Skill Comply
Measure whether your agents actually follow the skills, rules, and definitions you gave them, with auto-generated scenarios and compliance reports.
- What
- Measure whether your agents actually follow the skills, rules, and definitions you gave them, with auto-generated scenarios and compliance reports.
- Cost
- Free
- Needs
- Use "Skill Comply" with your Muse.
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Curated by Skill Harbor: an automated compliance measurement workflow for coding agents, from the ECC framework and adaptable to any Claude Code setup. It auto-generates expected behavioral sequences (specs) from any .md file (workflow skills, mandatory rules, agent definitions), builds test scenarios at three prompt strictness levels (supportive, neutral, competing), runs claude -p while capturing tool call traces via stream-json, classifies tool calls against spec steps with an LLM (not regex), checks temporal ordering deterministically, and produces self-contained reports with the spec, the scenario prompts, compliance scores, and full tool call timelines. Run via uv run python -m scripts.run (a no-cost --dry-run mode generates spec plus scenarios only; custom models supported). The key concept is prompt independence: whether a skill or rule is followed even when the prompt does not explicitly support it. From the affaan-m/ECC repository (MIT). Honest caveats: requires the skill's companion scripts and a claude -p style CLI with stream-json output; full runs consume model tokens, so start with --dry-run. Verifying internal agent-definition workflows is not yet supported. Skill Harbor never reviews the code, review it yourself before use.
Version:
Install
Copy the install package below, then paste it into MuseThe install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Use "Skill Comply" with your Muse. Prerequisites: uv, the skill's companion scripts (scripts/run from the repo), and a claude -p style CLI with stream-json output. 1. Open the skill: https://github.com/affaan-m/ECC/blob/main/skills/skill-comply/SKILL.md and copy the full SKILL.md text. 2. Paste it into a chat with Muse and add: "Check whether my agents actually follow [path to your SKILL.md or rules file]." 3. Start with a dry run: uv run python -m scripts.run --dry-run <path> (no cost), then run the full measurement. Tip: run it after adding new rules or skills, and periodically as quality maintenance; it measures compliance, it does not fix the skill. Safety: a skill is plain-text instructions; it runs nothing by itself. Full runs spend model tokens on scenario agents; scope the target file before launching.
Saved to your recent installs. Find it anytime on /connect.
Questions
How do I install a build?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the build up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a build before installing it.