code-quality-audit

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime LLM context is populated from the target repository’s files (e.g., it creates an audit worktree and then Codex runs with -C "$worktree" to read GOAL.md/PROGRESS.md and inspect repo contents), so any outsider-authored repo text (issues/PRs/wiki/discussion content or other third-party code/docs) can be ingested as readable prose during the audit.