never triggers a broader fallback.
158
175
 
159
176
  ## Risk and subagent policy
160
177
 
@@ -192,11 +209,38 @@ project-specific facts outside the managed ENGINEERING block. Root `AGENTS.md` i
192
209
  replaced as a whole; its previous content is archived under
193
210
  `docs/archive/workflow/` as historical, non-authoritative context.
194
211
 
212
+ ## What changed in 4.1.3
213
+
214
+ - Made explicit task kinds fail closed: `none` is internal-only and terminal
215
+ maintenance cannot be applied to code or unrelated documentation paths.
216
+ - Added guarded `--continue-direct` reclassification for already-owned changes;
217
+ unknown dirty paths still require isolation.
218
+ - Separated intent risk hints from path/rule-confirmed high risk so ordinary login
219
+ or payment UI work no longer automatically receives a heavy lifecycle.
220
+ - Aligned generated Agent plans with their JSON Schema and exposed optional
221
+ read-only Agent candidates without auto-dispatching them.
222
+ - Clarified delegated-fixer worktree ownership, reviewer identity, and explorer
223
+ source authority.
224
+
225
+ ## What changed in 4.1.2
226
+
227
+ - Added explicit read-only, change, and terminal-maintenance task kinds so
228
+ intent-only questions cannot accidentally trigger code-delivery machinery.
229
+ - Split mechanism planning into required, model-selectable, and forbidden sets;
230
+ read-only discovery agents remain available when they save time or add needed
231
+ expertise.
232
+ - Made 120/180-second gate budgets recommended defaults, added route-level
233
+ timeouts, and kept project-specific affected-scope overrides valid.
234
+ - Rejected unnecessary lifecycle creation before access checks, fetching, locks,
235
+ or branch work; corrected the review lifecycle examples.
236
+ - Added a release-package assertion and prepack cache cleanup so generated Python
237
+ caches cannot leak into the npm tarball.
238
+
195
239
  ## What changed in 4.1.1
196
240
 
197
241
  - Added a machine-readable minimum mechanism plan and rejected unnecessary clean
198
242
  serial task lifecycles unless explicitly requested.
199
- - Added hard per-command and total time ceilings for the final changed-scope gate.
243
+ - Added bounded per-command and total time budgets for the final changed-scope gate.
200
244
  - Made root `AGENTS.md` the sole behavioral protocol, reduced `SKILL.md` to
201
245
  invocation guidance, and reduced `ENGINEERING.md` to project configuration and
202
246
  facts.
@@ -8,7 +8,7 @@ developer_instructions = """
8
8
 
9
9
  工作要求:
10
10
  1. 优先用 rg / git / 项目脚本快速定位真实入口,再向下追踪调用链。
11
- 2. 先读项目已有文档、配置、路由、任务记录;涉及结构/优化时,仅在结构标准、健康基线或优化 backlog 已存在时读取,不因可选文档缺失而阻塞。
11
+ 2. 以当前代码、测试、schema、migration 和运行配置为事实源;仅按任务需要读取 ENGINEERING 或当前接口文档,任务记录只在明确追溯历史时读取。涉及结构/优化时,仅在结构标准、健康基线或 optimization backlog 已存在时读取,不因可选文档缺失而阻塞。
12
12
  3. 必要时运行 `ap.py classify --scope auto`,说明执行模式、风险和是否确实需要设计/reviewer;普通任务默认主代理直接完成,本地 gate 始终保持 changed-scope。
13
13
  4. 输出根因时说明当前代码应归属的层级、现有同类模块、可复用 helper / 组件 / 脚本,以及是否触及 `structure.layer_rules`。
14
14
  5. 如果根因依赖外部库、框架或版本行为,标记需要 docs_researcher / Context7 继续核对,不要凭记忆猜。
@@ -7,9 +7,9 @@ developer_instructions = """
7
7
  你负责小而准的实现修改。
8
8
 
