rule-delivery-gate
Pre-commit gate over staged trait and role injections. Use when composing a role that carries the `ai-hats-dev` trait, or when diagnosing why the rule-delivery hook blocked a commit.
- License 许可
- MIT
- 所属仓库
- muratovv/ai-hats
- 最近更新
- 2026-11-05
- Stars
- ★ 3
- Forks
- 1
01 前言 Front Matter
- name
- rule-delivery-gate
- ai_hats
- { "type": "skill", "schema_version": 1 }
- license
- MIT
- description
- Pre-commit gate over staged trait and role injections. Use when composing a role that carries the `ai-hats-dev` trait, or when diagnosing why the rule-delivery hook blocked a commit.
02 资源文件 1
03 SKILL.md 正文
Rule Delivery Gate
Pure-infrastructure hook-carrier skill. It contributes one git pre-commit hook
that enforces the rule-delivery contract: a see rule X pointer in a
delivered trait/role injection must name a rule the agent can actually read.
There is no agent-side decision logic here — the value is delivered entirely
through composition.
What it gates
- Scope: fires only when a commit stages a trait/role injection —
{ai_hats_library,library}/**/config.yaml: the monorepo package, or the pre-monorepolibrary/the e2e fixtures use. The check then scans that library root at the checkout, because a pointer's validity depends on rule existence across the library, not a single diff. With no library root at the checkout the gate is a loud, journalled skip. - Not a consumer's
libraries/: this skill ships in theai-hats-devlayer, which is addressed to the ai-hats repository, not to a consuming project. - Invariant: every
see rule Xmust resolve to a rule that exists in the library. A pointer to a non-existent rule blocks the commit. - Effect: the same pure function the G2 unit test uses
(
python -m ai_hats.rule_delivery <library root>) returns non-zero → commit blocked with the offendingsource: see rule Xlines.
Because the scope is the changed files, the gate only fires on commits that touch an injection — it never retro-blocks the pre-existing library.
Who gets it
Whatever composes this skill, and nothing else: the hook is written into
.githooks/pre-commit.d/ at composition time, so a role that does not carry
it never sees the gate. The server-side counterpart is the G2 unit test, run by CI on every PR/push to master.
How to bypass
Fix the pointer — create the missing rule in the library or drop the pointer — or, after
confirming the pointer is intentional, ask the supervisor to skip the gate for a
single commit: AI_HATS_RULE_DELIVERY_ACK=1, exported in the shell that runs the
commit. Writing it onto your own command line is refused — see safety-guard.
If python / the ai_hats package is unavailable the hook is a loud no-op
(fail-open): it prints a SKIPPED notice and allows the commit, so a missing dev
tool never wedges work. Inside an ai-hats dev/agent env the package is present,
so the gate is live.
Overrides
AI_HATS_RULE_DELIVERY_ACK=1— allow the current commit despite findings.AI_HATS_RULE_DELIVERY_CMD— override the checker invocation (default:-m ai_hats.rule_deliveryunder the checkout's own.venv/bin/python3, elsepython3on PATH); used by the test suite to inject a stub.
04 同仓库更多 Skills
tool-evaluation-protocol
Time-bounded protocol for evaluating new tools, libraries, or methodologies before adoption. Use when evaluating a candidate dependency for the project, researching a methodology (e.g. ML/LLM frameworks, testing approaches), or following up on a hype-driven recommendation.
systematic-debugging
4-phase bug-fix protocol (evidence, pattern, hypothesis, verify). Use when investigating any error, failure, or unexpected behavior, handling bug reports from users or monitoring, or diagnosing test failures with non-obvious causes.
skill-template
Canonical template and validation guide for ai-hats skills. Use when creating a new skill, reviewing an existing one for structural compliance, or deciding which pattern fits a new behavior.
skill-optimization
Audit and refactor library components to eliminate redundancy and staleness. Use during periodic library maintenance, after adding multiple new skills or rules, when a role's token budget is too high, or after retrospectives identify component issues.
silent-fallback-gate
PreToolUse hook — refuses an Edit or Write that adds an inert broad except. Use when a Python edit was blocked for a silent fallback.
retro-to-framework
Convert an observation about agent behavior — from a retrospective or from a single incident — into a framework-level change (hook, rule, skill, skill update). Use after a retrospective identifies problems that are not project-specific, after an incident where an agent misapplied a component it had read, when CLAUDE.md band-aids accumulate (more than 3 per-project rules that could be generic), or when the same problem recurs across multiple projects.