video-script-collaborial

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • PROMPT_INJECTION (SAFE): No instructions to override safety protocols or system behavior were found.\n- DATA_EXFILTRATION (SAFE): No commands related to network requests or sensitive file access are present.\n- NO_CODE (SAFE): The skill is entirely composed of markdown instructions and contains no programming code.\n- PROMPT_INJECTION (SAFE): While the skill ingests user script data, it has no tool-use capabilities, eliminating the risk of indirect prompt injection attacks.\n
  • Ingestion points: Processes user-provided text for video scripts in file SKILL.md.\n
  • Boundary markers: Not specified.\n
  • Capability inventory: None; the skill provides text instructions only.\n
  • Sanitization: Not applicable.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:11 PM