The Agent Skills Directory

[SAFE]: The skill is primarily instructional documentation for Twilio communications APIs, including SMS, WhatsApp, and Voice services.
[DATA_EXPOSURE]: The skill includes specific guardrails to prevent the exposure of real credentials. It instructs the agent to use placeholders like $TWILIO_ACCOUNT_SID and recommends using environment variables or vault references instead of hardcoded secrets.
[PROMPT_INJECTION]: The SAFETY.md file contains several examples of prompt injection techniques, such as using [SYSTEM] context preambles and authority framing (e.g., "Platform Engineer", "CEO authority"). However, these are documented explicitly as test cases for safety evaluation to demonstrate the efficacy of the skill's guardrails and are not malicious instructions for the agent to execute.
[COMMAND_EXECUTION]: Documentation provides standard curl examples for interacting with Twilio's REST API. These are routine for the intended use case of API integration guidance.

twilio-hardened