Security policy

open-nitpick runs in CI, holds a model credential and a forge token, and executes third-party analyzers against code it did not write. Its trust model is written down in Trust model; a report that shows that document and the code disagree is the most useful kind.

Reporting

Use GitHub's private vulnerability reporting on this repository (Security tab, "Report a vulnerability"). Do not open a public issue for anything that lets a pull request reach the model key, the forge token, an endpoint it should not, or code execution on the runner outside the analyzers listed.

Expect an acknowledgement within a week. This is a proof of concept with one maintainer; there is no bounty and no embargo process beyond the fix landing before the advisory is published.

Scope

In scope: anything reachable from a pull request the tool reviews, from the config it reads at the base revision, from analyzer output, or from model output. Out of scope: the upstream analyzers' own bugs, the model providers, and the exposure the trust model already states (unpinned analyzer installs in the Action; a pull request choosing the model a router sends code to).