Hey Product Hunt - we're launching Trestle today, a local secret scanner for AI-written code. AI coding agents have changed the development loop. They don't just autocomplete a line anymore. They create files, refactor code, move config around, update workflows, and can accidentally leave secrets in places they should never be. Trestle catches API keys, tokens, passwords, private keys, and certificates before they leak. It's code-aware and built to protect the AI coding loop. Trestle is designed to fit into that loop instead of sitting outside it: * check code inside AI coding workflows * surface findings in editors like VS Code and Cursor * scan locally from the CLI * catch issues before commit * run in GitHub Actions for PRs and CI Trestle is local-first by design: no account, no signup, no telemetry, and no need to upload code or secrets anywhere. The Community edition is open source, and Pro adds remediation guidance, rotation steps, and a commercial license. A big focus was signal over noise. Trestle uses code context instead of relying only on pattern matching, so it can catch secrets accurately without burying developers in false positives. I'd love feedback from developers using AI coding tools: - Where would secret scanning be most useful in your workflow? - What would make you add a scanner like this to every repo? - Which integrations should Trestle improve first? Thanks for taking a look.

Trestle

Stop AI-written code from leaking secrets

Stop AI-written code from leaking secrets