cinematic-audiovisual-language

Pass

Audited by Gen Agent Trust Hub on Oct 5, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill serves as a comprehensive instructional baseline for cinematic language tasks. A thorough analysis of all instructions and reference files revealed no evidence of prompt injection, credential harvesting, or malicious intent.
  • [EXTERNAL_DOWNLOADS]: The references/source-map.md file includes several external links to educational platforms, university course syllabi, and professional filmmaking blogs (e.g., Adobe, StudioBinder). These URLs are provided strictly for research and source-mapping purposes and do not involve dynamic code execution or automated data exfiltration.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a "拉片 Protocol" for analyzing external video references. While this establishes a surface for processing untrusted data, the skill mitigates risk by providing rigorous structural requirements and diagnostic checklists that constrain the agent's output to technical cinematic analysis.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 5, 2026, 01:13 PM