The Agent Skills Directory

[SAFE]: The skill serves as a reference and template for developers deploying Deepgram integrations. Analysis of the provided instructions and code snippets revealed no malicious patterns or security risks.- [CREDENTIALS_UNSAFE]: No hardcoded credentials were found. The provided code templates correctly demonstrate fetching secrets from a dedicated manager (e.g., getSecret('DEEPGRAM_API_KEY')).- [DATA_EXPOSURE]: The skill includes a sanitize function to strip potentially sensitive information from audio URLs before they are logged, which is a recommended security practice.- [COMMAND_EXECUTION]: While the skill's metadata allows Bash(curl:*), the skill body does not contain any commands that execute external scripts or perform unauthorized network operations.

deepgram-prod-checklist