Replace custom DictCheckpointStorage with built-in agent-framework 1.2.1 storages by Copilot · Pull Request #423 · microsoft/OpenAIWorkshop

Copilot · 2026-04-28T18:12:36Z

agent-framework 1.2.1 ships InMemoryCheckpointStorage, FileCheckpointStorage, and (via agent-framework-azure-cosmos) CosmosCheckpointStorage, all conforming to a standardized CheckpointStorage protocol. The two hand-rolled DictCheckpointStorage / _DictCheckpointStorage classes in the multi-agent modules predate these and carry ~150 lines of bespoke storage code plus ~80 lines of defensive plumbing to bridge legacy/current method shapes.

Changes

New multi_agent/checkpoint_storage.py — single create_checkpoint_storage(session_id) factory that picks memory (default) / file / cosmos from WORKFLOW_CHECKPOINT_BACKEND (with WORKFLOW_CHECKPOINT_DIR and standard AZURE_COSMOS_* env vars). Caches one storage instance per (backend, session) in-process. Also exposes prune_checkpoints / purge_checkpoints retention helpers built on the public protocol, and a _coerce_checkpoint_storage duck-type validator preserving the existing magentic_checkpoint_storage override hook.
magentic_group.py — drop DictCheckpointStorage, _workflow_name_for_storage, _purge_checkpoint_storage, mark_pending_prompt/consume_pending_prompt, and the multi-fallback _get_latest_checkpoint_id. The pending prompt was already mirrored in state_store; the storage-side copy was dead weight. Resume is now a direct await storage.get_latest(workflow_name=...) against the workflow name captured from workflow.name after build and persisted in state_store.
handoff_multi_domain_agent.py — drop _DictCheckpointStorage. Resume now calls storage.get_latest(workflow_name=f"handoff-{session_id}"). Per-turn prune_checkpoints(retain=5) replaces the in-storage _RETENTION cap.
Tests — replaced the four class-pinned tests with six covering the factory, backend selection (file via monkeypatch), protocol surface, save/load roundtrip, the new resume helper, and the host-injected override hook. Migrated asyncio.get_event_loop() / new_event_loop() calls in this file to asyncio.run() (also silences a DeprecationWarning).
STATE_MANAGEMENT.md — new §6.0 lists the built-in storages first; the legacy hand-rolled Cosmos sample in §6.2 is kept with a "prefer the built-in" note.
Merged latest int-agentic (Fix backend 500s by renaming ChatOptions(model_id=…) to ChatOptions(model=…) for agent-framework 1.2.1 #421 ChatOptions(model_id=…) → ChatOptions(model=…)) into the branch and applied the rename inside the refactored modules.

PR #422 integration-test fix (rolled into this PR)

PR #422's CI/CD pipeline failed on tests/test_backend_api.py — 5 test_backend_chat_* tests returned HTTP 500 in ~0.5s. PR #421 had attempted to fix this by renaming ChatOptions(model_id=…) → ChatOptions(model=…), but the failure persisted because the wrong value was being passed. In agent-framework 1.2.1, the model field in default_options overrides the chat client's deployment name when building the Azure URL. We were passing self.openai_model_name (literal "gpt-5-chat" from bicep) while AZURE_OPENAI_CHAT_DEPLOYMENT holds the actual deployment name, so requests were routed to a non-existent deployment and Azure returned 5xx — bubbling up as /chat 500. Fixed in all 5 sites by passing self.azure_deployment instead:

agentic_ai/agents/agent_framework/single_agent.py
agentic_ai/agents/agent_framework/multi_agent/handoff_multi_domain_agent.py
agentic_ai/agents/agent_framework/multi_agent/magentic_group.py
agentic_ai/agents/agent_framework/multi_agent/reflection_agent.py (2 sites)

Robustness note

CosmosDBStateStore.__setitem__ only upserts on top-level reassignment, so the previous storages' nested-dict mutations were never actually persisted — checkpoints were effectively in-memory regardless of state-store backend. Setting WORKFLOW_CHECKPOINT_BACKEND=cosmos now finally enables real cross-process resume via the SDK's partitioned, durable container, with no behaviour change in the default path.

