Audit agent configuration and AI maintainability for instruction drift and control gaps.
npx skills add https://github.com/OiAnthony/.agents --skill health-oianthony
Please help me install this Agent Skill. Skill: health Source: https://github.com/OiAnthony/.agents/tree/main/skills/health Command: npx skills add https://github.com/OiAnthony/.agents --skill health-oianthony
This Skill finds when AI coding agents drift from project instructions, waste context on broken or redundant surfaces, or accumulate maintainability risks that make future work unreliable.
Ask the health skill to run a budget-aware audit of the current repository and report instruction drift, verification gaps, MCP or hook issues, and maintainability risks.
High-intent search queries and answers about installing and using this skill.
To audit agent configuration, run a budget-aware diagnostic pass that checks project and runtime instruction layers for conflicts, missing trust settings, hook gaps, and MCP issues. It separates control-layer failures from maintainability risks and produces copy-pasteable remediation actions.
Your Claude agent may ignore project instructions due to instruction drift, control gaps, or verification coverage issues. An agent configuration audit checks hooks, MCP servers, and skills to identify the exact misaligned layer and provides actionable fixes to restore rule adherence.
The best way to check AI maintainability is running a health audit that surfaces hotspots, oversized context, stale references, and documentation drift. This identifies maintainability risks causing code rot or repeated AI mistakes and outputs targeted remediation steps.
Yes, you need python3, jq, git, and npx installed as dependencies to run the agent health check. These tools execute the scripts required to audit repository agent configuration, verification coverage, and AI maintainability risks.
You can use this audit for Claude, Codex, and Pi workflows. The health check applies to hooks, MCP servers, skills, hotspots, and durable documentation across these agent platforms to detect instruction drift and control gaps during code-rot reviews.
Automated maintainability audits are limited to detecting configuration conflicts, hook gaps, MCP issues, stale references, and documentation drift. They require a summary-first, budget-aware diagnostic pass and cannot fix semantic logic errors or enforce business-specific coding standards automatically.
vinceferroIntegrate multi-agent orchestration into existing repositories with gap analysis.
gannonhDiagnose failing GitHub Actions CI checks and generate remediation plans.
dallayBuild stateful AI agents on Cloudflare Workers with the Agents SDK.
TLOGBenAudit AI project configuration and maintainability across five layers.
