video-hosting-draft

Pass

Audited by Gen Agent Trust Hub on Jun 20, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection surface. 1. Ingestion points: 'videos/' and 'transcripts/' folders (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: Google Sheets API, Google Drive API, subprocess execution for Whisper (SKILL.md). 4. Sanitization: Absent. The risk is mitigated by a mandatory human-in-the-loop (HITL) design where an editor must review and approve AI-generated content in a Google Sheet before any publication.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 20, 2026, 06:44 AM
Security Audit — agent-trust-hub — video-hosting-draft