Audit AGENTS.md files for token efficiency, actionability, and scope placement.
npx skills add https://github.com/eivindeb/agent-skills --skill audit-agents-md
Please help me install this Agent Skill. Skill: audit-agents-md Source: https://github.com/eivindeb/agent-skills/tree/main/audit-agents-md Command: npx skills add https://github.com/eivindeb/agent-skills --skill audit-agents-md
This Skill streamlines your agent's instructions by auditing AGENTS.md files for token efficiency, completeness, scope hygiene, and actionability, ensuring your AI agents operate with maximum impact and minimal wasted context.
AGENTS.md files, skills, and Cursor rules.AGENTS.md files to ensure all agent instructions are concise, relevant, and correctly scoped, preventing confusion and improving agent performance.Audit all AGENTS.md files in the current repository for token efficiency and actionability.
High-intent search queries and answers about installing and using this skill.
To optimize AGENTS.md files for token efficiency, you can audit them to identify redundant, vague, or duplicated directives, ensuring your AI agents operate with maximum impact and minimal wasted context.
Auditing AI agent instructions for actionability involves analyzing directives to ensure they are specific and unambiguous, checking that every instruction provides clear, actionable guidance for the agent to follow.
Determining correct scope placement requires analyzing directives to recommend optimal placement across AGENTS.md files, skills, and Cursor rules, preventing duplication and improving overall agent performance.
Your AI agent might ignore AGENTS.md instructions if they are verbose, lack unambiguous actionability, or duplicate existing skills, which causes a drop in token ROI and overall instruction effectiveness.
Yes, you can check AGENTS.md files for completeness to highlight missing high-value directives, such as verification matrices or architectural boundaries, ensuring all instructions are concise and relevant before refactoring.
k97Refactor CLAUDE.md files into modular progressive disclosure structures.
coco-researchCreate persistent Cursor rules with YAML frontmatter and glob patterns.
WrongStackCreate frontmatter-driven SKILL.md files for WrongStack skills.
axiomanticCompress long instruction files while preserving capabilities and safety constraints.