Execute a shell command in a subprocess and return the output. Use this tool when you need to run system commands, manage files, interact with command-line tools, or perform system operations.

This tool executes shell commands using subprocess with configurable timeout. Important notes: - Commands run through the system shell (/bin/sh or cmd.exe) - Captures both stdout and stderr - Default timeout is 30 seconds - Returns stdout, stderr, and return code - Supports piping, redirection, and other shell features Security considerations: - Commands execute with the same permissions as the agent process - Be extremely careful with destructive operations (rm, del, etc.) - Avoid executing untrusted commands - Commands will affect the actual system (not sandboxed) - Use absolute paths when possible for clarity

The shell command to execute. Can include pipes, redirections, and other shell syntax. Maximum execution time in seconds. Default is 300 seconds. Use higher values for long-running commands. The working directory in which to execute the command. If not specified, defaults to the current working directory of the agent. Response containing: - success: Whether the command executed successfully (return code 0) - message: Status message - Dictionary with stdout, stderr, and return_code # List files ls -la # Check disk space df -h # Using pipes cat /etc/hosts | grep localhost # With custom timeout sleep 5 && echo "Done waiting" 10 # Create directory structure mkdir -p project/{src,tests,docs} && ls -R project # Git operations git status # Execute in a specific directory ls -la /app/AIG-PromptSecurity