Repository navigation
.NET: Fix declarative workflows deep research sample - #7674
Peter Ibekwe (peibekwe) merged 2 commits into
Conversation
Sorry, something went wrong.
There was a problem hiding this comment.
Pull request overview
Fixes the .NET DeepResearch declarative workflow’s planning and completion behavior.
Changes:
- Adds structured team input templating for the planner.
- Clarifies valid negative-result completion criteria.
- Documents an example workflow prompt.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| dotnet/samples/03-workflows/Declarative/DeepResearch/Program.cs | Updates planner inputs and manager guidance. |
| declarative-agents/workflow-samples/DeepResearch.yaml | Adds an example input. |
💡 Add a code-review agent skill for context-aware, tailored reviews. Learn more in the docs.
Sorry, something went wrong.
There was a problem hiding this comment.
MAF Automated Review — Iteration 1
Result: Findings reported
Scope: full PR (1 commit(s)): 1349e2b79a4a
Model: gpt-5.6-sol
Overview
The change correctly replaces the planner's hardcoded team with a required structured input, and the initial planner invocation supplies a deterministic team description from the workflow's fixed agent list. The manager's revised completion guidance also distinguishes verified negative results from execution failures while retaining strict structured output. However, the stall-recovery planner invocation does not supply the newly required input, so recovery can fail when the workflow needs to replan.
Reviewed the supplied pull-request change set across correctness, security/reliability, architecture, and failure behavior.
1 verified finding remained after source verification (1 medium) across 1 file. Details are attached to the affected lines below.
Affected areas: dotnet/samples/03-workflows/Declarative/DeepResearch/Program.cs
Sorry, something went wrong.
There was a problem hiding this comment.
MAF Automated Review — Iteration 2
Result: No findings
Scope: 1 net-new commit(s): 17d4a5e62bfd
Model: gpt-5.6-sol
Overview
This incremental change fixes the stall-recovery planner invocation by passing the same required structured team input used during initial planning, while retaining the narrower replanning prompt. Both planner call sites now satisfy the declared contract, team data is initialized before either call, and the existing bounded restart and routing guards remain intact; no publishable regression was established.
Reviewed the supplied incremental change set across correctness, security/reliability, architecture, and failure behavior.
No publishable findings remained after source verification for this scope.
Sorry, something went wrong.
Motivation & Context
Fix broken DeepResearch declarative workflow sample and prevent valid negative results from being treated as incomplete.
Description & Review Guide
• Major changes: Use structured team inputs, clarify completion criteria, and add an example prompt.
• Impact: Improves planner reuse and manager decision accuracy.
• Reviewer focus: Planner input templating and completion guidance.
Contribution Checklist