Tear down your AI harness
Static checks in seconds, then an optional AI deep analysis. Every finding is tied to a line, with a fix.
Rather run it on your machine? Harness Doctor runs the same checks inside Claude Code →
What happens to your files
Free Static check
- Runs on our server. No AI provider sees them.
- We keep nothing after the result is shown.
- Analytics get counts and the score band — never file content.
Optional Deep analysis
- Only when you press its button. You see the file list first.
- Sent to Anthropic’s Claude API as our data processor under its Data Processing Addendum. Not used for training.
- Anthropic deletes its copy within 30 days; up to 2 years if its automated safety systems flag it; longer if the law requires.
- Before anything is sent, we scan your files for secrets. If the scan finds one, nothing is sent. Do not paste secrets. If you did, rotate them.
- Not available in the EU, the UK or Switzerland.
Common questions
- Where do my files go?
The static check runs on our server and keeps nothing. Deep analysis sends the files you confirm to Anthropic — see What happens to your files. Don’t paste secrets; if you did, rotate them.
- How is this different from the doctor plugin?
The doctor plugin runs inside Claude Code on your machine, against the files in your project. This page checks only the files you pick, and the checks run on our server.
- Why not just ask Claude to review it?
A model review changes from run to run. Static checks give the same findings for the same files; deep analysis adds the AI review on top, when you want it.
- What does the paid report add?
Every AI finding in full, a report.md, a patch you can apply, and a re-check runbook. Run the free deep analysis, then unlock it from the result.
Run this check locally, any time
Install the free Harness Doctor plugin to run these checks inside Claude Code, against your real project files.
Installation
Plug & play/plugin marketplace add ShipWithAI/shipwithai-plugins && /plugin install shipwithai-harness@shipwithai