# Opt into durable workflow resume across process restarts:
export WORKFLOW_CHECKPOINT_BACKEND=cosmos
export AZURE_COSMOS_ENDPOINT=https://my-account.documents.azure.com:443/
export AZURE_COSMOS_DATABASE_NAME=agent-db
export AZURE_COSMOS_CONTAINER_NAME=checkpoints
# (uses DefaultAzureCredential when AZURE_COSMOS_KEY is unset)

Verification

64/64 tests pass in tests/test_agent_framework_1_2_1_regression.py (after the deployment-name fix above).
CodeQL Security Scan: 0 alerts.
Code Review feedback addressed (asyncio.run migration).
Diff against int-agentic: 5 files changed, +440 / −378 lines.

* update fraud_detection_durable to feb 12 agent-framework * enhance fraud detection durable * update readme --------- Co-authored-by: James N. <james.nguyen@microsoft.com>

* update fraud_detection_durable to feb 12 agent-framework * enhance fraud detection durable * update readme * chore: reduce anomaly probability to 1% for controlled demo pace --------- Co-authored-by: James N. <james.nguyen@microsoft.com>

… into james-dev

…ped-contract workflow Scripts 6-8 for the MCP agent demo: - workflow_typed_contracts.py: 4-agent IT security incident response pipeline with Pydantic-enforced contracts at every boundary (SecurityAlert, ThreatAssessment, ImpactAnalysis, IncidentResponse) - mcp_server_hybrid.py: Single MCP endpoint exposing BOTH tool types: * Strict-schema: triage_alert, assess_threat, create_response (Pydantic) * Natural-language: ask_security_advisor, explain_for_customer (prose) * Shared session state across all tools - mcp_client_hybrid.py: 5-step incident flow using both tool types with explicit context passing between steps - README.md: Updated from 6 to 8 capabilities with Hybrid Mode architecture diagram and Quick Start sections 7-8

- Delete workflow_local_remote.py (old Script 3 with extra LLM call) - Add workflow_proxy_agent.py: MCPProxyAgent (BaseAgent, no LLM) calls MCP tools directly via call_tool() zero local LLM overhead - Simplify MCPProxyAgent: extract common _call() method, module-level _last_text() helper, remove verbose static methods - Fix workflow_typed_contracts.py: add async with context managers - Update README: merge Script 9 into Script 3 position, update architecture diagram, Quick Start, How It Works sections

- workflow_proxy_agent.py: remove 'traditional vs proxy' framing MCPProxyAgent is the standard approach for remote agent integration - Delete workflow_typed_contracts.py: not relevant to MCP integration - README: 7 scripts, renumber hybrid 6-7, remove typed-contract sections

- mcp_server_langgraph.py: LangGraph ReAct agent exposed as MCP server on port 8003 with architecture tools (pattern eval, migration estimate, tech stack recommendation) - workflow_group_chat.py: MAF GroupChatBuilder orchestration with LLM planner routing between local BusinessStrategist (MAF) and remote TechnicalArchitect (LangGraph via MCP) - pyproject.toml: add langgraph, langchain-openai, langchain-core deps - README.md: update to 9 scripts, add cross-framework architecture diagram, Quick Start sections, and dependency table entries

