Fix(designer): consumption SKU agent runAfter bug — restore runAfter when re-adding agent loop

[wgrooversoftie]

Commit Type

• feature - New functionality
• fix - Bug fix
• refactor - Code restructuring without behavior change
• perf - Performance improvement
• docs - Documentation update
• test - Test-related changes
• chore - Maintenance/tooling

Risk Level

• Low - Minor changes, limited scope
• Medium - Moderate changes, some user impact
• High - Major changes, significant user/system impact

What & Why

Run after property wasn't being set properly for A2A workflows in consumption. If the agent loop was deleted, re-adding it would not be valid because the runAfter would not get set back up

Impact of Change

• Users: Bug fixed for users
• Developers: Replaces equality-checks for workflowKind === 'agent' with isA2AWorkflow(state) helper; tests added. Reviewers should verify the helper's detection criteria and ensure no other code relies on the old equality check semantics.
• System: No new runtime dependencies. Behavior change could modify how runAfter edges are added in some workflows — recommend testing migration/updating of existing workflows in staging.

Test Plan

• Unit tests added/updated
• E2E tests added/updated
• Manual testing completed
• Tested in:

Contributors

Screenshots/Videos

[Copilot]

Pull request overview

This PR fixes a bug where the runAfter property wasn't being set correctly for Agent-to-Agent (A2A) workflows in Consumption SKU. The fix replaces a simple workflowKind check with a more comprehensive isA2AWorkflow() helper that detects A2A workflows across both Standard and Consumption SKUs.

Key Changes:

• Introduced isA2AWorkflow() helper function to properly detect A2A workflows across both SKUs
• Replaced direct equals(state.workflowKind, 'agent') checks with the new helper in workflow manipulation functions
• Added comprehensive test coverage for the new detection logic

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
libs/designer/src/lib/core/state/workflow/test/helper.spec.ts	Added comprehensive unit tests for isA2AWorkflow() function covering Standard SKU, Consumption SKU, and edge cases
libs/designer/src/lib/core/parsers/pasteScopeInWorkflow.ts	Updated to use isA2AWorkflow() helper instead of direct workflowKind check
libs/designer/src/lib/core/parsers/moveNodeInWorkflow.ts	Updated to use isA2AWorkflow() helper for both old and new graph locations
libs/designer/src/lib/core/parsers/deleteNodeFromWorkflow.ts	Updated to use isA2AWorkflow() helper when deleting nodes
libs/designer/src/lib/core/parsers/addNodeToWorkflow.ts	Updated to use isA2AWorkflow() helper when adding nodes
libs/designer-v2/src/lib/core/state/workflow/test/helper.spec.ts	Added identical comprehensive unit tests for designer-v2 library
libs/designer-v2/src/lib/core/parsers/pasteScopeInWorkflow.ts	Updated to use isA2AWorkflow() helper (designer-v2 version)
libs/designer-v2/src/lib/core/parsers/moveNodeInWorkflow.ts	Updated to use isA2AWorkflow() helper (designer-v2 version)
libs/designer-v2/src/lib/core/parsers/deleteNodeFromWorkflow.ts	Updated to use isA2AWorkflow() helper (designer-v2 version)
libs/designer-v2/src/lib/core/parsers/addNodeToWorkflow.ts	Updated to use isA2AWorkflow() helper (designer-v2 version)

[github-actions]

🤖 AI PR Validation Report

PR Review Results

Thank you for your submission! Here's detailed feedback on your PR title and body compliance:

✅ PR Title

• Current: Fix(designer): consumption SKU agent runAfter bug — restore runAfter when re-adding agent loop
• Issue: Title is generally descriptive and scoped. Small suggestion: follow conventional casing for the commit-type prefix (lowercase fix) and make the scope explicit if you primarily changed both designer and designer-v2 (optional).
• Recommendation: Keep as-is or consider: fix(designer): restore runAfter when re-adding agent loop for consumption SKU (A2A)

---

✅ Commit Type

• Properly selected (fix) in the body. Only one option checked which is correct.
• Note: The PR title also uses Fix(...) — consider lowercase fix(...) to match conventional commit style.

