zoom-through
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions for the user to execute a local Node.js script at
~/.claude/skills/video-effects/run.cjsto process video files and apply effects. - [INDIRECT_PROMPT_INJECTION]: The skill accepts user-supplied text via the
--wordsparameter, which is subsequently rendered into the video output. This represents an attack surface for indirect prompt injection if the rendering engine is susceptible to malicious input, though no evidence of exploitation is present in the skill instructions. - Ingestion points: User-supplied strings via the
--wordsargument and the input MP4 file inSKILL.md. - Boundary markers: The instructions do not define specific delimiters or warnings to ignore embedded instructions in the input text.
- Capability inventory: The skill facilitates execution of
run.cjswhich utilizesffmpeg,Node,Puppeteer, andtorchfor video manipulation and rendering as noted inSKILL.md. - Sanitization: There are no documented sanitization or validation routines for the input text provided by the user.
Audit Metadata