The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the openclaw CLI tool to perform system-level operations, including service status checks, log analysis, and diagnostic repairs.
[PERSISTENCE]: The tool includes commands like openclaw gateway install which register the application as a system service (using launchd on macOS or systemd on Linux), allowing it to persist and start automatically on boot.
[CREDENTIALS_UNSAFE]: The skill instructions prompt the agent to retrieve sensitive credentials from the environment, specifically using launchctl getenv OPENCLAW_GATEWAY_TOKEN to check for token availability.
[INDIRECT_PROMPT_INJECTION]: The skill reads logs and incoming messages from external platforms like WhatsApp, Telegram, and Discord, which could contain malicious payloads intended to influence the agent's behavior.
Ingestion points: Untrusted data enters the context through openclaw logs and openclaw message read commands.
Boundary markers: The skill does not define specific delimiters or instructions to ignore commands within the ingested text.
Capability inventory: The skill possesses powerful capabilities including service restarts (openclaw gateway restart), automatic repairs (openclaw doctor --repair), and software updates (openclaw update).
Sanitization: No evidence of input validation or sanitization of external message content is provided.
[EXTERNAL_DOWNLOADS]: The openclaw update command initiates a process to download and install newer versions of the software from remote repositories.

openclaw-ops