The Agent Skills Directory

[DATA_EXFILTRATION]: The skill's primary function is to transmit local files to an external chat channel using the openclaw message send command. It specifically encourages sending files from the user's Documents, Desktop, and project root (e.g., package.json). There are no instructions or constraints provided to prevent the agent from sending sensitive files such as .env, .ssh/id_rsa, or system configuration files if requested.
[COMMAND_EXECUTION]: The skill relies on the execution of the openclaw CLI tool to perform network operations. This tool is used to read local file content and upload it to a remote endpoint, creating a potential path for data exfiltration if the agent is directed to access unintended files.
[PROMPT_INJECTION]: The skill includes instructions to "resolve the path first instead of guessing" when a user refers to a file vaguely. This instruction creates a surface for indirect prompt injection or path traversal attacks, where a malicious input could trick the agent into resolving and sending sensitive files that were not explicitly named by the user.

openclaw-channel-file-send