Enji GuardThe Cursor AI code security checklist
Cursor edits across many files quickly, so one wrong assumption spreads wide. Run through this checklist after a broad Cursor change, and automate it so every edit gets the same check.
1 repo · setup in about 5 minutes · no card, no commitment
01
After a broad multi-file edit
Agentic editing makes mistakes bigger. When Cursor rewrites several files for one prompt, check:
- No auth or access-control check was dropped during the refactor
- Input handling is consistent across every file it touched
- No secret or key was written into a reachable route
- Behavior is unchanged where it should be (no silent regressions)
02
Dependencies Cursor added
Agents add packages to satisfy a change. Check:
- Every new package actually exists upstream
- No typosquatted or slopsquatted names slipped in
- No known-vulnerable or abandoned libraries were introduced
03
Tests and proof
Cursor may rewrite tests to pass. Check:
- Tests still verify behavior, not just turn green
- New code paths and error handling are actually covered
04
Runtime and rules
Some risk only shows at runtime, and Cursor's own rules can drift. Check:
- A route Cursor changed still requires auth in the running app
- Your .cursorrules / Cursor rules still match the real repo
05
Automate it
A broad change needs a broad check, every time. Enji Guard audits the repo and the running app after Cursor edits, ranks what matters, and returns reviewable GitHub issues or pull requests.
Quick questions
Does Guard slow Cursor down?
No. Guard runs out of band on a schedule through a revocable GitHub App. You keep using Cursor; Guard audits the result and only opens issues or pull requests a person reviews.
Can it run the checklist automatically after each change?
Yes. Connect the repo and Guard re-audits as it changes, so broad Cursor edits get the same check every time. The first audit is free.
Verify what Cursor wrote, automatically.
Connect a repo and Guard audits every broad change. No card.