Feature

A security checkyou can scope.

You pick what to check, a separate security agent looks at exactly that, and you decide about each fix.

Pro
// the problem

What goes wrong

Security is easy to deferThere is always a feature to finish first. Auth gaps and open endpoints are quiet until the day they are not.
A vague audit finds vague thingsAsking an AI to check everything gives a long list with no order. A named scope gives a short list you can act on.
// what adams does

What Adams does

Asks the scopeInput handling, access control, secrets and dependencies, or the whole change. You choose.
Uses a separate agentThe check is run by an agent that did not write the code.
Reports by severity, never prints secretsEach finding has a file, a line and a level. A secret shows as a kind and a place, never a value.
// worked example

An example

Secrets and dependencies

Illustrative session.

~/project~/your-project · example
/adams:secure
Scope: secrets and dependencies
P0: a live key in an old commit
P2: one package is far behind
Replace the key first, then clean history
// faq

Questions

Is it a penetration test?

No. It is a code level check by an AI agent. It does not replace a professional test for high risk systems.

Does it fix things?

Only after you say yes.

Is it Pro?

Yes. The secret and git guards that run while you work are Lite.

// related

Keep reading