Creates and maintains concise AGENTS.md files that reference local agent skills.
npx skills add https://github.com/klh/skills --skill agents-md-klh
Please help me install this Agent Skill. Skill: agents-md Source: https://github.com/klh/skills/tree/main/.well-known/agent-skills/agents-md Command: npx skills add https://github.com/klh/skills --skill agents-md-klh
.claude/skills and plugin directories for SKILL.md files so AGENTS.md can reference skills instead of duplicating their content. - Style enforcement: Applies writing rules (headers and bullets only, no filler, reference-don't-duplicate) and lists anti-patterns to avoid. - Use Case: When setting up a new repository, ask the agent to create an AGENTS.md; it discovers installed skills, writes concise sections, and symlinks CLAUDE.md to AGENTS.md. ## Quick Start Ask the agent to create an AGENTS.md for this project that references the local skills and follows the minimal documentation style.High-intent search queries and answers about installing and using this skill.
Create AGENTS.md at the project root with required sections: Package Manager, Commit Attribution, Key Conventions, and Local Skills. Use headers and bullets only, reference skills instead of duplicating content, and symlink CLAUDE.md to AGENTS.md.
Required sections are Package Manager, Commit Attribution, Key Conventions, and Local Skills referencing discovered SKILL.md files. Optional sections like API route templates or CLI command tables should be added only when truly needed.
Discover skills by finding SKILL.md files under .claude/skills or plugin directories, read each frontmatter, then add short references like "Use `db-migrate` skill. See `.claude/skills/db-migrate/SKILL.md`" instead of copying skill content.
No. AGENTS.md is the canonical agent-facing documentation, and CLAUDE.md should be a symlink to it created with `ln -s AGENTS.md CLAUDE.md`. This keeps a single source of truth for agent instructions.
Avoid welcome intros, pleasantries, long prose paragraphs, obvious instructions like "write clean code", and content duplicated from skills. Keep instructions as headers and bullets, state what to do without explaining why, and trust agent capabilities.
silentFellowGenerates and updates AGENTS.md and CONTEXT.md documentation through interactive repo analysis.
intellectronicaInitialize a Git repository with optional agent instructions and .gitignore templates.
parisgroup-aiDiagnose and fix drift between memory sources and the codebase.
fractaryDetect hybrid agents by analyzing execution patterns in .claude/agents prompts.