…nt, finalize group chat with inline Planner

…language

…tterns (Agent-as-Tool + Agent Adapter), add conceptual architecture

…schemas not

Breaking changes migrated: - agent-framework 1.0.0rc1: ChatAgent->Agent, AgentThread->AgentSession, run_stream->run(stream=True), model->default_options(ChatOptions), WorkflowOutputEvent/AgentRunEvent/RequestInfoEvent->unified WorkflowEvent, MagenticBuilder now uses constructor kwargs, orchestrations moved to agent_framework_orchestrations package - fastmcp 3.0.2: removed mcp version pin (now pulled by fastmcp), no code changes needed (fully compatible) Files updated: - agentic_ai/agents/agent_framework/single_agent.py - agentic_ai/agents/agent_framework/multi_agent/handoff_multi_domain_agent.py - agentic_ai/agents/agent_framework/multi_agent/reflection_agent.py - agentic_ai/agents/agent_framework/multi_agent/magentic_group.py - agentic_ai/applications/pyproject.toml (agent-framework==1.0.0rc1) - agentic_ai/workflow/fraud_detection_durable/pyproject.toml (agent-framework==1.0.0rc1) - mcp/pyproject.toml (fastmcp==3.0.2) - tests/test_agent_framework_rc1_regression.py (51 regression tests, all passing)

The Docker build uses requirements.txt (not pyproject.toml/uv.lock), and it still had agent-framework-core==1.0.0b260130 pinned, which conflicts with agent-framework==1.0.0rc1 requiring agent-framework-core==1.0.0rc1.

Auto-merged after successful CI/CD pipeline run 22315210980

Pushes made by GITHUB_TOKEN don't trigger other workflows (GitHub Actions limitation to prevent infinite loops). This meant the auto-merge into int-agentic never triggered promote-to-main.yml. Now uses secrets.GH_PAT with fallback to GITHUB_TOKEN so: 1. If GH_PAT is configured: merge triggers promote-to-main.yml 2. If not: merge still works, but promotion PR must be created manually

Allows manual triggering when auto-trigger is missed (e.g. after GITHUB_TOKEN-based merges that don't fire downstream workflows).

entic' of https://github.com/microsoft/OpenAIWorkshop into int-agentic st "test" t config --global core.pager cat; git fetch origin main int-agentic; git checkout int-agentic o hello t log --oneline -3 t branch --show-current t branch --show-current 2>&1 | Out-String t fetch origin main int-agentic; git checkout int-agentic; git branch --show-current t -C c:\testing\OpenAIWorkshop -c core.pager=cat fetch origin main int-agentic t -C "c:\testing\OpenAIWorkshop" config --global core.pager cat iines starting with '#' will be ignored, and an empty message aborts

# Conflicts: # agentic_ai/workflow/fraud_detection_durable/pyproject.toml # agentic_ai/workflow/fraud_detection_durable/uv.lock

1. MCP Cosmos DB: Replace AzureCliCredential (not available in containers) with ManagedIdentityCredential (when AZURE_CLIENT_ID is set) or DefaultAzureCredential as fallback. This fixes the 'Azure CLI not found on path' error in production Container Apps. 2. Backend: Remove reflection_workflow_agent from DEFAULT_AVAILABLE_AGENTS and descriptions - the module file doesn't exist, causing it to show as a broken option in the UI agent selector.

Auto-merged after successful CI/CD pipeline run 22321497108

1. promote-to-main.yml: Use GH_PAT instead of GITHUB_TOKEN so the workflow can create PRs (GITHUB_TOKEN lacks permission for this in repos with branch protection). 2. orchestrate.yml: Add Step 8 auto-destroy that tears down the integration-* environment after successful auto-merge to int-agentic. Only dev environments are destroyed; production is retained.

james-tn · 2026-04-28T18:16:58Z

@copilot PR 422 failed integration test, I want you to include the fix for that in this PR

