The Agent Skills Directory

[DATA_EXPOSURE_AND_EXFILTRATION]: The skill securely handles the API key by retrieving it from environment variables using os.environ.get('TELNYX_API_KEY'), preventing credential exposure in the code.
[INDIRECT_PROMPT_INJECTION]: 1. Ingestion points: The url parameter in the transcription methods within SKILL.md allows processing of external audio files. 2. Boundary markers: None are present to delimit the transcription output. 3. Capability inventory: The skill performs network requests to the Telnyx API but does not execute the returned text. 4. Sanitization: No sanitization is performed on the transcription output. As a standard transcription wrapper, this represents a low-risk attack surface.
[SAFE]: The skill relies on the official telnyx Python package and the telnyx-agent CLI, which are legitimate vendor resources. No obfuscation or persistence mechanisms were detected.

telnyx-stt-python