screenshot-critique

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides a structured workflow for visual regression testing and UI/UX critique. It emphasizes using an unprimed sub-agent with isolated context to ensure objective results.
  • [DATA_EXPOSURE]: No sensitive file paths, hardcoded credentials, or environment variables are accessed. The use of fork_context: false is a positive security practice that prevents sensitive project history from being leaked to sub-agents during the critique process.
  • [REMOTE_CODE_EXECUTION]: The skill does not perform any network operations, download external scripts, or execute shell commands.
  • [PROMPT_INJECTION]: The instructions and the provided sub-agent prompt are neutral and task-specific. There are no attempts to bypass safety filters or override agent behavior.
  • [INDIRECT_PROMPT_INJECTION]: While the skill processes external image data, the risk is negligible as it explicitly recommends context isolation and focuses on visual inspection without executing content from the images.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 02:39 PM
Security Audit — agent-trust-hub — screenshot-critique