skill-visual-feedback
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data in the form of screenshots and text descriptions to drive code changes.
- Ingestion points: The workflow relies on screenshots and descriptions provided by users as described in SKILL.md.
- Boundary markers: The instructions do not define clear boundaries or warnings to the agent to ignore potentially malicious instructions embedded within the user-provided images or text.
- Capability inventory: The skill leverages file discovery tools (Glob, Grep) and code modification tools (Edit) as detailed in the implementation phases of SKILL.md.
- Sanitization: There is no evidence of input validation or sanitization for the external visual feedback before it influences the agent's file system operations.
Audit Metadata