ai-podcast-creation
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides templates that process external document content and script lines to generate podcast audio and text. * Ingestion points: User-provided document text and script segments are interpolated into command-line arguments for the belt CLI in the SKILL.md workflows. * Boundary markers: While using JSON for structure, the templates lack explicit delimiters or instructions to prevent the underlying AI models from executing instructions embedded within the text inputs. * Capability inventory: The skill uses the Bash tool to execute the belt CLI for audio processing and LLM tasks. * Sanitization: No sanitization, escaping, or validation logic for the external content is present in the provided templates.
- [COMMAND_EXECUTION]: The skill operates by executing the belt CLI tool through shell commands. This behavior is consistent with its stated purpose as a wrapper for the inference.sh platform tools and is restricted to the allowed binary.
- [EXTERNAL_DOWNLOADS]: The skill references the installation of the belt-sh/cli package and provides links to installation scripts hosted on the vendor's official GitHub repository, which are recognized as legitimate vendor-owned resources.
Audit Metadata