render-podcast-skit

Pass

Audited by Gen Agent Trust Hub on Aug 14, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided configuration files to drive a multi-stage video assembly pipeline, which presents an injection risk. \n
  • Ingestion points: Data enters via fields like scenes[].text, set_description, cta_text, and url_text as seen in scripts/config.example.json. \n
  • Boundary markers: No explicit delimiters or instructions are provided to help the agent ignore commands embedded in script text. \n
  • Capability inventory: The system coordinates file writing, network API calls to ElevenLabs and Fal.ai, and the execution of FFmpeg and Playwright commands. \n
  • Sanitization: No sanitization or validation of the configuration data is described in the agent's instructions. \n- [COMMAND_EXECUTION]: The skill requires running multiple Python scripts and system utilities that operate on data sourced from the potentially untrusted configuration file.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 14, 2026, 05:22 AM
Security Audit — agent-trust-hub — render-podcast-skit