Claude Health audits six-layer Claude Code hygiene
Claude Health is an open-source Claude Code skill that audits configuration health across six layers: CLAUDE.md, rules, skills, hooks, subagents, and verifiers. It detects project complexity, runs parallel diagnostics, and returns a prioritized report so you can fix the highest-risk issues first.
Hot take: this is the kind of utility that only feels essential after your Claude setup has gotten messy enough to hurt. It’s less a flashy AI demo and more a practical maintenance lens for people who depend on Claude Code every day.
- –Strong fit for teams with layered Claude workflows, especially where config sprawl, prompt bloat, or inconsistent safeguards are becoming real problems.
- –The six-layer framework is a useful mental model because it pushes the audit beyond just `CLAUDE.md` and into the surrounding tooling surface.
- –Running two parallel diagnostic agents is a nice touch for speed and suggests the tool is opinionated about actionable output, not just passive scanning.
- –The scope is deliberately narrow: it targets Claude Code only, so it is best viewed as a specialist maintenance skill rather than a general-purpose config manager.
DISCOVERED
80d ago
2026-03-21
PUBLISHED
80d ago
2026-03-21
RELEVANCE
AUTHOR
Github Awesome