social-media-carousel
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the
belt-sh/clitool using the commandnpx skills add belt-sh/cli. This is a prerequisite for using the image generation features described in the instructions. - [COMMAND_EXECUTION]: The skill documentation provides examples of using the
beltcommand-line interface to authenticate (belt login) and execute remote applications (belt app run). These commands are used to render HTML to images and generate AI-based visuals for slides. - [INDIRECT_PROMPT_INJECTION]: The skill acts as a template for generating content where user-provided text or agent-generated summaries are interpolated into HTML structures before being sent to a remote rendering service. This constitutes an indirect prompt injection surface.
- Ingestion points: Slide titles, body text, and carousel descriptions are formatted into HTML strings within shell commands in
SKILL.md. - Boundary markers: No specific delimiters or safety instructions are used to separate the content from the HTML/JSON structure in the provided examples.
- Capability inventory: The skill uses the
Bash(belt *)tool, allowing it to invoke remote rendering and AI generation services. - Sanitization: No explicit sanitization or validation of the interpolated content is mentioned in the guides.
Audit Metadata