fix: push_repo_memory should not run when agent job is skipped or failed by Copilot · Pull Request #24363 · github/gh-aw

Copilot · 2026-04-03T19:55:00Z

push_repo_memory was unconditionally guarded by always(), so it ran even when the agent job failed or was skipped — pointlessly consuming a runner and risking writing stale/incomplete memory.

Changes

pkg/workflow/repo_memory.go: Added needs.agent.result == 'success' to the push_repo_memory job condition, matching the identical guard already present on update_cache_memory
Lock files: Recompiled all 183 workflows

Before / After (with threat detection):

# Before
if: always() && (needs.detection.result == 'success' || needs.detection.result == 'skipped')

# After
if: always() && (needs.detection.result == 'success' || needs.detection.result == 'skipped') && needs.agent.result == 'success'

Without threat detection:

# Before
if: always()

# After
if: always() && needs.agent.result == 'success'

Agent-Logs-Url: https://github.com/github/gh-aw/sessions/111dcd6f-85f7-4979-a14a-c0a71ff548da Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>

Copilot

Pull request overview

Updates the generated GitHub Actions workflow logic so push_repo_memory only runs when the agent job succeeds, preventing wasted runner usage and avoiding writing stale/incomplete repo memory.

Changes:

Tightened push_repo_memory job-level if: to additionally require needs.agent.result == 'success' (and still gate on detection pass/skip when threat detection is enabled).
Regenerated compiled workflow lock files to reflect the updated condition across workflows.

Show a summary per file

File	Description
pkg/workflow/repo_memory.go	Updates `push_repo_memory` job condition to require agent success (and detection pass/skip when enabled).
.github/workflows/workflow-health-manager.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/weekly-blog-post-writer.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/technical-doc-writer.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/security-compliance.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/pr-triage-agent.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/metrics-collector.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/glossary-maintainer.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/firewall-escape.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/discussion-task-miner.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/developer-docs-consolidator.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/delight.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/deep-report.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/daily-testify-uber-super-expert.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/daily-news.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/daily-copilot-token-report.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/daily-community-attribution.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/daily-code-metrics.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/daily-cli-performance.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/copilot-session-insights.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/copilot-pr-prompt-analysis.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/copilot-pr-nlp-analysis.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/copilot-cli-deep-research.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/copilot-agent-analysis.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/code-scanning-fixer.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/audit-workflows.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.
.github/workflows/agent-performance-analyzer.lock.yml	Regenerated lock to include `needs.agent.result == 'success'` in `push_repo_memory` condition.

Copilot's findings

Tip

Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Files reviewed: 27/27 changed files
Comments generated: 2

Copilot · 2026-04-03T19:59:42Z

pkg/workflow/repo_memory.go

+	// failed or was skipped). Using always() so the job still runs even when upstream jobs
+	// are skipped (e.g. detection is skipped when agent produces no outputs).


The function-level doc comment for buildPushRepoMemoryJob earlier in this file still states the job runs even if the agent job fails, but the updated jobCondition now requires needs.agent.result == 'success'. Please update that doc comment to match the new behavior so future readers aren’t misled.

Suggested change

// failed or was skipped). Using always() so the job still runs even when upstream jobs

// are skipped (e.g. detection is skipped when agent produces no outputs).

// failed or was skipped). Using always() so the condition is still evaluated consistently

// when other upstream jobs are skipped (e.g. detection is skipped when agent produces no

// outputs), but the job still requires the agent job result to be "success".

Copilot · 2026-04-03T19:59:42Z

pkg/workflow/repo_memory.go

+	agentSucceeded := BuildEquals(
+		BuildPropertyAccess(fmt.Sprintf("needs.%s.result", constants.AgentJobName)),
+		BuildStringLiteral("success"),
+	)
+	jobNeeds := []string{string(constants.AgentJobName)}
+	var jobCondition string
 	if threatDetectionEnabled {
-		// When threat detection is enabled, run only if detection succeeded (no threats found)
-		// or was skipped (agent produced no outputs or patch — nothing to detect against).
-		jobCondition = RenderCondition(BuildAnd(BuildFunctionCall("always"), buildDetectionPassedCondition()))
+		// When threat detection is enabled, also require detection passed (succeeded or skipped).
+		jobCondition = RenderCondition(BuildAnd(BuildAnd(BuildFunctionCall("always"), buildDetectionPassedCondition()), agentSucceeded))
 		jobNeeds = append(jobNeeds, string(constants.DetectionJobName))
+	} else {
+		jobCondition = RenderCondition(BuildAnd(BuildFunctionCall("always"), agentSucceeded))
 	}


This change updates the generated job-level if: for push_repo_memory, but there doesn’t appear to be a test asserting the new condition (agent success required, plus detection passed/skipped when enabled). Adding a unit/integration test that validates pushRepoMemoryJob.If contains the expected clauses would help prevent regressions (similar to the existing update_cache_memory condition coverage).

fix: push_repo_memory should not run if agent job is skipped or failed

37ef717

Agent-Logs-Url: https://github.com/github/gh-aw/sessions/111dcd6f-85f7-4979-a14a-c0a71ff548da Co-authored-by: pelikhan <4175913+pelikhan@users.noreply.github.com>

Copilot AI assigned Copilot and pelikhan Apr 3, 2026

Copilot created this pull request from a session on behalf of pelikhan April 3, 2026 19:55 View session

pelikhan marked this pull request as ready for review April 3, 2026 19:56

Copilot AI review requested due to automatic review settings April 3, 2026 19:56

pelikhan merged commit 84049b3 into main Apr 3, 2026
51 checks passed

pelikhan deleted the copilot/fix-agent-job-run-conditions branch April 3, 2026 19:56

Copilot started reviewing on behalf of pelikhan April 3, 2026 19:56 View session

Copilot AI reviewed Apr 3, 2026

View reviewed changes

github-actions bot mentioned this pull request Apr 3, 2026

Smoke Test: Claude - 23960187828 #24366

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix: push_repo_memory should not run when agent job is skipped or failed#24363

fix: push_repo_memory should not run when agent job is skipped or failed#24363
pelikhan merged 1 commit intomainfrom
copilot/fix-agent-job-run-conditions

Copilot AI commented Apr 3, 2026

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Apr 3, 2026

Uh oh!

Copilot AI Apr 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

		// failed or was skipped). Using always() so the job still runs even when upstream jobs
		// are skipped (e.g. detection is skipped when agent produces no outputs).

-	// failed or was skipped). Using always() so the job still runs even when upstream jobs
-	// are skipped (e.g. detection is skipped when agent produces no outputs).
+	// failed or was skipped). Using always() so the condition is still evaluated consistently
+	// when other upstream jobs are skipped (e.g. detection is skipped when agent produces no
+	// outputs), but the job still requires the agent job result to be "success".

Conversation

Copilot AI commented Apr 3, 2026

Changes

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Copilot's findings

Uh oh!

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 3, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants