The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl, openssl, grep, and sed within a bash script to perform diagnostic checks and redact sensitive data. These commands are localized to the user's environment and target official ElevenLabs API endpoints (api.elevenlabs.io).
[DATA_EXPOSURE_&_EXFILTRATION]: While the skill collects configuration and logs, it implements active redaction strategies using sed to mask ELEVENLABS_API_KEY and other sensitive .env values before saving them to the debug bundle. The diagnostic output is intended for manual review by the user before submission to support.
[REMOTE_CODE_EXECUTION]: The skill provides a TypeScript implementation that uses the official @elevenlabs/elevenlabs-js SDK. This is a standard integration for debugging within a Node.js environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes local log files and API responses to generate reports. It uses boundary markers and redaction logic to ensure that sensitive data is not leaked into the diagnostic bundle. The risk of injection via log files is minimal as the output is treated as diagnostic data, not as executable instructions.

elevenlabs-debug-bundle