1. Habit
  2. Skills
  3. Health skill
tw93 · 16K installs

Health skill
Runs by itself, once a week.

Your agent follows your rules more often.

Install the Health skill and make it a habit: follow https://gethabit.dev/skills/tw93/health.md

  1. 1Copy
  2. 2Paste into Claude Code, Codex or Cursor
  3. 3It runs by itself

16K installs7K starsMIT licenceSource on GitHub ↗

Health

by tw93 · 16K installs

The skill

Checks if your agent setup is healthy and finds what is broken.

Only when you ask.

Better agents

Your agent follows your rules more often. Odd mistakes found every week.

↻ Runs by itselfonce a week

How the habit works.
You don't have to.

Three questionsWhen does it start? How does it check its work? What does it remember?

01

It starts by itself

weekly health audit. Runs in Claude Code routine.

  • ↻ weekly health audit
  • 0 9 * * 1

02

It checks its work

collector + finding format. If a check fails, it tells you.

  • ✓collect-data.sh runs and prints the summary sections
  • ✓every finding names its layer, evidence and a copy-pasteable action
  • ✓no secret appears unless shown as [REDACTED]
  • ✓no project test, build or installer was run

03

It remembers

last health report. The next run reads it first: open findings and global items already reported once.

loops/agent-health-audit-loop/report.md

One run, step by step.
Then it stops.

Stop ruleOne summary audit per run, with at most three recent sessions scanned unless the user asks for a deep audit.

  1. 01

    Set the scope

    Use instructions for a static-only request, otherwise auto, and keep it when escalating.

  2. 02

    Collect

    Run collect-data.sh, or run-health.ps1 on Windows, in summary mode.

  3. 03

    Check runtime

    Inventory hooks and MCP servers and probe only servers that are already connected.

  4. 04

    Check safety

    Run the deny-list, environment override and skill supply chain checks.

  5. 05

    Ask before escalating

    Tell the user before a deep audit, and ask before calling a hook broken.

  6. 06

    Analyse

    Sort findings into critical, structural and incremental, with evidence for each.

  7. 07

    Report

    Write the health report and report global findings only once.

What it may do.
Nothing more.

Safe by defaultIt writes only in its own folder and asks you before anything goes out.

Is it for you?
Use it when it fits.

Quick checkRead both lists. If the left one sounds like you, the Health skill helps.

Use it when

  • ✓ Your agent ignores instructions
  • ✓ Hooks or MCP seem broken
  • ✓ You suspect AI coding code rot

Skip it when

  • – You need to fix an app bug
  • – You want a pull request review

Questions

Health skill, answered.

Does it fix things?

No. A health request is report-only. Repairs need your explicit request.

Which agents does it cover?

Claude, Codex and Pi setups, plus shared project instruction files.

Why ask before a deep audit?

Deep audits read many past sessions and can use a lot of tokens.

How do I install the Health skill?

Copy the prompt at the top into your agent. It installs the skill from tw93/Waza and sets up the habit. Or run npx skills add tw93/Waza --skill health.

Is the Health skill free?

Yes. It is published under MIT. Habit is free too.

Health. Now as a habit.

Install the Health skill and make it a habit: follow https://gethabit.dev/skills/tw93/health.md

  • Free
  • No account
  • Undo in one line

All tw93 skills →More ops skills →

Works in

  • Claude Code
  • Codex
  • Cursor
  • ChatGPT
  • Claude app

Made by tw93, published under MIT. Habit links to the original and never changes it. Is this your skill?

Copying didn't work here. Select the line and copy it by hand.