agentic-legibilitylisted

# Agentic Legibility ## Overview Measure how legible a repository is to coding agents using seven repo-visible metrics: bootstrap self-sufficiency, task entrypoints, validation harness, lint and format gates, agent repo map, structured docs, and decision records. The scorer can auto-discover nested scopes such as `client/`, `server/`, or language-specific subtrees when the repo clearly routes work into a self-contained subsystem with its own manifests, commands, or agent docs. Keep the main score limited to evidence present in version control. Do not claim branch protection, CI reliability, team habits, or other external-system properties unless the user asks for a separate operational review. Use metric selection when you want to split scoring across multiple hosted-shell or Responses API runs and aggregate the JSON results upstream. ## Hosted Use This folder is the skill bundle you upload or register when you want to mount `agentic-legibility` into hosted shell. Public docs for that workflow: - [Skills guide](https://developers.openai.com/api/docs/guides/tools-skills/) - [Shell guide: Attach skills](https://developers.openai.com/api/docs/guides/tools-shell/#attach-skills) - [Skills in OpenAI API cookbook](https://developers.openai.com/cookbook/examples/skills_in_api/) ## Workflow 1. Run `scripts/score_repo.py` against the target repository. 2. Check the reported `evaluated_scope` and `discovered_scopes`. 3. If the repo has multiple plausible sub-scopes, inspect t