What problem does it solve? AGENTS.md files often become stale, incomplete, or missing key commands, leaving AI coding agents without the project context they need to work effectively. This Skill systematically finds every AGENTS.md file in a repository, grades its quality, and applies targeted improvements after your approval. ## Core Features & Use Cases - Repository-wide discovery: Locates project root, package-level, local override, and global AGENTS.md files, including monorepo setups. - Quality scoring and reporting: Grades each file A-F across criteria like commands, architecture clarity, currency, and actionability, then outputs a structured quality report before any changes. - Targeted updates with diffs: Proposes minimal, high-value additions (missing build commands, gotchas, environment setup) shown as diffs, and applies them only after user confirmation. - Use Case: After a major refactor, run an audit to find that your root AGENTS.md still references an old test command and is missing the new environment variable setup, then apply the suggested fixes in one pass. ## Quick Start Ask the agent to audit all AGENTS.md files in this repository, produce a quality report, and suggest improvements.
