Skill Comply

1---2name: skill-comply3description: Visualize whether skills, rules, and agent definitions are actually followed — auto-generates scenarios at 3 prompt strictness levels, runs agents, classifies behavioral sequences, and reports compliance rates with full tool call timelines4origin: ECC5tools: Read, Bash6---7 8# skill-comply: Automated Compliance Measurement9 10Measures whether coding agents actually follow skills, rules, or agent definitions by:111. Auto-generating expected behavioral sequences (specs) from any .md file122. Auto-generating scenarios with decreasing prompt strictness (supportive → neutral → competing)133. Running `claude -p` and capturing tool call traces via stream-json144. Classifying tool calls against spec steps using LLM (not regex)155. Checking temporal ordering deterministically166. Generating self-contained reports with spec, prompts, and timelines17 18## Supported Targets19 20- **Skills** (`skills/*/SKILL.md`): Workflow skills like search-first, TDD guides21- **Rules** (`rules/common/*.md`): Mandatory rules like testing.md, security.md, git-workflow.md22- **Agent definitions** (`agents/*.md`): Whether an agent gets invoked when expected (internal workflow verification not yet supported)23 24## When to Activate25 26- User runs `/skill-comply <path>`27- User asks "is this rule actually being followed?"28- After adding new rules/skills, to verify agent compliance29- Periodically as part of quality maintenance30 31## Usage32 33```bash34# Full run35uv run python -m scripts.run ~/.claude/rules/common/testing.md36 37# Dry run (no cost, spec + scenarios only)38uv run python -m scripts.run --dry-run ~/.claude/skills/search-first/SKILL.md39 40# Custom models41uv run python -m scripts.run --gen-model haiku --model sonnet <path>42```43 44## Key Concept: Prompt Independence45 46Measures whether a skill/rule is followed even when the prompt doesn't explicitly support it.47 48## Report Contents49 50Reports are self-contained and include:511. Expected behavioral sequence (auto-generated spec)522. Scenario prompts (what was asked at each strictness level)533. Compliance scores per scenario544. Tool call timelines with LLM classification labels55 56### Advanced (optional)57 58For users familiar with hooks, reports also include hook promotion recommendations for steps with low compliance. This is informational — the main value is the compliance visibility itself.