…_group 1) tests/test_agent_framework_1_2_1_regression.py::test_chat_options_model_id incorrectly asserted that ChatOptions(model_id=...) works. ChatOptions is a TypedDict(total=False) so unknown keys are silently accepted at construction and only blow up downstream when forwarded as kwargs to responses.create(). Renamed to test_chat_options_model and added an __annotations__ check so the test fails loudly if the field is renamed again. 2) magentic_group._validate_configuration required AZURE_OPENAI_API_KEY even though _build_chat_client supports managed identity. The other agents (single_agent, handoff_multi_domain_agent, reflection_agent) only require the deployment/endpoint/version and accept either api_key or credential. Aligned magentic_group with that contract so MI-only deployments don't raise spurious RuntimeError on every /chat call. These follow-ups were originally pushed to PR #421 but dropped by its squash merge. PR #423 already touches both files, so re-applying here. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

james-tn · 2026-04-28T18:19:32Z

@copilot Reviewed PR #423 against the actual agent-framework==1.2.1 PyPI wheels (the vendored reference/agent-framework/ is 1.0.0b260107, not 1.2.1, so it's not a reliable reference). The new checkpoint_storage.py module and the InMemoryCheckpointStorage / FileCheckpointStorage / CosmosCheckpointStorage wiring all line up with agent_framework._workflows._checkpoint.CheckpointStorage (protocol with save / load / list_checkpoints / delete / get_latest / list_checkpoint_ids). LGTM on the storage refactor itself.

While reviewing I noticed two unrelated CI-blockers that PR #423 already touches but doesn't fix. I had pushed these as follow-ups to PR #421, but they were dropped by the squash-merge. Re-applied them here as c40a5a9:

tests/test_agent_framework_1_2_1_regression.py::test_chat_options_model_id — incorrectly asserts that ChatOptions(model_id=...) works. ChatOptions is a TypedDict(total=False), so unknown keys are silently accepted at construction and only blow up later when forwarded as kwargs to responses.create() (TypeError: unexpected keyword argument 'model_id'). That's exactly the failure mode that bit the int-agentic CI run (https://github.com/microsoft/OpenAIWorkshop/actions/runs/25067397984). Renamed to test_chat_options_model with an __annotations__ check so a future rename fails loudly instead of silently.
magentic_group._validate_configuration — still required AZURE_OPENAI_API_KEY, even though _build_chat_client already supports managed identity. The other three agents (single_agent, handoff_multi_domain_agent, reflection_agent) only require deployment / endpoint / version and accept either api_key or credential. Aligned magentic_group with that contract so MI-only deployments don't raise spurious RuntimeError on every /chat call.

Both changes are scoped to files PR #423 already modifies, so no extra surface area.

…E) to ChatOptions Agent-Logs-Url: https://github.com/microsoft/OpenAIWorkshop/sessions/a202414e-c0c2-481c-abb6-e986d34c845e Co-authored-by: james-tn <25941658+james-tn@users.noreply.github.com>

Copilot · 2026-04-28T18:32:49Z

@copilot PR 422 failed integration test, I want you to include the fix for that in this PR

Investigated PR #422's integration-test failure (run 25069733984: 5 × test_backend_chat_* returning HTTP 500). Root cause: PR #421 renamed model_id → model in ChatOptions, but in agent-framework 1.2.1 the model field in default_options overrides the chat client's deployment name when building the Azure URL. We were passing self.openai_model_name (literal "gpt-5-chat" from bicep), so requests routed to a non-existent deployment and Azure returned 5xx — bubbled up as /chat 500.

Fixed in e8ee036 by passing self.azure_deployment (the actual AZURE_OPENAI_CHAT_DEPLOYMENT) in all 5 sites: single_agent.py, handoff_multi_domain_agent.py, magentic_group.py, and both calls in reflection_agent.py. 64/64 regression tests still pass.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

ca-be-002.wonderfulfield-a7295646.eastus2.azurecontainerapps.io
- Triggering command: /usr/bin/curl curl -sS -X POST -H Content-Type: application/json -d {"session_id":"diag-1","prompt":"hi"} --max-time 60 REDACTED (dns block)
example.com
- Triggering command: /usr/bin/python python -c import asyncio, sys sys.path.insert(0, '.') from agents.agent_framework.single_agent import Agent async def main(): agent = Agent({}, 'test-session', access_token=None) await agent._setup_single_agent() print('Setup OK') asyncio.run(main()) bin/rustc (dns block)
example.openai.azure.com
- Triggering command: /usr/bin/python python -c import asyncio, os os.environ.update({'AZURE_OPENAI_CHAT_DEPLOYMENT':'gpt-4', 'OPENAI_MODEL_NAME':'gpt-4', 'AZURE_OPENAI_ENDPOINT':'REDACTED', 'AZURE_OPENAI_API_VERSION':'2024-02-01', 'AZURE_OPENAI_API_KEY':'dummy'}) import sys; s (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

