The Agent Skills Directory

[DATA_EXPOSURE_AND_EXFILTRATION]: The skill manages credentials using environment variables (HA_URL, HA_TOKEN) rather than hardcoding them, which is a standard security practice for managing access to services like Home Assistant.
[COMMAND_EXECUTION]: The skill utilizes the curl binary to interact with local or remote Home Assistant API endpoints. The execution is scoped to specific automation and scene management services (/api/services/scene/turn_on, /api/services/automation/trigger) necessary for its functional purpose.
[INDIRECT_PROMPT_INJECTION]: The skill defines an interface for processing user-supplied natural language requests. It mitigates injection risks by requiring the agent to inventory actual system capabilities first, restate success conditions and boundaries before execution, and maintain the original system's permission model.
[REMOTE_CODE_EXECUTION]: No remote script execution or unverifiable package installations were detected. The skill relies on pre-installed system binaries and environment configuration.

awaken