LLDB MCP Server

Instructions

Implementation Reference

• lldb_mcp_server.py:502-527 (handler)

  The main async handler function for the 'lldb_run_command' tool. It takes RunCommandInput parameters, executes the LLDB command via _run_lldb_command helper, and formats the markdown output.
  async def lldb_run_command(params: RunCommandInput) -> str: """Execute an arbitrary LLDB command and return the output. This is a flexible tool for running any LLDB command. Use this when other specialized tools don't cover your specific need. Common commands: - 'help' - Show help for commands - 'version' - Show LLDB version - 'settings list' - Show all settings - 'type summary list' - List type summaries - 'platform list' - List available platforms Args: params: RunCommandInput containing the command and optional target Returns: str: Command output or error message """ result = _run_lldb_command(params.command, target=params.target, working_dir=params.working_dir) if result["success"]: return f"```\n{result['output']}\n```" else: return f"**Error:** {result['error']}\n\n**Output:**\n```\n{result['output']}\n```"
• lldb_mcp_server.py:177-192 (schema)

  Pydantic BaseModel defining the input schema for the tool, including required 'command' field and optional 'target' and 'working_dir'.
  class RunCommandInput(BaseModel): """Input for running arbitrary LLDB commands.""" model_config = ConfigDict(str_strip_whitespace=True) command: str = Field( ..., description="The LLDB command to execute (e.g., 'help', 'version', 'breakpoint list')", min_length=1, max_length=2000, ) target: str | None = Field( default=None, description="Path to the executable to debug (optional)" ) working_dir: str | None = Field(default=None, description="Working directory for the command")
• lldb_mcp_server.py:492-501 (registration)

  The @mcp.tool decorator that registers the lldb_run_command function as an MCP tool with the specified name and annotations/hints.
  @mcp.tool( name="lldb_run_command", annotations={ "title": "Run LLDB Command", "readOnlyHint": False, "destructiveHint": False, "idempotentHint": False, "openWorldHint": False, }, )
• lldb_mcp_server.py:66-116 (helper)

  Supporting helper function that executes a single LLDB command using subprocess.run in batch mode, capturing output and handling errors/timeouts. Called by the handler.
  def _run_lldb_command( command: str, target: str | None = None, args: list[str] | None = None, working_dir: str | None = None, timeout: int = 30, ) -> dict[str, Any]: """ Execute an LLDB command and return the output. This runs LLDB in batch mode for simple commands. """ cmd = [LLDB_EXECUTABLE] if target: cmd.extend(["--file", target]) # Add batch commands cmd.extend(["--batch", "-o", command]) if args: cmd.append("--") cmd.extend(args) try: result = subprocess.run( cmd, capture_output=True, text=True, timeout=timeout, cwd=working_dir or os.getcwd() ) return { "success": result.returncode == 0, "output": result.stdout, "error": result.stderr if result.returncode != 0 else None, "return_code": result.returncode, } except subprocess.TimeoutExpired: return { "success": False, "output": "", "error": f"Command timed out after {timeout} seconds", "return_code": -1, } except FileNotFoundError: return { "success": False, "output": "", "error": f"LLDB executable not found at '{LLDB_EXECUTABLE}'. Please ensure LLDB is installed and in PATH.", "return_code": -1, } except Exception as e: return {"success": False, "output": "", "error": str(e), "return_code": -1}