The Agent Skills Directory

[SAFE]: The skill documents how to build communication features while explicitly warning against security anti-patterns like hardcoding credentials or ignoring delivery status callbacks.
[DATA_EXPOSURE_AND_EXFILTRATION]: The provided code samples adhere to security best practices by accessing sensitive credentials via environment variables (os.environ) rather than hardcoding them in the source. No patterns of data exfiltration were identified.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns, unauthorized downloads, or dynamic script executions were detected. The skill uses standard, well-known libraries (twilio, flask).
[PROMPT_INJECTION]: No prompt injection or behavior override instructions were found in the skill text or code comments.
[COMMAND_EXECUTION]: The skill does not contain any shell command execution or subprocess spawning.
[INDIRECT_PROMPT_INJECTION]: The skill defines clear boundaries for processing external data by providing a decorator for Twilio request validation (validate_twilio_request), ensuring that webhook endpoints only process authentic requests from the service provider.

twilio-communications