Python: fix AG-UI service-session snapshot replay

[pratik wayase (PratikWayase)]

Motivation & Context

When AgentFrameworkAgent combines use_service_session=True with a snapshot_store, the provider service session must remain the sole authority for prior provider history. Replaying the AG-UI snapshot alongside an existing Foundry conversation or response ID can send historical assistant messages or tool results twice, causing Responses API validation failures.

AG-UI threadId and runId values are client-owned protocol correlation IDs; they are not inherently provider conversation identifiers. This change keeps complete backend-owned snapshots for AG-UI hydration, sends only validated incremental input to the provider, and persists the provider continuation privately.

Description & Review Guide

• What are the major changes?

  • Split service-session requests into a validated provider suffix and a complete persisted snapshot using the existing canonical snapshot reconstruction and trust filtering.
  • Support full, incremental, ID-less, stale, truncated, tool-result, approval, and generic input_required continuation shapes.
  • Persist provider conversation/response IDs as private snapshot continuation state instead of conflating them with the AG-UI Thread ID.
  • In provider-owned mode, create a provider conversation through create_conversation() when supported; otherwise persist the provider-issued response ID for subsequent previous_response_id continuation.
  • Add explicit service_session_id_from_thread_id=True compatibility behavior for applications that intentionally use a provider continuation ID as their AG-UI Thread ID.
  • Add deterministic AG-UI → OpenAIChatClient → local Responses host coverage, explicit Agent Framework Message/Content history coverage, and live Foundry Hosted Agent tests for pre-created and backend-created conversations.

• What is the impact of these changes?

  • Foundry Hosted Agents and other service-session providers receive only the current continuation input.
  • Native AG-UI clients can retain their own UUID or application-defined Thread IDs while the backend privately manages provider conversations or response chains.
  • Applications intentionally sharing a provider ID as threadId can preserve that behavior through the explicit compatibility option.
  • AG-UI clients retain complete snapshot hydration and persistence, while non-service-session behavior remains snapshot-authoritative.

• What do you want reviewers to focus on?

  • Provider suffix validation for ID-less, stale, truncated, tool-result, and resume continuations.
  • Separation between client-owned AG-UI IDs, provider continuation IDs, provider-bound messages, and persisted UI snapshots.
  • Conversation-mode and response-mode behavior, including private restoration of provider continuation state.
  • The deterministic local Responses boundary tests and live two-turn Foundry Hosted Agent coverage.

Related Issue

Fixes #7710

Contribution Checklist

• The code builds clean without any errors or warnings
• All unit tests pass, and I have added new tests where possible
• The PR follows the Contribution Guidelines
• This PR is linked to an issue and there is no other open PR for this issue.
• This is not a breaking change.

[Copilot]

Pull request overview

Separates service-managed provider history from AG-UI snapshot history.

Changes:

• Prevents stored snapshot replay into service-session provider calls.
• Reconstructs full snapshots for UI hydration.
• Adds split-authority regression coverage.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
_agent_run.py	Separates provider input and persisted snapshots.
test_snapshots.py	Tests incremental provider input and full snapshot history.

---

💡 Add a code-review agent skill for context-aware, tailored reviews. Learn more in the docs.

[chetantoshniwal]

/review

[github-actions]

MAF Automated Review — Iteration 1

Result: Findings reported
Scope: full PR (3 commit(s)): 3f0ac0ba31ed, 7d51557ce79c, 35527333f272
Model: gpt-5.6-sol

Overview

The review found 3 verified inline finding(s).

Reviewed the supplied pull-request change set across correctness, security/reliability, architecture, and failure behavior.
3 verified findings remained after source verification (1 high, 2 medium) across 1 file. Details are attached to the affected lines below.

Affected areas: python/packages/ag-ui/agent_framework_ag_ui/_agent_run.py

[github-actions]

Python Test Coverage Report •

File	Stmts	Miss	Cover	Missing
packages/ag-ui/agent_framework_ag_ui
_agent.py	48	2	95%	82, 145
_agent_run.py	1474	147	90%	161, 221, 240, 287, 305, 321–322, 444, 497–498, 505, 614, 618, 620, 637, 664–665, 719, 748, 762, 766, 770, 773, 778, 783, 795, 802–808, 819, 833, 840, 850, 865, 868, 900, 935, 959, 972, 1021, 1039, 1041, 1067, 1069, 1071, 1074–1078, 1102, 1133, 1169, 1224–1225, 1267, 1281, 1326–1332, 1341, 1349, 1377, 1399, 1403, 1406, 1408, 1468–1469, 1475–1481, 1584, 1588–1589, 1592, 1597–1598, 1601, 1607–1608, 1612–1614, 1632–1634, 1659–1660, 1683–1684, 1697, 1739–1741, 1752–1754, 1756, 1770–1772, 1900, 1908, 1938, 1942, 1987, 2147–2148, 2230, 2339, 2344, 2416, 2418, 2425, 2441, 2446, 2503, 2508, 2512, 2556, 2602, 2677, 2822, 2860, 2868, 2870, 2873, 2879, 2934, 2937, 2947–2948, 2955, 3002
_run_common.py	662	30	95%	100, 175–176, 178, 180, 183, 189, 191, 205, 212, 236–237, 250, 269, 292–294, 305, 349, 361, 369, 372, 560, 581, 875–876, 1200–1201, 1267, 1270
_snapshots.py	83	4	95%	139, 222, 233–234
TOTAL	48227	4484	90%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
9760	36 💤	0 ❌	0 🔥	2m 38s ⏱️