carousel-creator
Pass
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the shell tool to run local scripts (export-carousel-pdf.zsh, export-carousel-mp4.zsh, and export-carousel-hyperframes.zsh) for media processing. It also performs SVG validation using a Python one-liner via the shell. These operations are intended for converting generated SVG slides into final social media assets.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface (Category 8) as it processes untrusted user inputs to generate content while possessing significant capabilities. 1. Ingestion points: User-provided 'topic', 'target audience', and 'handle' are ingested in SKILL.md and README.md. 2. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the generation prompts. 3. Capability inventory: The skill can write files (SVGs and markdown) using the Write tool and execute shell scripts via the shell tool across all zsh scripts provided. 4. Sanitization: Structural SVG validation is performed via Python, but no semantic sanitization of user-provided strings is mentioned.
Audit Metadata