ai-podcast-creation
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches installation documentation and requirements from the official
inference-shGitHub repository. - [COMMAND_EXECUTION]: Utilizes the
beltCLI tool to run remote AI applications for text-to-speech, music generation, and audio editing. The skill's permissions are scoped to thebeltcommand via theallowed-toolsconfiguration. - [INDIRECT_PROMPT_INJECTION]: The skill features a 'NotebookLM-Style' workflow that ingests external document content into an LLM prompt to generate dialogue scripts.
- Ingestion points: Ingests user-provided documents through the
your-document-contentplaceholder in theSKILL.mdfile. - Boundary markers: Employs clear natural language headers (e.g., "Document:") to separate user data from the system prompt.
- Capability inventory: The skill is capable of executing shell commands (
belt) and performing remote audio processing. - Sanitization: No automated escaping or filtering is applied to the document content, relying on model-level guardrails and structural delimiters.
Audit Metadata