ai-podcast

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted user data (podcast scripts and image URLs) and passes them as parameters to external AI tools.
  • Ingestion points: User-provided images for identity training (phota/train) and text content for the podcast script turns used in TTS (inworld/text-to-speech-2) and video generation (pruna/p-video-avatar).
  • Boundary markers: The instructions do not define delimiters or specific "ignore embedded instructions" headers when interpolating user content into tool parameters.
  • Capability inventory: The skill has the capability to invoke external AI services for image, audio, and video processing, as well as file merging via the infsh application runner.
  • Sanitization: No sanitization or validation steps are defined to check the script content for malicious instructions before it is processed by the downstream generation tools.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 05:54 PM
Security Audit — agent-trust-hub — ai-podcast