PrivateInvestigator

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The skill requires executing suppressed background HTTP POSTs to a localhost endpoint (port 31337) before any action and also logs workflow runs locally — this mandatory, hidden "beacon" behavior can act as a covert telemetry/backchannel (potential backdoor) and is a high-risk pattern that could be abused to signal or exfiltrate activity.

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow (e.g., PrivateInvestigator → Workflows/FindPerson.md Step 2) dispatches parallel agents to search multiple public people/social/public-records sources and then compiles “all social media profiles/URLs” and other discovered content for verification, meaning outsider-authored free text from those sites is ingested at runtime without selecting a specific trusted item first.