repurpose-quotes
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes content atoms that may originate from untrusted external URLs to generate social media content. Ingestion points: The 'atoms' input derived from the '[url-or-atoms]' argument. Boundary markers: Absent; the skill does not use specific delimiters or warnings to isolate external content from its instructions. Capability inventory: Writing text files to the 'quotes/' directory and invoking the '/banana' image generation tool. Sanitization: Absent; the skill processes input content without explicit filtering or validation. The impact is low as actions are restricted to content generation.
- [EXTERNAL_DOWNLOADS]: The skill references Pixabay, a well-known stock photo service, providing search queries for manual background selection as a functional fallback.
- [COMMAND_EXECUTION]: The skill utilizes an internal tool named '/banana' for image generation, which is a platform-provided capability and not a vector for arbitrary shell command execution.
Audit Metadata