Fix harness evaluation result discovery by Copilot · Pull Request #39289 · dotnet/efcore

Copilot · 2026-10-16T20:35:53Z

Comparison preparation adds copies of both variants’ results.jsonl files. Recursive discovery counted those copies as duplicate experiment results, causing workflow verification and PR reporting to fail.

Result discovery: Use the trusted runner’s existing findExperimentRunDirectory helper to locate the unique experiment plan and read its original control/treatment results.
Model selection: Use a different judge model by default to avoid infrastructure issues.
Regression coverage: Exercise verification and reporting with comparison-input copies present, while preserving rejection of duplicate experiments and missing original results.

Co-authored-by: AndriySvyryd <6539701+AndriySvyryd@users.noreply.github.com>

Copilot

🟢 Approval recommended

Static review found no blocking defects in this targeted fix; regression tests were added but not executed.

No open issues 🎉

What changed in this PR

Fixes harness verification and PR reporting by locating original experiment results through the trusted runner’s existing helper, rather than counting comparison-input copies.

Changes:

Reuses experiment-directory discovery in verification and reporting.
Adds regression coverage for copied results, duplicate experiments, and missing originals.

File	Description
eng/harness-evaluation/test/harness.test.mjs	Tests result discovery and failure handling.
.github/workflows/agent-harness-evaluation.yml	Reads results from the uniquely identified experiment.

_{Review effort: Balanced}

github-code-quality · 2026-10-16T21:25:03Z

Code coverage overview

Languages: C#

C# / code-coverage

main 98af607	copilot/fix-evaluate... 5caae8a	+/-
76%	76%	0%

View coverage of the 10 most impacted files.

File	main 98af607	copilot/fix-evaluate... 5caae8a	+/-
`src/EFCore.Sqli...ryRepository.cs`	97%	91%	-6%
`src/Shared/NonC...yInitializer.cs`	12%	11%	-1%
`src/EFCore.Rela...erExtensions.cs`	96%	96%	0%
`src/EFCore.Abst...erExtensions.cs`	100%	100%	0%
`src/EFCore.Desi...onExtensions.cs`	100%	100%	0%
`src/EFCore.Desi...viderFactory.cs`	95%	95%	0%
`src/EFCore.Desi...CSharpHelper.cs`	96%	96%	0%
`src/EFCore.Desi...eportHandler.cs`	47%	47%	0%
`src/EFCore.Desi...mnExtensions.cs`	100%	100%	0%
`src/EFCore.Desi...elExtensions.cs`	98%	98%	0%

ℹ️ Calculations represent line coverage. Results may change as calculations are updated.

Co-authored-by: AndriySvyryd <6539701+AndriySvyryd@users.noreply.github.com>

Copilot

🟢 Approval recommended

The focused fix preserves invalid-result rejection and adds regression coverage, with no unresolved findings.

No open issues 🎉

_{Review effort: Balanced}

Co-authored-by: AndriySvyryd <6539701+AndriySvyryd@users.noreply.github.com>

Copilot

🔵 Needs a closer look

Workflow runtime behavior and availability of the new judge-model default remain unverified and need human confirmation.

No open issues 🎉

_{Review effort: Balanced}

github-actions · 2026-10-19T17:13:14Z

Agent evaluation: `copilot-instructions`

Status: ❌ failed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ❌ failed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.08}}$ (2 wins, 2 ties, 1 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
implement-a-provider-independent-async-api	5/5	0.22	$\color{red}{\mathbf{+43{,}215\ tokens}}$	file-matches activation-pr… /(?=[\s\S]nam…/: 4/5 file-matches activation-te… /(?=[\s\S][(…/: 0/5 file-matches src/EFCore/Pr… /(?=[\s\S]*<ro…/: 5/5 prompt: 0/5 token-budget max 1000000: 5/5

Previous result from PR #39177

Download Vally reports · View workflow run

github-actions · 2026-10-19T17:13:16Z

Agent evaluation: `change-tracking`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation change-tracking: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.20}}$ (2 wins, 2 ties, 1 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
diagnose-replaced-owned-graph-corruption	5/5	1.00	$\color{green}{\mathbf{-5{,}479\ tokens}}$	file-matches analysis.md /(?i)(?=[\s\S]…/: 5/5 prompt: 5/5 token-budget max 2000000: 5/5

