snap

Create, update or improve the local AGENTS.md for the current package. Use when setting up a new package, after significant code changes, or when new process knowledge should be captured. Imports `AGENTS-shared.md` and extends it with package-specific context.

Audit `AGENTS-shared.md` and all convention files for quality and correctness. Checks for internal consistency, gaps, redundancies and alignment between conventions and agent instructions.

Suite-level audit of skill files and guidelines. Runs snap-check-skills, snap-check-guidelines and reflect in sequence.

Audit all skill files for quality and consistency. Checks each skill against project conventions and checks across all skills for gaps, redundancies, conflicts, ambiguities, scope issues, stale references and naming discrepancies.

Review the current conversation for learnings that should improve existing skills, AGENTS.md files or memory. Use at the end of a session to capture non-obvious findings before they're lost.

Scan the whole project for improvements in implementation, architecture, API, clarity, refactoring and guideline compliance. Records findings as TODO comments in source or as bullets in the README TODO section.

Check all source comments in the package for typos, correctness, clarity, brevity and policy compliance. Fixes typos immediately; lists content changes for per-item decision.

Review and improve the README of the current project. Use when the user asks to check, audit, verify or update a README, especially when correctness against the public API matters.

Full audit of the current package: comments, README and AGENTS.md. Runs check-comments, check-readme, check-agents-md and reflect in sequence.