microsoft/autogen · error · ValueError

No code blocks to execute.

Error message

No code blocks to execute.

What it means

Raised by DockerCommandLineCodeExecutor._execute_code_dont_check_setup (and the public execute_code_blocks path) when the code_blocks list passed in is empty. It is a simple fail-fast validation: there is nothing to execute, and the API refuses rather than returning an empty result.

Source

Thrown at python/packages/autogen-ext/src/autogen_ext/code_executors/docker/_docker_code_executor.py:334

                    logging.debug(f"Kill command scheduled, future: {future!r}")
                except RuntimeError as e:
                    logging.error(f"Failed to schedule kill command on loop {self._loop!r}: {e}")
                except Exception as e:
                    logging.exception(f"Unexpected error scheduling kill command: {e}")
            else:
                logging.warning(
                    f"Cannot schedule kill command: Executor loop is not available or closed (loop: {self._loop!r})."
                )
            return "Code execution was cancelled.", 1

    async def _execute_code_dont_check_setup(
        self, code_blocks: List[CodeBlock], cancellation_token: CancellationToken
    ) -> CommandLineCodeResult:
        if self._container is None or not self._running:
            raise ValueError("Container is not running. Must first be started with either start or a context manager.")

        if len(code_blocks) == 0:
            raise ValueError("No code blocks to execute.")

        outputs: List[str] = []
        files: List[Path] = []
        last_exit_code = 0
        try:
            for code_block in code_blocks:
                lang = code_block.language.lower()
                code = silence_pip(code_block.code, lang)

                # Check if there is a filename comment
                try:
                    filename = get_file_name_from_content(code, self.work_dir)
                except ValueError:
                    outputs.append("Filename is not in the workspace")
                    last_exit_code = 1
                    break

                if not filename:

View on GitHub (pinned to 027ecf0a37)

Solutions

1. Guard the call: only invoke the executor when code_blocks is non-empty
2. Fix the extraction logic so an empty result is handled before reaching the executor (log the raw message to see why extraction failed)
3. Return a neutral CodeResult yourself if your agent protocol expects one for empty turns

Example fix

# before
result = await executor.execute_code_blocks(extract_blocks(msg), ct)  # may be []

# after
blocks = extract_blocks(msg)
if not blocks:
    return CodeResult(exit_code=0, output="No code blocks to execute.")
result = await executor.execute_code_blocks(blocks, ct)

from autogen_core.code_executor import CodeResult

async def execute_safe(executor, blocks, ct) -> CodeResult:
    if not blocks:
        return CodeResult(exit_code=0, output="No code blocks to execute.")
    return await executor.execute_code_blocks(blocks, ct)

def has_code_blocks(blocks) -> bool:
    return isinstance(blocks, list) and len(blocks) > 0

null

When it happens

Trigger: Calling execute_code_blocks([], cancellation_token); passing a filtered list where every block was dropped (e.g. selecting only 'python' blocks from an LLM message that contained none); splitting an empty string into blocks programmatically.

Common situations: Agents whose last message contained no code fences; upstream code that extracts code blocks via regex and forwards an empty list when extraction fails; loops over messages where some messages legitimately have no code.

Related errors

• Functions failed to load: {exec_result.output.strip()}
• Timeout must be greater than or equal to 1.
• Module name must be a valid Python identifier
• Functions failed to load: {exec_result.output}
• Custom image {image_name} does not exist

AI-assisted analysis of microsoft/autogen@027ecf0a37 (2026-08-15). Data as JSON: /api/errors/5e5d64ab9d7c4c8f. Report an issue: GitHub.