assemblyai-deploy-integration

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected. The skill provides standard deployment procedures using official platform CLIs and encourages the use of secure secret managers (Vercel Environment Variables, Google Secret Manager, and Fly Secrets).
  • [INDIRECT_PROMPT_INJECTION]: The skill includes code for API endpoints that process external data inputs.
  • Ingestion points: The audioUrl and features fields ingested via req.json() in the transcription route within SKILL.md.
  • Boundary markers: None present in the instructional code snippets.
  • Capability inventory: The skill utilizes deployment capabilities via vercel, gcloud, and fly CLI tools, as well as network transcription requests via the assemblyai SDK.
  • Sanitization: The implementation relies on the structured API of the AssemblyAI SDK for data handling. The skill explicitly promotes security best practices by advising the use of temporary streaming tokens and preventing the exposure of the primary API key to client-side environments.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 04:49 AM
Security Audit — agent-trust-hub — assemblyai-deploy-integration