Configure Actions setup steps to set up my environment, which run before the firewall is enabled
Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

james-tn · 2026-04-28T18:34:27Z

@copilot Reviewed PR #423 against the actual agent-framework==1.2.1 PyPI wheels (the vendored reference/agent-framework/ is 1.0.0b260107, not 1.2.1, so it's not a reliable reference). The new checkpoint_storage.py module and the InMemoryCheckpointStorage / FileCheckpointStorage / CosmosCheckpointStorage wiring all line up with agent_framework._workflows._checkpoint.CheckpointStorage (protocol with save / load / list_checkpoints / delete / get_latest / list_checkpoint_ids). LGTM on the storage refactor itself.

While reviewing I noticed two unrelated CI-blockers that PR #423 already touches but doesn't fix. I had pushed these as follow-ups to PR #421, but they were dropped by the squash-merge. Re-applied them here as c40a5a9:

tests/test_agent_framework_1_2_1_regression.py::test_chat_options_model_id — incorrectly asserts that ChatOptions(model_id=...) works. ChatOptions is a TypedDict(total=False), so unknown keys are silently accepted at construction and only blow up later when forwarded as kwargs to responses.create() (TypeError: unexpected keyword argument 'model_id'). That's exactly the failure mode that bit the int-agentic CI run (https://github.com/microsoft/OpenAIWorkshop/actions/runs/25067397984). Renamed to test_chat_options_model with an __annotations__ check so a future rename fails loudly instead of silently.

magentic_group._validate_configuration — still required AZURE_OPENAI_API_KEY, even though _build_chat_client already supports managed identity. The other three agents (single_agent, handoff_multi_domain_agent, reflection_agent) only require deployment / endpoint / version and accept either api_key or credential. Aligned magentic_group with that contract so MI-only deployments don't raise spurious RuntimeError on every /chat call.

Both changes are scoped to files PR #423 already modifies, so no extra surface area.

@copilot how about these comments?

Copilot · 2026-04-28T18:36:43Z

@copilot Reviewed PR #423 against the actual agent-framework==1.2.1 PyPI wheels (the vendored reference/agent-framework/ is 1.0.0b260107, not 1.2.1, so it's not a reliable reference). The new ...

Both fixes from c40a5a9 are already on the branch (between 95e7ac5 and my deployment-name fix e8ee036) — no further action needed:

test_chat_options_model (with the __annotations__ assertion) is in place and the silently-passing model_id test is gone.
magentic_group._validate_configuration now matches the other three agents — only deployment/endpoint/version are mandatory, and either api_key or credential satisfies auth.

Re-ran tests/test_agent_framework_1_2_1_regression.py against agent-framework==1.2.1 — 64/64 pass.

james-tn · 2026-04-28T18:37:46Z

@copilot resolve conflicts

…-checkpoint-methods # Conflicts: # agentic_ai/agents/agent_framework/multi_agent/handoff_multi_domain_agent.py # agentic_ai/agents/agent_framework/multi_agent/magentic_group.py # agentic_ai/agents/agent_framework/multi_agent/reflection_agent.py # agentic_ai/agents/agent_framework/single_agent.py Co-authored-by: james-tn <25941658+james-tn@users.noreply.github.com>

Copilot · 2026-04-28T18:41:00Z

@copilot resolve conflicts

Resolved in 4349eba (merge of origin/int-agentic). All four conflicts were on the same line — default_options=ChatOptions(model=...) — where HEAD has self.azure_deployment and int-agentic has self.openai_model_name. Kept HEAD in all four files, since azure_deployment is the fix that resolves the PR #422 /chat 500s (passing the literal "gpt-5-chat" model name routes to a non-existent deployment).

