crypto-secrets

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow runs scripts/scan.py, which reads and ingests outsider-authored repository files from a user-supplied <target-dir-or-file> path (via read_text() and then detect_crypto.scan_file() / detect_secrets.scan_file()), so arbitrary text from outside the operating user’s chosen workspace can be fed into the agent’s LLM context through the emitted findings/snippets.