technical-video-script
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted source material (e.g., blog posts, documentation) to generate scripts and maintain project context.
- Ingestion points: External data is ingested through the 'Interview' section and the 'Transpose the source' step in the 'Workflow' section of SKILL.md.
- Boundary markers: The skill does not define or use specific delimiters or 'ignore' instructions when processing external content provided by the user.
- Capability inventory: The skill writes project-specific metadata to a context file (devrel-context.md) and generates structured script artifacts based on external inputs.
- Sanitization: There are no instructions for sanitizing or validating the contents of the source material before it is used for script generation.
Audit Metadata