Files resolved:

single_agent.py
multi_agent/handoff_multi_domain_agent.py
multi_agent/magentic_group.py
multi_agent/reflection_agent.py (both sites)

james-tn · 2026-04-28T18:46:49Z

@copilot resolve the merge conflicts in this pull request

…oint-methods # Conflicts: # agentic_ai/agents/agent_framework/multi_agent/handoff_multi_domain_agent.py # agentic_ai/agents/agent_framework/multi_agent/magentic_group.py # agentic_ai/agents/agent_framework/multi_agent/reflection_agent.py # agentic_ai/agents/agent_framework/single_agent.py Co-authored-by: james-tn <25941658+james-tn@users.noreply.github.com>

Copilot · 2026-04-28T18:48:26Z

@copilot resolve the merge conflicts in this pull request

Resolved against origin/main in 754dbb6. Same conflict pattern as before: default_options=ChatOptions(model=self.azure_deployment) (HEAD) vs …=self.openai_model_name (main) on a single line in all four agent files (5 sites total). Kept HEAD — azure_deployment is the fix for the /chat 500s; using openai_model_name (literal "gpt-5-chat") routes to a non-existent Azure deployment.

…es 'preview' Live deployment was returning 500 'API version not supported' from Azure OpenAI because the agents pinned api_version='2025-04-01-preview' on every OpenAIChatClient construction. agent-framework 1.2.1 already reads AZURE_OPENAI_API_VERSION from env and otherwise defaults to the Responses-API magic value 'preview' (v1 endpoint), which is the right behaviour for Azure deployments like gpt-5.2-chat. Changes: - Remove api_version=self.api_version from all four agent_framework agents (single_agent, handoff_multi_domain_agent, magentic_group, reflection_agent). - Drop api_version from each agent's _validate_configuration. - Stop injecting AZURE_OPENAI_API_VERSION into ca-be Terraform; remove the openai_api_version variable and tfvars entries. - Also fix two latent CI issues that were already noted in PR #423 review: * tests/test_agent_framework_1_2_1_regression.py::test_chat_options_model_id asserted the wrong field name; renamed to test_chat_options_model and added an __annotations__ check. * magentic_group._validate_configuration no longer hard-requires the API key; managed identity is now accepted, matching the other three agents. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

James N. and others added 30 commits February 17, 2026 12:02

update fraud_detection_durable to feb 12 agent-framework

e52d463

enhance fraud detection durable

91369f9

update readme

ee3f9f2

Updating Durable Agent Implementation (#404)

391b2d3

* update fraud_detection_durable to feb 12 agent-framework * enhance fraud detection durable * update readme --------- Co-authored-by: James N. <james.nguyen@microsoft.com>

chore: reduce anomaly probability to 1% for controlled demo pace

85c574f

Merge branch 'int-agentic' into james-dev

ea5bc41

add mcp_agent_demo

6c8527c

Merge branch 'james-dev' of https://github.com/microsoft/OpenAIWorkshop…

3132693

… into james-dev

MCP agent demo: rewrite README as MCP-vs-A2A thesis, delete proxy age…

e93bc7c

…nt, finalize group chat with inline Planner

README: Mermaid diagrams, professional tone, remove LinkedIn teasing …

c4af733

…language

README: reframe around multi-framework interop problem, two design pa…

8dfcd4c

…tterns (Agent-as-Tool + Agent Adapter), add conceptual architecture

README: fix Mermaid diagrams - use br tags instead of \n for line breaks

b72c14b

README: fix A2A comparison table - elicitation supported, structured …

5515120

…schemas not

chore: merge james-dev into int-agentic (auto)

0cd61f9

Auto-merged after successful CI/CD pipeline run 22315210980

Add workflow_dispatch trigger to promote-to-main

c081130

Allows manual triggering when auto-trigger is missed (e.g. after GITHUB_TOKEN-based merges that don't fire downstream workflows).

