PromptsEdge
FreeGitHubOtherScripts

Measure Agent Compliance

Your AI checks if agents truly follow given skills, rules, and definitions by running scenarios at different prompt strictness levels. It captures and classifies tool call sequences and generates detailed compliance reports with timelines.

affaan-m on GitHub

Curated by PromptsEdge from a public repo · MIT license. All credit goes to the author.

What is this skill?

Measures whether coding agents actually follow skills, rules, or agent definitions by:

When to use

  • User runs /skill-comply <path>
  • User asks "is this rule actually being followed?"
  • After adding new rules/skills, to verify agent compliance
  • Periodically as part of quality maintenance
  1. Auto-generating expected behavioral sequences (specs) from any .md file
  2. Auto-generating scenarios with decreasing prompt strictness (supportive → neutral → competing)
  3. Running claude -p and capturing tool call traces via stream-json
  4. Classifying tool calls against spec steps using LLM (not regex)
  5. Checking temporal ordering deterministically
  6. Generating self-contained reports with spec, prompts, and timelines

Supported Targets

  • Skills (skills/*/SKILL.md): Workflow skills like search-first, TDD guides
  • Rules (rules/common/*.md): Mandatory rules like testing.md, security.md, git-workflow.md
  • Agent definitions (agents/*.md): Whether an agent gets invoked when expected (internal workflow verification not yet supported)

Usage

bash
# Full run
uv run python -m scripts.run ~/.claude/rules/common/testing.md

# Dry run (no cost, spec + scenarios only)
uv run python -m scripts.run --dry-run ~/.claude/skills/search-first/SKILL.md

# Custom models
uv run python -m scripts.run --gen-model haiku --model sonnet <path>

Key Concept: Prompt Independence

Measures whether a skill/rule is followed even when the prompt doesn't explicitly support it.

Report Contents

Reports are self-contained and include:

  1. Expected behavioral sequence (auto-generated spec)
  2. Scenario prompts (what was asked at each strictness level)
  3. Compliance scores per scenario
  4. Tool call timelines with LLM classification labels

Advanced (optional)

For 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.

Related skills

View all →