Check a change for security holes

Reads what a change touches, then checks input shapes, injection, server requests to supplied addresses, stray secrets and access control on each route and query. For a developer about to ship.

What it asks for

  • about (Optional, words): What this run is for, in the person's words.
  • known (Optional, words): What is already known, decided or done, so the run starts from it instead of asking again.

The steps your AI walks

  1. Read what the change touches
  2. Check every input is validated by a declared shape
  3. Check that outside input never becomes code, a query or markup
  4. Check every request the server makes to an address someone supplied
  5. Check that no secret value is in the change and each one lives in the sanctioned place
  6. Check each route and data query applies access control itself
  7. Record what was found and what matters most

Use it from your AI

Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this workflow by what it does and runs it.

Say to your AI: “Use Trillion to reads what a change touches, then checks input shapes, injection, server requests to supplied addresses, stray secrets and access control on each route and query. For a developer about to ship”