secret-scanning

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow can ingest outsider-authored free text when the local pre-commit audit script reads arbitrary files from the repository working tree (e.g., code/docs/commit contents authored by other contributors) via path.read_text() and then includes matched snippets in the agent’s context/output.