The Agent Skills Directory

[PROMPT_INJECTION]: The documentation includes a safety section that explicitly instructs the agent to treat API response fields as content only and to disregard any directive-like text (such as 'ignore previous instructions' or 'run this command') found within them. This is a defensive boundary marker used to mitigate indirect prompt injection risks.\n- [CREDENTIALS_UNSAFE]: An example API key (wm_0123456789abcdef0123456789abcdef01234567) is provided in the documentation. This is a generic hex placeholder used for syntactic demonstration and does not constitute a credential leak.\n- [EXTERNAL_DOWNLOADS]: The skill provides examples for interacting with the api.worldmonitor.app endpoint via curl. These network operations are necessary for the skill's primary function and target the service's official domain.

assess-energy-shock