audit-codebase

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

• Third-party content exposure detected (medium risk: 0.30). Runtime path scripts/_skill_protocol_runtime.py:_run_stateless runs scripts/seal_audit.py, which reads the agent-supplied --bundle JSON from {input.bundle} and renders audit-handoff.md using scripts/audit_handoff.py:render without any filtering of outsider-authored free-text fields inside that JSON.