screenshot-critique
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides a set of instructions for a visual testing workflow. It does not include any scripts, commands, or network operations.
- [INDIRECT_PROMPT_INJECTION]: The skill processes external visual data (screenshots and crops). It includes specific isolation measures by instructing the agent to spawn sub-agents with
fork_context: falseand neutral prompts, preventing context leakage between the main task and the visual critique.
Audit Metadata