The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses curl to interact with the PiKVM API and jq to process JSON output. These commands are integral to the skill's functionality for remote hardware control, such as taking screenshots, typing text, and managing power states.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves screenshots and device information from the PiKVM API (e.g., via /api/streamer/snapshot and /api/info). This data represents an ingestion point for external content that could potentially influence the agent's behavior if processed without validation, though the skill itself contains no malicious instructions.
[DATA_EXPOSURE]: Authentication and configuration are managed through the $PIKVM_AUTH and $PIKVM_URL environment variables. This follows security best practices by ensuring sensitive credentials are not hardcoded within the skill instructions.

pikvm