ketchup-video
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data sources including screenshots, recordings, diffs, and project data to generate video slides. This represents an ingestion surface for indirect prompt injection.
- Ingestion points: The skill explicitly instructs the agent to use 'screenshots, recordings, diffs, and data the work produced' located in the environment (SKILL.md).
- Boundary markers: There are no explicit instructions or delimiters provided to prevent the agent from following instructions that might be embedded within the ingested data or diffs.
- Capability inventory: The skill involves file reading (sibling directory SKILL.md files), file writing ('outside the repo'), and video rendering which typically involves external tool invocation.
- Sanitization: No sanitization or validation steps are defined for the content extracted from screenshots or data files.
- [COMMAND_EXECUTION]: The instruction to 'Render one silent landscape video, such as an mp4' and 'match the file's frames to the slide' implies the use of media processing CLI tools (e.g., FFmpeg) or specialized scripts, which expands the agent's execution footprint.
Audit Metadata