skill-comply

Automate compliance testing by generating observable behavior specs and scenario-based tests. The skill executes agent sessions with Claude, records tool-call timelines, and produces a final compliance report with hook-promotion guidance.

Scenario-based compliance verification quantifies whether skills and rules are actually followed by generating three-tier prompts with varying levels of support, executing them, and classifying tool calls against spec steps to compute per-scenario compliance rates.

Yes, you can measure compliance rates by running Claude-based agent sessions. The workflow classifies tool calls against the spec, computes per-scenario compliance rates, and reports hook-promotion recommendations in the final output.

Generate observable behavior specs automatically from skill content using the skill's internal scripts. It requires pyyaml and a Claude-based workflow to parse the definitions and produce the three-tier scenario prompts for testing.

Skill compliance testing requires the pyyaml dependency and a Claude-based workflow to generate specs, scenarios, classify tool calls, and produce the final report. It applies to testing skills, rules, and agent definitions.

The best way to visualize rule adherence is by generating scenario-based tests that record tool-call timelines. This approach classifies actions against the spec steps and produces clear, transparent compliance reports with hook-promotion guidance.