draft-then-enhance-workflow
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions were found to be benign and focused on legitimate service usage.
- [EXTERNAL_DOWNLOADS]: The skill provides configuration steps for an external Model Context Protocol (MCP) server at
https://mcp.apimage.org/mcp. This domain is an official vendor resource for the apimageorg service. - [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided text for video generation, which is the intended function. 1. Ingestion points: User prompts in the
generate_videotool. 2. Boundary markers: None specified. 3. Capability inventory: Video generation, enhancement, and listing tools. 4. Sanitization: Not specified.
Audit Metadata