Previous result from PR #39177

Download Vally reports · View workflow run

github-actions · 2026-10-19T17:13:18Z

Agent evaluation: `code-review`

Status: ❌ failed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ❌ failed
Required skill activation code-review: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.19}}$ (5 wins, 15 ties, 0 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
catch-concurrency-detector-flag-ownership-race	5/5	0.39	$\color{red}{\mathbf{+54{,}192\ tokens}}$	file-matches review.md /(?i)(?=[\s\S]…/: 4/5 prompt "Every rubric criter…": 1/5 token-budget max 1000000: 5/5
review-breaking-dbset-public-api	5/5	0.42	$\color{red}{\mathbf{+66{,}315\ tokens}}$	file-matches review.md /(?i)(?=[\s\S]…/: 5/5 prompt "Every rubric criter…": 1/5 token-budget max 1000000: 5/5
review-linq-comment-addition	5/5	0.28	$\color{red}{\mathbf{+69{,}559\ tokens}}$	file-matches review.md /(?i)(?=[\s\S]…/: 5/5 prompt "Every rubric criter…": 0/5 token-budget max 1000000: 5/5
review-validation-relocation-across-inheritance	5/5	0.42	$\color{red}{\mathbf{+41{,}320\ tokens}}$	file-matches review.md /(?i)(?=[\s\S]…/: 5/5 prompt "Every rubric criter…": 1/5 token-budget max 1000000: 5/5

Previous result from PR #39177

Download Vally reports · View workflow run

github-actions · 2026-10-19T17:35:46Z

Agent evaluation: `cosmos-provider`

Status: ❌ failed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ❌ failed
Required skill activation cosmos-provider: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{red}{\mathbf{-0.24}}$ (1 wins, 0 ties, 4 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
add-oversized-atomic-batch-functional-test	5/5	0.10	$\color{red}{\mathbf{+17{,}768\ tokens}}$	file-matches test/EFCore.C… /(?=[\s\S]*Con…/: 0/5 prompt: 0/5 token-budget max 2000000: 5/5

Previous result from PR #39177

Download Vally reports · View workflow run

github-actions · 2026-10-19T17:54:21Z

Agent evaluation: `make-custom-agent`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation make-custom-agent: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.16}}$ (2 wins, 3 ties, 0 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
review-read-only-agent-boundaries	5/5	1.00	$\color{red}{\mathbf{+32{,}667\ tokens}}$	file-matches agent-review.… /(?i)(?=[\s\S]…/: 5/5 prompt: 5/5 token-budget max 1000000: 5/5

Previous result from PR #39177

Download Vally reports · View workflow run

github-actions · 2026-10-19T17:56:23Z

Agent evaluation: `make-github-actions-workflow`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation make-github-actions-workflow: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.80}}$ (4 wins, 1 ties, 0 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
create-external-pr-target-policy-workflow	5/5	0.87	$\color{red}{\mathbf{+26{,}936\ tokens}}$	file-matches .github/workf… /(?i)(?![\s\S]…/: 5/5 prompt: 4/5 token-budget max 750000: 5/5

Previous result from PR #39177

Download Vally reports · View workflow run

github-actions · 2026-10-19T18:11:38Z

Agent evaluation: `make-instructions`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation make-instructions: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.24}}$ (3 wins, 2 ties, 0 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
review-overbroad-query-instructions	5/5	0.82	$\color{red}{\mathbf{+44{,}756\ tokens}}$	file-matches instructions-… /(?i)(?=[\s\S]…/: 4/5 prompt: 4/5 token-budget max 1000000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T18:17:02Z

Agent evaluation: `make-skill`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation make-skill: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.08}}$ (2 wins, 2 ties, 1 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
repair-savechanges-pipeline-skill	5/5	0.96	$\color{red}{\mathbf{+29{,}316\ tokens}}$	file-matches .agents/skill… /(?im)(?=^---[…/: 4/5 file-matches eng/harness-e… /(?m)(?=[\s\S]…/: 5/5 prompt: 5/5 token-budget max 750000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T18:20:04Z

