ai-marketing-videos

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The workflow for generating explainer videos chains the output of a script-generation model into a text-to-speech model. This creates a potential surface where instructions embedded in the generated script could influence the behavior of the downstream voiceover tool.
  • Ingestion points: The generated content from openrouter/claude-sonnet-45 (stored in script.json) is passed directly to the infsh/kokoro-tts app.
  • Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the interpolated script data.
  • Capability inventory: The skill utilizes the Bash tool to execute belt CLI commands for AI model inference and media merging.
  • Sanitization: No sanitization or validation logic is applied to the script content before it is processed.
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to automate multi-step video production processes, including the use of shell loops and variables to manage multiple assets and API calls.
  • [EXTERNAL_DOWNLOADS]: The skill references external installation resources hosted on the GitHub repository for the inference-sh organization.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 02:20 PM
Security Audit — agent-trust-hub — ai-marketing-videos