tts
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Node.js script (scripts/speech.mjs) to generate speech files from text. The script uses standard Node.js libraries for argument parsing and file system operations.
- [EXTERNAL_DOWNLOADS]: The script communicates with the official Vercel AI Gateway endpoint (https://ai-gateway.vercel.sh) to perform text-to-speech generation.
- [INDIRECT_PROMPT_INJECTION]: The skill includes functionality to process external text for narration, which presents a surface for indirect prompt injection.
- Ingestion points: Text provided by the user or content from files specified via the --text-file argument in scripts/speech.mjs.
- Boundary markers: The instructions in SKILL.md advise the agent not to send unrelated private context to the provider, but the script itself does not implement technical delimiters for the text payload.
- Capability inventory: The scripts/speech.mjs script has the capability to read local files and transmit their contents to the Vercel AI Gateway.
- Sanitization: The script enforces character count limits (1–5000 characters) and strictly validates the structure and headers of the returned WAV audio data to ensure it is valid media.
Audit Metadata