The Agent Skills Directory

[SAFE]: The skill enables an agent to control its iteration loop using standard local inter-process communication (IPC) mechanisms. Analysis of the scripts confirms that communication is restricted to a local Unix domain socket specified by the environment variable TARIBOY_TOOLS_SOCKET. No evidence of prompt injection, data exfiltration, or unauthorized remote code execution was found. The scripts/loop.sh wrapper is a safe mechanism for calling the included Python logic, and the script's access to local manifest files for versioning is consistent with its stated purpose. All commands and logic are directed toward internal agent management without external dependencies or risky behaviors.

loop