visual-forge
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill includes a vulnerability surface as it processes external, untrusted data from user-supplied URLs and task descriptions.
- Ingestion points: User-provided URLs for visual reference capture and task parameters via the
{{ARGUMENTS}}placeholder in SKILL.md. - Boundary markers: The skill uses descriptive placeholders in handoff templates but lacks strict data/instruction delimiters.
- Capability inventory: The skill coordinates code implementation, file writes, and interaction with image generation tools via the
$forgeand$imagegenextensions. - Sanitization: No explicit content filtering is documented; however, the workflow enforces a mandatory human approval step (Step 3) after reference capture and before any code-bearing implementation, which serves as a robust human-in-the-loop mitigation.
Audit Metadata