9
9
  工作要求:
10
- 1. 只在问题已经相对明确,并收到 task_id、execution_mode、owned_paths、depends_on 和 acceptance 后再动手;只有 delegated/isolated 工作才要求完整任务元数据。
11
- 2. direct 模式只用于单写入者且主代理确认当前 checkout 干净的当前分支;isolated 模式必须确认当前目录和 task_branch 与分配的注册 worktree 一致。不满足时停止并上报。
12
- 3. 一个 worktree 只能有一个写入者。只修改 owned_paths,不进入或修改其他任务工作树;发现未知改动立即停止,不 restore/reset/stash/clean。direct 模式期间主代理不得并发编辑同一 checkout。
10
+ 1. 只在问题已经相对明确,并收到 task_id、execution_mode、owned_paths、depends_on 和 acceptance 后再动手。
11
+ 2. fixer 属于 delegated writer,必须使用注册的 isolated worktree;direct 模式保留给干净 checkout 中独立开发的主代理。确认当前目录和 task_branch 与分配一致,否则停止并上报。
12
+ 3. 一个 worktree 只能有一个写入者。只修改 owned_paths,不进入其他任务工作树;发现未知改动立即停止,不 restore/reset/stash/clean。
13
13
  4. 坚持“最小可辩护修改”原则。
14
14
  5. 除非明确要求,否则不要重构、不做大面积改名、不改无关文件。
15
15
  6. 先跟随项目现有模式、helper、组件、脚本和文档约束,不新造不必要抽象。
@@ -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 after material change. For already-direct work, add `--continue-direct`
35
+ with every task path and reuse it on `task-start` if needed; unknown dirt isolates.
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] [--continue-direct] [--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> --reviewer <REVIEWER_ID>
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 }
@@ -382,6 +384,7 @@
382
384
  "policies",
383
385
  "assignment_contract",
384
386
  "result_contract",
387
+ "optional_roles",
385
388
  "stages",
386
389
  "constraints"
387
390
  ],
@@ -407,6 +410,9 @@
407
410
  "result_contract": {
408
411
  "$ref": "#/$defs/stringArray"
409
412
  },
413
+ "optional_roles": {
414
+ "$ref": "#/$defs/stringArray"
415
+ },
410
416
  "stages": {
411
417
  "type": "array",
412
418
  "items": {
@@ -452,9 +458,12 @@
452
458
  "classification_files",
453
459
  "intent_provided",
454
460
  "intent_categories",
461
+ "intent_risk_candidates",
462
+ "task_kind",
455
463
  "profile",
456
464
  "categories",
457
465
  "recommended_agents",
466
+ "optional_agents",
458
467
  "agent_plan",
459
468
  "mechanism_plan"
460
469
  ],
@@ -480,6 +489,12 @@
480
489
  "intent_categories": {
481
490
  "$ref": "#/$defs/stringArray"
482
491
  },
492
+ "intent_risk_candidates": {
493
+ "$ref": "#/$defs/stringArray"
494
+ },
495
+ "task_kind": {
496
+ "enum": ["none", "read_only", "change", "terminal_maintenance"]
497
+ },
483
498
  "profile": {
484
499
  "enum": [
485
500
  "micro",
@@ -493,6 +508,9 @@
493
508
  "recommended_agents": {
494
509
  "$ref": "#/$defs/stringArray"
495
510
  },
511
+ "optional_agents": {
512
+ "$ref": "#/$defs/stringArray"
513
+ },
496
514
  "agent_plan": {
497
515
  "$ref": "#/$defs/agentPlan"
498
516
  },
@@ -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": "4.1.3",
5
+ "engineering": "4.1.3"
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: "4.1.3"
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=4.1.3 -->
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=4.1.3 -->
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
 
@@ -49,11 +51,11 @@ failure may continue in the same conversation/task without a second lifecycle.
49
51
 
50
52
  ## Git and parallel work
51
53
 
52
- - One writer in a clean checkout works directly on the current branch without a
53
- task lifecycle.
54
+ - One writer in a clean checkout works directly without a lifecycle. On
55
+ reclassification, `--continue-direct` requires every dirty path to be declared.
54
56
  - Existing unrelated changes, another writer, or configured mandatory isolation
55
57
  requires a registered task branch/worktree.
56
- - Every writing agent owns a distinct task ID, worktree, writer lease,
58
+ - Every delegated fixer or parallel writer owns a task ID/worktree, writer lease,
57
59
  non-overlapping `owned_paths`, and prerequisite commit SHAs.
58
60
  - Never let multiple writing agents share a checkout. Never restore, reset, stash,
59
61
  clean, overwrite, or commit unknown changes.
@@ -64,14 +66,15 @@ 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.
72
75
  - Use parallel fixers only for bounded, dependency-free, non-overlapping units.
73
- - Require fingerprinted review only for high-risk, contract-crossing, parallel,
74
- or explicitly configured work.
76
+ - Require fingerprinted review only for path/rule-confirmed high-risk,
77
+ contract-crossing, parallel, or configured work; intent words are candidates only.
75
78
  - Review blocks only defects introduced or worsened in the promised scope.
76
79
  Adjacent existing issues are non-blocking follow-ups.
77
80
  - Semantic changes invalidate approval. Mechanical documentation-only corrections
@@ -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.