crewAIInc/crewAI · error · ValueError

seconds must be a number, got NaN.

Error message

seconds must be a number, got NaN.

What it means

WaitTool normalizes requested wait durations through a validation helper. It explicitly rejects NaN (math.isnan) because NaN comparisons are always False, so an uncapped NaN would silently pass the max_seconds cap check; the ValueError names the exact defect.

Source

Thrown at lib/crewai-tools/src/crewai_tools/tools/wait_tool/wait_tool.py:175

    def _resolve_duration(self, seconds: float) -> tuple[float, bool]:
        """Validate and clamp the requested duration to ``max_seconds``.

        ``BaseTool.run`` skips ``args_schema`` validation when called with
        positional arguments, so the bounds are enforced here too rather than
        left to ``time.sleep`` to reject. Infinity is a valid request: it clamps
        to the cap like any other oversized wait.

        Args:
            seconds: The requested wait duration.

        Returns:
            A tuple of the duration to actually wait and whether it was capped.

        Raises:
            ValueError: If ``seconds`` is negative or not a number.
        """
        if math.isnan(seconds):
            raise ValueError("seconds must be a number, got NaN.")
        if seconds < 0:
            raise ValueError(f"seconds must be zero or greater, got {seconds:g}.")
        if seconds > self.max_seconds:
            return self.max_seconds, True
        return seconds, False

    def _format_result(
        self, waited: float, requested: float, reason: str | None
    ) -> str:
        """Describe the completed wait back to the model.

        Args:
            waited: The duration actually waited.
            requested: The duration the model asked for.
            reason: Optional note on what is being waited for.

        Returns:
            A summary of how long was waited and whether the request was capped.

View on GitHub (pinned to 754d7323be)

Solutions

1. Sanitize numeric inputs before calling the tool: use a helper that rejects/replaces NaN (e.g., 0.0) with math.isnan.
2. Fix the upstream computation producing NaN (guard zero divisors, fillna() in pandas).
3. Tighten the tool schema/format so model output cannot be NaN (strict JSON parsing: json.loads rejects NaN unless parse_constant allows it).

Example fix

# before
import math
seconds = float('nan')  # e.g. from model output 'NaN'
tool.run(seconds=seconds)  # ValueError: got NaN

# after
seconds = seconds if (seconds is not None and not math.isnan(seconds)) else 0.0
tool.run(seconds=seconds)

import math

def safe_seconds(v) -> float:
    try:
        f = float(v)
    except (TypeError, ValueError):
        return 0.0
    return 0.0 if math.isnan(f) else f

# seconds = safe_seconds(model_output) before WaitTool

import math

def is_valid_seconds(v) -> bool:
    try:
        f = float(v)
    except (TypeError, ValueError):
        return False
    return not math.isnan(f) and f >= 0

try:
    tool.run(seconds=raw)
except ValueError as e:
    if 'NaN' in str(e):
        tool.run(seconds=0.0)  # sane default, log the bad input
    else:
        raise

When it happens

Trigger: An LLM emitting a tool call like wait(seconds='NaN') that parses to float('nan'), or user code computing seconds as 0/0, float('nan'), or a NaN from math operations (math.inf - math.inf) and passing it to WaitTool.

Common situations: Model-generated JSON with the literal token NaN (accepted by Python's float() but invalid strict JSON); arithmetic upstream producing NaN (division by zero in a stat that feeds the wait); NaN leaking through pandas/numpy values passed unconverted.

Related errors

• seconds must be zero or greater, got {seconds:g}.
• Project name '{name}' would generate folder name '{folder_na
• Project name '{name}' contains no valid characters for a Pyt
• Project name '{name}' would generate class name '{class_name
• Project name '{name}' would generate class name '{class_name

AI-assisted analysis of crewAIInc/crewAI@754d7323be (2026-08-15). Data as JSON: /api/errors/1f685bab6a21d306. Report an issue: GitHub.