deepgram-core-workflow-a
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides implementation examples for audio transcription using Deepgram, a well-known technology service. No malicious patterns or security risks were identified.
- [DATA_EXFILTRATION]: The skill accesses local files and remote URLs to provide data to Deepgram's official API for processing. This behavior is necessary for the skill's primary function and targets an established service provider.
- [CREDENTIALS_UNSAFE]: The implementation examples correctly demonstrate the use of environment variables (e.g.,
process.env.DEEPGRAM_API_KEY) for API authentication instead of hardcoding sensitive credentials. - [COMMAND_EXECUTION]: The provided code uses standard file system interfaces in Node.js and Python to read audio data from the local environment for processing.
Audit Metadata