Using Pydantic state with aliased fields

Finding: In langgraph issue #2555, a graph whose state is a Pydantic model with Field(alias='bar') reportedly fails with a validation error when invoked with the alias key {'bar': 'hello'}; the langgraph and pydantic versions are not stated (UNKNOWN). Source: https://github.com/langchain-ai/langgraph/issues/2555. This is an external report with no KFA reproduction. Reported behavior: the issue author expects the alias to be respected, with the field updateable by its alias, but observes a validation error. They note aliases matter at IO boundaries and are sometimes needed to avoid name collisions, such as a field named schema, which Pydantic already uses. The issue gives no exact error text. Candidate causes (hypotheses, not maintainer-confirmed): A 2026-04-13 comment said _coerce_state builds the model with schema(**input) keyed by field name, while Pydantic expects the alias for aliased fields, and pointed to PR #6634. A comment dated 2026-10-03 reported tracing the failure on current main: the START writer filters updates against field-name channels, so the alias key is dropped before the node runs, and changing _coerce_state alone does not fix it. After mapping input to the field name, constructing an alias-only model also needs the validation alias. These are contributor analyses, not maintainer statements. Status: The issue is open. Contributors on the thread mention PRs #5719, #6634 (described as closed as stale) and #8675 (described as auto-closed for lack of assignment). The latest comment links an unmerged reference branch for string validation aliases at graph input and model construction. The contributor says it fails on the base revision and passes with the patch in a parameterized test, with 7 tests passing in tests/test_pydantic.py. That is self-reported and not independently verified. The contributor states that outputs and checkpoints stay field-name keyed, and that AliasChoices/AliasPath, alias-keyed node or Command updates, and output/stream/snapshot naming are out of scope pending an agreed contract. No merged commit or released fix is evidenced in the supplied sources. Practical implication: Treat alias-keyed input to Pydantic-state graphs as an unresolved, unsupported area. No verified workaround is given in the thread. Anyone affected should track the issue rather than rely on the unmerged branch. This record is a summary of external reports and was not reproduced by KFA.