langchain-ai/deepagents · error · TypeError

system_prompt must be str or None, got {type(system_prompt).

Error message

system_prompt must be str or None, got {type(system_prompt).__name__}

What it means

MemoryMiddleware's `system_prompt` must be a str or None. A value of any other type (dict, list, object, etc.) raises TypeError in `__init__`. The prompt is later used for string formatting, so non-string values cannot be accepted.

Source

Thrown at libs/deepagents/deepagents/middleware/memory.py:236

                turns (memory content would otherwise shift after every update
                and invalidate the prefix cache).

                No-ops on non-Anthropic models; Bedrock and Vertex wrappers do
                not qualify.
            system_prompt: System-prompt fragment template. Must contain a
                `{agent_memory}` slot for runtime memory substitution. Pass
                `None` to skip appending entirely (memory is still loaded
                into `state["memory_contents"]`).

        Raises:
            TypeError: If `system_prompt` is not `str` or `None`.
            ValueError: If `system_prompt` is a string missing the
                `{agent_memory}` format slot.
        """
        if system_prompt is not None:
            if not isinstance(system_prompt, str):
                msg = f"system_prompt must be str or None, got {type(system_prompt).__name__}"
                raise TypeError(msg)
            if "{agent_memory}" not in system_prompt:
                msg = "system_prompt must contain the `{agent_memory}` format slot"
                raise ValueError(msg)
        self._backend = backend
        self.sources = sources
        self._add_cache_control = add_cache_control
        self.system_prompt = system_prompt

    def _format_agent_memory(self, contents: dict[str, str], template: str = MEMORY_SYSTEM_PROMPT) -> str:
        """Format memory with locations and contents paired together.

        Substitutes loaded memory into the `{agent_memory}` slot of the
        supplied template.

        Args:
            contents: Dict mapping source paths to content.
            template: Surrounding template; must contain `{agent_memory}`.

View on GitHub (pinned to a1af029e6e)

Solutions

1. Pass a plain string (containing the `{agent_memory}` slot) or None
2. Render your template object to a string before passing it, e.g. template.format(...) or str(...)

Example fix

// before
mw = MemoryMiddleware(system_prompt=ChatPromptTemplate.from_messages([...]))
// after
mw = MemoryMiddleware(system_prompt="... {agent_memory} ...")

def check_system_prompt(p):
    if p is not None and not isinstance(p, str):
        raise TypeError(f"system_prompt must be str or None, got {type(p).__name__}")
    return p

def is_str_or_none(v) -> bool:
    return v is None or isinstance(v, str)

try:
    mw = MemoryMiddleware(system_prompt=prompt)
except TypeError as e:
    prompt = str(prompt)  # or render your template to a string
    mw = MemoryMiddleware(system_prompt=prompt)

When it happens

Trigger: Passing system_prompt=<non-string>, e.g. a dict of prompt parts, a LangChain prompt template object, or bytes, to MemoryMiddleware(system_prompt=...).

Common situations: Reusing an existing prompt-template object (e.g. ChatPromptTemplate or a config dict) instead of rendering it to a string first.

Understand the failure class

Background: Invalid argument type errors: "must be of type string", "expected X, got Y", and ERR_INVALID_ARG_TYPE explained — this error's family across 15 libraries.

Related errors

• -32002
• SHELL_ALLOW_ALL should not be used with ShellAllowListMiddle
• interpreter_ptc must be False, 'safe', 'all', or a list of t
• {name} must be a table
• max_retries must be an int, got {type(max_retries).__name__}

AI-assisted analysis of langchain-ai/deepagents@a1af029e6e (2026-08-29). Data as JSON: /api/errors/a2bc41d3712d8575. Report an issue: GitHub.