king-hu-perspective

Pass

Audited by Gen Agent Trust Hub on May 8, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill acts as an analysis tool for user-provided creative content, creating a surface for indirect prompt injection. \n
  • Ingestion points: Visual storyboards, scene descriptions, and creative concepts provided by the user are processed for analysis. \n
  • Boundary markers: Absent. The skill does not define specific delimiters or instructions for the agent to isolate untrusted user content. \n
  • Capability inventory: The skill is limited to roleplay and aesthetic analysis; it does not invoke tools for network access, file writing, or command execution. \n
  • Sanitization: There is no explicit logic for sanitizing or escaping the content provided by the user. \n- [CREDENTIALS_UNSAFE]: The documentation references local research files using absolute paths (e.g., /Users/baimengke/Documents/...). While these are informational citations and not instructions to access the system, exposing specific local directory structures reveals information about the author's environment.
Audit Metadata
Risk Level
SAFE
Analyzed
May 8, 2026, 03:47 PM
Security Audit — agent-trust-hub — king-hu-perspective