deepgram-go-speech-to-text

Pass

Audited by Gen Agent Trust Hub on Jun 29, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the official Deepgram Go SDK library hosted on GitHub at github.com/deepgram/deepgram-go-sdk/v3.
  • [EXTERNAL_DOWNLOADS]: Provides a command to install additional vendor-owned skills from deepgram/skills via the npx skills add utility.
  • [PROMPT_INJECTION]: An indirect prompt injection surface exists because the skill facilitates transcribing audio from external sources that could contain malicious instructions.
  • Ingestion points: Audio content is processed through FromURL, FromFile, and FromStream methods in SKILL.md.
  • Boundary markers: No explicit delimiters or warnings for handling the resulting transcribed text are provided in the code samples.
  • Capability inventory: The demonstrated Go code includes capabilities for network communication and file system access.
  • Sanitization: The transcription output is not sanitized or validated before being printed or processed in the examples.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 29, 2026, 02:41 PM
Security Audit — agent-trust-hub — deepgram-go-speech-to-text