The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for executing the openclaw CLI tool to perform messaging and agent-to-agent coordination.
[DATA_EXFILTRATION]: The skill facilitates sending messages and files to external messaging services. This is the intended and documented purpose of the skill.
[PROMPT_INJECTION]: The inclusion of the openclaw message read command allows the agent to ingest content from external communication channels, creating a surface for indirect prompt injection.
Ingestion points: openclaw message read commands within SKILL.md.
Boundary markers: The provided examples do not include delimiters or instructions to ignore embedded commands.
Capability inventory: The skill has access to the openclaw CLI for sending messages and initiating agent sessions.
Sanitization: No sanitization steps for external message content are described in the skill.

openclaw-messaging