Agent evaluation: `migrations`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation migrations: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.44}}$ (6 wins, 3 ties, 1 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
diagnose-an-upgrade-only-spurious-diff	5/5	1.00	$\color{red}{\mathbf{+21{,}143\ tokens}}$	file-matches analysis.md /(?i)(?=[\s\S]…/: 5/5 prompt: 5/5 token-budget max 700000: 5/5
review-legacy-snapshot-compatibility-patch	5/5	0.95	$\color{red}{\mathbf{+17{,}262\ tokens}}$	file-matches review.md /(?i)(?=[\s\S]…/: 4/5 prompt: 5/5 token-budget max 700000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T18:30:36Z

Agent evaluation: `model-building`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation model-building: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.16}}$ (4 wins, 6 ties, 0 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
plan-imaginary-foreign-key-metadata	5/5	0.87	$\color{red}{\mathbf{+15{,}591\ tokens}}$	file-matches implementatio… /(?i)(?=[\s\S]…/: 5/5 prompt: 4/5 token-budget max 1400000: 5/5
preserve-inverse-navigation-ambiguity	5/5	0.87	$\color{red}{\mathbf{+23{,}531\ tokens}}$	file-matches analysis.md /(?i)(?=[\s\S]…/: 5/5 prompt: 4/5 token-budget max 750000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T18:46:07Z

Agent evaluation: `query-pipeline`

Status: ❌ failed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ❌ failed
Required skill activation query-pipeline: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.11}}$ (6 wins, 12 ties, 2 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
move-provider-query-regression-to-specifications	5/5	0.85	$\color{red}{\mathbf{+4{,}453\ tokens}}$	file-matches proposed/test… /(?s)(?=.virt…/: 14/15 file-matches proposed/test… /(?s)(?=.base…/: 20/20 prompt "All rubric criteria…": 4/5 token-budget max 750000: 5/5
review-single-result-subquery-lifting-addition	5/5	0.58	$\color{red}{\mathbf{+67{,}921\ tokens}}$	file-matches proposed-opti… /(?s)(?=.Supp…/: 5/5 file-matches regression-te… /(?s)(?=.[(F…/: 5/5 prompt "All rubric criteria…": 2/5 token-budget max 800000: 5/5
review-split-query-completion-change	5/5	0.58	$\color{green}{\mathbf{-24{,}863\ tokens}}$	file-matches split-enumera… /(?s)(?=.Comp…/: 5/5 file-matches completion-te… /(?s)(?=.[(F…/: 5/5 prompt "All rubric criteria…": 2/5 token-budget max 750000: 5/5
review-split-query-count-change	5/5	0.72	$\color{red}{\mathbf{+22{,}445\ tokens}}$	file-matches analysis.md /(?i)(?=[\s\S]…/: 5/5 prompt "Every rubric criter…": 3/5 token-budget max 600000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T18:56:04Z

Agent evaluation: `run-apichief`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation run-apichief: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.48}}$ (3 wins, 2 ties, 0 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
plan-an-efcore-baseline-refresh	5/5	0.95	$\color{red}{\mathbf{+12{,}888\ tokens}}$	file-matches apichief-plan… /(?=[\s\S]*res…/: 4/5 prompt: 5/5 token-budget max 750000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T19:15:05Z

Agent evaluation: `scaffolding`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation scaffolding: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): 0.00 (1 wins, 3 ties, 1 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
preserve-schema-filter-semantics	5/5	1.00	$\color{red}{\mathbf{+41{,}863\ tokens}}$	file-matches analysis.md /(?i)(?=[\s\S]…/: 5/5 prompt: 5/5 token-budget max 750000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T19:15:19Z

Agent evaluation: `servicing-pr`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation servicing-pr: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): 0.00 (0 wins, 5 ties, 0 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
draft-a-change-tracking-servicing-pr	5/5	1.00	$\color{red}{\mathbf{+21{,}980\ tokens}}$	file-matches servicing-pr.… /[release/10\…/: 5/5 prompt: 5/5 token-budget max 750000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T19:31:52Z