---

⚠️ Risk Level

• The PR body marks this as Low and the PR has the risk:low label.
• Assessment: Based on the code changes, I advise risk:medium.
  • Reason: The change replaces multiple direct equality checks for workflowKind === 'agent' with a new helper isA2AWorkflow(state). This affects runAfter logic across several parser flows (add/delete/move/paste) and could change behavior for existing workflows (edge assignment and runAfter restoration). Although the PR adds comprehensive unit tests, the change has cross-cutting effects on workflow topology and could have user-visible behavior changes that warrant broader validation (integration/E2E / migration testing).
• Recommendation: Update the PR label to risk:medium and include a short note in the body explaining why a medium risk rating is chosen (detection logic change impacting runAfter behavior broadly).

---

✅ What & Why

• Current: Run after property wasn't being set properly for A2A workflows in consumption. If the agent loop was deleted, re-adding it would not be valid because the runAfter would not get set back up
• Issue: The section is concise and sufficient.
• Recommendation: Optionally expand one sentence to mention the new helper isA2AWorkflow and that detection now supports both Standard SKU and Consumption SKU heuristics (metadata and trigger patterns). Example: "Replaced direct equality checks with isA2AWorkflow helper to correctly detect A2A workflows across Standard and Consumption SKUs and restore runAfter edges when re-adding agent loops."

---

✅ Impact of Change

• The Impact section is present and mentions Users/Developers/System impacts.
• Suggestion: Be explicit about the areas to validate in staging: e.g., existing workflows that may have runAfter edges removed/added during schema migrations, and any tools that rely on the old workflowKind equality check.
• Recommendation:
  • Users: Bug fix for runAfter restoration in A2A (agent) workflows. Recommend testing a small set of production flows in staging to verify no regressions.
  • Developers: Note changed semantics: isA2AWorkflow(state) replaces workflowKind === 'agent' checks. Review other usages of workflowKind elsewhere to ensure no assumptions were relying on previous equality checks.
  • System: No new runtime deps, but behavior change to how runAfter edges are created — validate migration/updating of existing workflows in staging and consider adding an E2E test to catch regression in graph edge behavior.

---

✅ Test Plan

• The PR claims unit tests added/updated and manual testing completed.
• Verification: The diff includes substantial unit tests added for isA2AWorkflow across both designer and designer-v2 test files. Good coverage for detection logic and many edge cases.
• Recommendation: Add or run an integration/E2E scenario that exercises adding/deleting/re-adding an agent loop in real workflows to ensure runAfter restoration works end-to-end. If E2E isn't possible in this PR, document the manual test steps and results explicitly in the body.

---

⚠️ Contributors

• Current: Section is empty.
• Assessment: This is optional. If others contributed, please add them. If none, consider adding a short note acknowledging reviewers or saying "No additional contributors." This is not a blocker.

---

⚠️ Screenshots/Videos

• Current: Not provided. This is fine because changes are internal/logic-focused.
• Assessment: No screenshots required. If you test UI behaviors that changed visually, consider adding a short recording or screenshots in future PRs.

---

Summary Table

Section	Status	Recommendation
Title	✅	Optionally normalize to conventional commit casing (fix(...))
Commit Type	✅	Good
Risk Level	⚠️	Change label to risk:medium and state reasoning in PR body
What & Why	✅	Optionally mention the new isA2AWorkflow helper in the section
Impact of Change	✅	Add explicit staging/migration validation steps
Test Plan	✅	Add E2E/integration test or document manual E2E steps/results
Contributors	⚠️	Add contributors or a note if none
Screenshots/Videos	⚠️	Not required for this logic change

---

Final message
This PR body is largely well-formed and compliant with the template. The primary item to address is the risk level: because you changed a cross-cutting detection used by multiple parsers (which can alter runAfter behavior across workflows), I recommend labeling this risk:medium and adding a short justification in the PR body explaining the broader impact and the testing you ran.

Actionable next steps (recommended):

• Update PR label to risk:medium (and explain in the body why the increase). ✔️
• Add a brief sentence in "What & Why" mentioning isA2AWorkflow helper to make the change explicit. ✔️
• Add (or document) an integration/E2E test or manual testing steps that reproduce adding/deleting/re-adding an agent loop end-to-end and verify runAfter restoration. If E2E cannot be added here, add the manual test steps + results to the Test Plan. ✔️
• Consider a quick cross-check to ensure no other places rely on equals(state.workflowKind, 'agent') semantics that were intentionally different. If any exist, either update them or call out why they are intentionally unchanged. ✔️

Please update the PR label and optionally the PR body with the suggested clarifications and testing notes, then mark ready. Thank you for the thorough tests and clear description — this looks like a strong fix; the only change is to increase the risk categorization to medium and document broader validation.

---

Last updated: Fri, 06 Feb 2026 17:54:22 GMT