Merge remote-tracking branch 'origin/main' into int-agentic

ae60f8a

# Conflicts: # agentic_ai/workflow/fraud_detection_durable/pyproject.toml # agentic_ai/workflow/fraud_detection_durable/uv.lock

chore: merge james-dev into int-agentic (auto)

cbadc7d

Auto-merged after successful CI/CD pipeline run 22321497108

Trigger rebuild: deploy Cosmos DB auth fix and agent list cleanup

50efa21

Copilot AI assigned Copilot and james-tn Apr 28, 2026

Copilot created this pull request from a session on behalf of james-tn April 28, 2026 18:14 View session

Copilot AI requested a review from james-tn April 28, 2026 18:14

Copilot started work on behalf of james-tn April 28, 2026 18:17 View session

james-tn temporarily deployed to production April 28, 2026 18:19 — with GitHub Actions Inactive

Fix integration test 500: pass Azure deployment (not OPENAI_MODEL_NAM…

e8ee036

…E) to ChatOptions Agent-Logs-Url: https://github.com/microsoft/OpenAIWorkshop/sessions/a202414e-c0c2-481c-abb6-e986d34c845e Co-authored-by: james-tn <25941658+james-tn@users.noreply.github.com>

Copilot finished work on behalf of james-tn April 28, 2026 18:34

Copilot started work on behalf of james-tn April 28, 2026 18:34 View session

Copilot finished work on behalf of james-tn April 28, 2026 18:37

Copilot started work on behalf of james-tn April 28, 2026 18:38 View session

Copilot finished work on behalf of james-tn April 28, 2026 18:41

james-tn marked this pull request as ready for review April 28, 2026 18:45

Copilot started work on behalf of james-tn April 28, 2026 18:46 View session

james-tn merged commit e51ae14 into main Apr 28, 2026
5 checks passed

Copilot stopped work on behalf of james-tn due to an error April 28, 2026 18:49
The session was cancelled by the user.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Replace custom DictCheckpointStorage with built-in agent-framework 1.2.1 storages#423

Replace custom DictCheckpointStorage with built-in agent-framework 1.2.1 storages#423
james-tn merged 51 commits intomainfrom
copilot/review-checkpoint-methods

Copilot AI commented Apr 28, 2026 •

edited

Loading

Uh oh!

james-tn commented Apr 28, 2026

Uh oh!

james-tn commented Apr 28, 2026 •

edited

Loading

Uh oh!

Copilot AI commented Apr 28, 2026 •

edited

Loading

I tried to connect to the following addresses, but was blocked by firewall rules:

Uh oh!

james-tn commented Apr 28, 2026

Uh oh!

Copilot AI commented Apr 28, 2026

Uh oh!

james-tn commented Apr 28, 2026

Uh oh!

Copilot AI commented Apr 28, 2026

Uh oh!

james-tn commented Apr 28, 2026

Uh oh!

Copilot AI commented Apr 28, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Conversation

Copilot AI commented Apr 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Changes

PR #422 integration-test fix (rolled into this PR)

Robustness note

Verification

Uh oh!

james-tn commented Apr 28, 2026

Uh oh!

james-tn commented Apr 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI commented Apr 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

I tried to connect to the following addresses, but was blocked by firewall rules:

Uh oh!

james-tn commented Apr 28, 2026

Uh oh!

Copilot AI commented Apr 28, 2026

Uh oh!

james-tn commented Apr 28, 2026

Uh oh!

Copilot AI commented Apr 28, 2026

Uh oh!

james-tn commented Apr 28, 2026

Uh oh!

Copilot AI commented Apr 28, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Copilot AI commented Apr 28, 2026 •

edited

Loading

james-tn commented Apr 28, 2026 •

edited

Loading

Copilot AI commented Apr 28, 2026 •

edited

Loading