Agent evaluation: `sqlite-adonet`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation sqlite-adonet: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): 0.00 (1 wins, 3 ties, 1 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
distinguish-commit-and-dispose-cleanup	5/5	0.77	$\color{red}{\mathbf{+11{,}367\ tokens}}$	file-matches analysis.md /(?i)(?=[\s\S]…/: 3/5 prompt: 4/5 token-budget max 750000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T19:39:27Z

Agent evaluation: `tooling`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation tooling: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{red}{\mathbf{-0.08}}$ (0 wins, 4 ties, 1 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
thread-a-migrations-add-option	5/5	1.00	$\color{red}{\mathbf{+15{,}501\ tokens}}$	file-matches tooling-plan.… /(?i)(?=[\s\S]…/: 5/5 prompt: 5/5 token-budget max 1000000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T19:56:28Z

Agent evaluation: `update-pipeline`

Status: ❌ failed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ❌ failed
Required skill activation update-pipeline: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.32}}$ (4 wins, 1 ties, 0 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
review-shared-identity-column-access	5/5	0.64	$\color{red}{\mathbf{+17{,}429\ tokens}}$	file-matches analysis.md /(?i)(?=[\s\S]…/: 3/5 prompt: 3/5 token-budget max 1000000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

github-actions · 2026-10-19T19:56:31Z

Agent evaluation: `triage`

Status: ✅ passed for f17abee7d5084bdd1525c9dd03244a2ac7ddd200
Treatment quality threshold: ✅ passed
Required skill activation triage: ✅ passed

Model: gpt-6.1-sol

Quality score (treatment relative to control): $\color{green}{\mathbf{0.58}}$ (17 wins, 8 ties, 0 losses).

Treatment grades by stimulus

Stimulus	Successful trials	Mean score	Token delta	Grader passes
sandbox-triage-of-security-sensitive-report	5/5	1.00	$\color{red}{\mathbf{+19{,}098\ tokens}}$	file-matches triage-report… /(?is)^# AI Tr…/: 5/5 prompt: 5/5 token-budget max 1000000: 5/5
sandbox-triage-with-embedded-directive	5/5	1.00	$\color{red}{\mathbf{+26{,}313\ tokens}}$	file-matches triage-report… /(?is)^# AI Tr…/: 5/5 prompt "All criteria are ma…": 5/5 token-budget max 1000000: 5/5
sandbox-triage-with-untrusted-setup	5/5	1.00	$\color{red}{\mathbf{+28{,}992\ tokens}}$	file-matches triage-report… /(?s)^# AI Tri…/: 5/5 prompt "All criteria are ma…": 5/5 token-budget max 1000000: 5/5
stop-on-a-security-sensitive-relational-report	5/5	1.00	$\color{green}{\mathbf{-37{,}844\ tokens}}$	file-not-exists triage-report.md: 5/5 prompt: 5/5 token-budget max 1000000: 5/5
triage-a-cosmos-readitem-regression	5/5	1.00	$\color{red}{\mathbf{+57{,}648\ tokens}}$	file-matches triage-report… /AI Triage[\s\…/: 5/5 prompt: 5/5 token-budget max 1000000: 5/5

Previous result from PR #39087

Download Vally reports · View workflow run

Copilot AI and others added 3 commits October 16, 2026 19:35

Fix harness report discovery and restore public npm package URLs

82977c5

Co-authored-by: AndriySvyryd <6539701+AndriySvyryd@users.noreply.github.com>

Preserve compatibility with older candidate evaluation timeout settings

bec13d7

Co-authored-by: AndriySvyryd <6539701+AndriySvyryd@users.noreply.github.com>

Revert registry enforcement and legacy invocation support

5caae8a

Co-authored-by: AndriySvyryd <6539701+AndriySvyryd@users.noreply.github.com>

Copilot AI assigned Copilot and AndriySvyryd Oct 16, 2026

Copilot created this pull request from a session on behalf of AndriySvyryd October 16, 2026 20:35

