creativeclaw-elevenlabs-v2
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied narration scripts for speech synthesis, which constitutes an ingestion surface for potentially malicious instructions embedded in the text.
- Ingestion points: The
textparameter in thegenerate_speechtool call, which is populated with script content provided by the user (SKILL.md). - Boundary markers: The skill instructs the agent to preserve the script as written and use the selected model, but it does not employ explicit boundary delimiters or security wrappers for the script text (workflow-basics.md).
- Capability inventory: The skill can invoke tools for speech generation, file management, and media assembly, involving network requests to the vendor's API (SKILL.md, media-assembly.md).
- Sanitization: No explicit sanitization or filtering is applied to the input script, as the Text-to-Speech model is expected to process the content literally.
- [EXTERNAL_DOWNLOADS]: The skill includes a dependency on a Model Context Protocol (MCP) tool hosted at
https://app.creativeclaw.co/mcp/chatgpt. This external connection is to the vendor's official domain and is necessary for the skill's primary functionality (agents/openai.yaml).
Audit Metadata