ai-podcast-creation
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends installing external packages and tools, including
belt-sh/cliand other skills from theinference-shorganization on GitHub, which are required for the podcast creation workflow. - [COMMAND_EXECUTION]: The skill uses the
beltCLI tool to execute various tasks, such as logging into the service, running AI inference models (Kokoro TTS, AI Music, Claude), and merging media files. These commands are constrained to thebeltutility as specified in the skill's metadata. - [INDIRECT_PROMPT_INJECTION]: The skill outlines workflows that ingest untrusted external data, such as user-provided documents or script segments, which are subsequently processed by Large Language Models (LLMs) to generate podcast content.
- Ingestion points: User-provided document content and script lines referenced in
SKILL.mdshell commands. - Boundary markers: None provided in the examples.
- Capability inventory: The skill invokes the
beltCLI to run remote AI applications for text and audio processing. - Sanitization: No explicit input validation or sanitization steps are documented for the processed content.
Audit Metadata