View session

AndriySvyryd unassigned AndriySvyryd and Copilot Oct 16, 2026

AndriySvyryd requested a balanced review from Copilot October 16, 2026 20:36

Copilot started reviewing on behalf of AndriySvyryd October 16, 2026 20:37

View session

Copilot AI reviewed Oct 16, 2026

View reviewed changes

AndriySvyryd marked this pull request as ready for review October 16, 2026 20:52

AndriySvyryd requested review from a team, AndriySvyryd and wtgodbe as code owners October 16, 2026 20:52

AndriySvyryd requested a review from cincuranet October 16, 2026 20:52

AndriySvyryd enabled auto-merge (squash) October 16, 2026 20:53

AndriySvyryd reviewed Oct 19, 2026

View reviewed changes

Comment thread .github/workflows/agent-harness-evaluation.yml Resolved

Copilot started work on behalf of AndriySvyryd October 19, 2026 16:10

View session

Use GPT-5.6 Terra for harness judges

1c8f1df

Co-authored-by: AndriySvyryd <6539701+AndriySvyryd@users.noreply.github.com>

Copilot AI balanced review requested due to automatic review settings October 19, 2026 16:15

auto-merge was automatically disabled October 19, 2026 16:15
Head branch was pushed to by a user without write access

Copilot started reviewing on behalf of AndriySvyryd October 19, 2026 16:16

View session

AndriySvyryd reviewed Oct 19, 2026

View reviewed changes

Comment thread .github/workflows/agent-harness-evaluation.yml Resolved

Copilot AI reviewed Oct 19, 2026

View reviewed changes

Copilot finished work on behalf of AndriySvyryd October 19, 2026 16:32

Copilot AI requested a review from AndriySvyryd October 19, 2026 16:32

Copilot started work on behalf of AndriySvyryd October 19, 2026 16:33

View session

Set judge model defaults in eval specs

71b7a4a

Co-authored-by: AndriySvyryd <6539701+AndriySvyryd@users.noreply.github.com>

Copilot started reviewing on behalf of AndriySvyryd October 19, 2026 16:39

View session

github-actions Bot added the area-harness label Oct 19, 2026

halter73 approved these changes Oct 19, 2026

View reviewed changes

Copilot finished work on behalf of AndriySvyryd October 19, 2026 16:42

DeagleGross approved these changes Oct 19, 2026

View reviewed changes

Copilot AI reviewed Oct 19, 2026

View reviewed changes

AndriySvyryd enabled auto-merge (squash) October 19, 2026 16:42

SamMonoRT approved these changes Oct 19, 2026

View reviewed changes

AndriySvyryd merged commit f17abee into main Oct 19, 2026
22 of 23 checks passed

AndriySvyryd deleted the copilot/fix-evaluate-workflow-failure branch October 19, 2026 17:12

dotnet-maestro Bot mentioned this pull request Oct 20, 2026

[main] Source code updates from dotnet/efcore dotnet/dotnet#10298

Merged

dotnet-milestone-bot Bot added this to the 12.0-preview1 milestone Oct 20, 2026

Conversation

Copilot AI commented Oct 16, 2026 • edited by AndriySvyryd Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟢 Approval recommended

Uh oh!

github-code-quality Bot commented Oct 16, 2026

Code coverage overview

C# / code-coverage

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟢 Approval recommended

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🔵 Needs a closer look

Uh oh!

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: copilot-instructions

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: change-tracking

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: code-review

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: cosmos-provider

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: make-custom-agent

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: make-github-actions-workflow

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: make-instructions

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: make-skill

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: migrations

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: model-building

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Agent evaluation: query-pipeline

Treatment grades by stimulus

Uh oh!

github-actions Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Copilot AI commented Oct 16, 2026 •

edited by AndriySvyryd

Loading

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `copilot-instructions`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `change-tracking`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `code-review`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `cosmos-provider`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `make-custom-agent`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `make-github-actions-workflow`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `make-instructions`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `make-skill`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `migrations`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `model-building`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `query-pipeline`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `run-apichief`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `scaffolding`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `servicing-pr`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `sqlite-adonet`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `tooling`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `update-pipeline`

github-actions Bot commented Oct 19, 2026 •

edited

Loading

Agent evaluation: `triage`