deepgram-java-speech-to-text

Pass

Audited by Gen Agent Trust Hub on Jul 16, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The documentation provides an installation command (npx skills add deepgram/skills) to add related capabilities from the official vendor repository.\n- [DATA_EXFILTRATION]: Demonstrates reading local files (e.g., audio.wav) using java.nio.file.Files to provide audio data for transcription services.\n- [EXTERNAL_DOWNLOADS]: References and downloads sample audio and documentation from official Deepgram domains, such as static.deepgram.com and developers.deepgram.com.\n- [PROMPT_INJECTION]: The skill establishes a data ingestion surface by processing audio data which could contain instructions intended for the AI agent.\n
  • Ingestion points: Audio data supplied via remote URLs or local file paths in SKILL.md.\n
  • Boundary markers: Absent in the provided code examples.\n
  • Capability inventory: Network communication with Deepgram APIs and local file system access.\n
  • Sanitization: Resulting transcripts are processed without explicit sanitization in the example code snippets.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 16, 2026, 09:08 PM
Security Audit — agent-trust-hub — deepgram-java-speech-to-text