ai-motion-director

Pass

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Risk. The skill's primary function involves ingesting and processing user-provided 'topic or source material' to generate video motion plans and execute production workflows.
  • Ingestion points: Topics, articles, scripts, or briefs provided by the user (SKILL.md).
  • Boundary markers: No delimiters or isolation instructions are present to prevent the agent from executing commands embedded within the user-supplied source material.
  • Capability inventory: The agent is instructed to create project directories, render MP4 files, and archive artifacts using renderer-specific toolchains on the local filesystem (SKILL.md).
  • Sanitization: No explicit sanitization or validation of the input content is defined in the instructions.
  • [DATA_EXFILTRATION]: Local Path Exposure. The skill hardcodes absolute filesystem paths for project storage and archiving, which exposes the system's internal directory structure and the username ('rain') to the AI model.
  • Evidence: Hardcoded paths such as /Users/rain/Desktop/内容创作/01-内容生产/视频工作台/动效项目/ and /Users/rain/Desktop/AI视频成品库/ in the 'Implementation Route' section of SKILL.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 30, 2026, 11:47 AM
Security Audit — agent-trust-hub — ai-motion-director