The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the openclaw CLI and system utilities such as launchctl to manage service lifecycles, install system-level services, and modify configuration parameters. These commands operate directly on the host environment to perform administrative tasks.
[PROMPT_INJECTION]: The skill demonstrates a surface for indirect prompt injection by design, as it instructs the agent to retrieve and analyze external data from channel logs and messages.
Ingestion points: Untrusted content is ingested into the agent's context through the openclaw logs and openclaw message read commands.
Boundary markers: No explicit delimiters or instructions to disregard instructions within the ingested data were identified in the skill's directives.
Capability inventory: The agent has broad execution capabilities, including service management, configuration editing, and the ability to send messages across integrated communication channels.
Sanitization: The provided files do not describe any mechanisms for sanitizing or validating the contents of retrieved logs or messages prior to processing.

openclaw-ops