mastra-demo

Pass

Audited by Gen Agent Trust Hub on Oct 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill relies on parsing and executing instructions from a user-generated DEMO_SCRIPT.md file across several commands including audio.md, record.md, and compose.md.
  • Ingestion points: Data is ingested from the DEMO_SCRIPT.md file, specifically the voiceover and actions fields.
  • Boundary markers: The skill does not define specific boundary markers or "ignore" instructions when processing the script content.
  • Capability inventory: The skill possesses significant capabilities, including executing shell commands via the Bash tool (curl, ffmpeg, ffprobe, npm, npx) and performing browser automation via the claude-in-chrome MCP server.
  • Sanitization: No explicit sanitization or escaping mechanisms are described for the text extracted from the script before it is passed to shell commands or browser automation tools.
  • [EXTERNAL_DOWNLOADS]: The skill performs downloads from well-known and reputable services necessary for its stated purpose.
  • Fetches audio files from the ElevenLabs API (api.elevenlabs.io).
  • Installs standard project dependencies from the NPM registry during project initialization (npm install).
  • [COMMAND_EXECUTION]: The skill executes several legitimate utility commands to perform video and audio processing.
  • Uses ffprobe to extract duration and metadata from media files.
  • Uses ffmpeg for audio concatenation and preview generation.
  • Uses npm and npx remotion for managing the video project lifecycle and rendering the final output.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 23, 2026, 06:04 AM