Delete msbuild-server skill

[ViktorHofer]

msbuild-server is now on by default since .NET 11 Preview 7. We don't want to recommend turning it on, automatically in other scenarios as critical fixes could be missing.

Summary

Related issue

Validation

Checklist

• I searched existing issues and pull requests to avoid duplicates.
• I kept this pull request focused and avoided unrelated refactors.
• I added or updated tests, evals, or documentation when changing skill or agent behavior.
• I updated CODEOWNERS when adding or moving owned content.
• I updated all marketplace manifests when plugin metadata changed.
• I updated eng/known-domains.txt for any new external domains referenced by skill content.

[github-actions]

Skill Coverage Report

	Plugin	Skill	Covered	Coverage
❔	dotnet-msbuild	msbuild-server	—	error

[Copilot]

Copilot review overview

🟡 Changes recommended

Another remaining dotnet-msbuild skill still recommends “MSBuild Server” usage, which conflicts with the PR’s stated intent to stop recommending enabling it.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Review tier: Lite
Findings: 1

New issues introduced by this change (1)

Severity	Finding
	eng/​eval-quality/​underpowered-allowlist.txt — PR description says we no longer want to recommend turning MSBuild Server on, but the…

What changed in this PR

This PR removes the msbuild-server skill and its associated evaluation fixture/test content, reflecting the new guidance that MSBuild Server should not be recommended as an opt-in tweak now that behavior is changing/standardizing in newer .NET SDKs.

Changes:

• Deleted the plugins/dotnet-msbuild/skills/msbuild-server skill content.
• Removed the tests/dotnet-msbuild/msbuild-server eval fixture (eval, sample project, and developer report).
• Dropped the removed eval from the underpowered allowlist.

File	Description
tests/​dotnet-msbuild/​msbuild-server/​SlowCliBuilds.csproj	Removes the sample project used by the msbuild-server eval fixture.
tests/​dotnet-msbuild/​msbuild-server/​eval.yaml	Removes the msbuild-server skill evaluation spec.
tests/​dotnet-msbuild/​msbuild-server/​developer-report.md	Removes the scenario prompt context used by the eval fixture.
tests/​dotnet-msbuild/​msbuild-server/​Calculator.cs	Removes the sample source file used by the fixture.
plugins/​dotnet-msbuild/​skills/​msbuild-server/​SKILL.md	Removes the msbuild-server skill documentation and activation guidance.
eng/​eval-quality/​underpowered-allowlist.txt	Removes the deleted eval from the underpowered allowlist.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Copilot review overview

🟢 Approval recommended

The PR cleanly removes the skill and associated eval artifacts and also eliminates remaining guidance that would have recommended enabling msbuild-server.

Review tier: Lite
Findings: 1

Pre-existing issues (1)

Severity	Finding
	eng/​eval-quality/​underpowered-allowlist.txt — PR description says we no longer want to recommend turning MSBuild Server on, but the… View comment

[ViktorHofer]

/evaluate

[github-actions]

📊 Skill Evaluation Results

2 model/skill results across 1 skill and 2 models — ✅ 0 improved, ➖ 0 not proven improved, ⚠️ 2 invalid or underpowered, ⛔ 0 activation contract failures, 📉 0 preference losses (report only).

Measurement identity: evaluated commit 6e275edf39c033643f2afc2d4f2c6d5f085e227b; 2 judge models.

Measurement health: 2 expected / 2 observed / 2 written; 0 missing, 0 unexpected, 2 invalid; 0 recovered comparison error slots and 0 unresolved comparison error slots.

Objective completion gate: not enabled. Aggregate completion transitions are telemetry only, so this report does not claim that zero objective regressions were proven.

A result passes only when preference-eligible distinct-stimulus votes have aggregate net win of at least 20% and an exact one-sided sign-test result of p ≤ 0.05, and every explicit dormancy activation contract passes. Repeated runs measure reliability only.

Skill	Model	Verdict	Gate evidence	Overfit	Warnings	Next action
build-perf-baseline	claude-sonnet-4.6	⚠️ Underpowered	n=1; 0W/0T/1L; d=1; p=0.500; net -100.0%	🔴 0.59	—	Predeclare more independent, discriminating stimuli; repeated runs do not add power.
build-perf-baseline	gpt-5.6-luna	⚠️ Underpowered	n=1; 1W/0T/0L; d=1; p=0.500; net +100.0%	🟡 0.35	—	Predeclare more independent, discriminating stimuli; repeated runs do not add power.

ℹ️ How to read this report

• ✅ Improved — the result passed both the statistical gate and the 20% practical net-win floor.
• ➖ Not proven improved — the result is valid but did not pass both gates. This is not automatically a regression.
• ⚠️ Invalid / underpowered — the gate withheld a quality verdict. Fix the measurement before judging the skill.
• ⛔ Activation contract failed — the isolated target skill activated on an explicit dormancy scenario. Dormancy preference is excluded, but this routing failure still blocks a pass.
• 📉 Preference loss — the LLM judge credibly preferred baseline. It is report-only, not objective completion proof.
• Gate evidence — n preference-eligible distinct-stimulus votes, W/T/L stimulus votes, d discordant votes, exact one-sided p, net win, and the count of separately retained dormancy stimuli. The p value applies to one model/skill result; no matrix-wide multiple-comparison correction is applied.
• Overfit — overfitting-judge severity (✅ Low, 🟡 Moderate, 🔴 High, — none) and score.
• Warnings — activation, timeout, retry recovery, or unresolved comparison conditions that need attention.
• Do not add repeated runs to increase statistical power. Do not add stimuli after seeing a near-pass unless the new breadth is predeclared for a new experiment.

⚠️ Underpowered — build-perf-baseline (claude-sonnet-4.6)

Why: Net win -100.0% (0W/0T/1L over 1 preference-eligible stimulus vote(s), sign test p=0.500), mean preference -40.0% across 1 paired run(s) — underpowered (1 preference-eligible stimulus vote(s); a credible verdict needs at least 5) — add distinct, discriminating stimuli; repeated runs do not increase task breadth

Next action: Predeclare more independent, discriminating stimuli; repeated runs do not add power.

State: INVALID_INCONCLUSIVE (underpowered)

Gate evidence: n=1; 0W/0T/1L; d=1; p=0.500; net -100.0%

Overfit: High (score 0.59)

Repeated-run reliability (not used by the gate): 1 paired run (0W/0T/1L).

Weak or warning scenarios:

Scenario	Preference gate	Net win	Δ Pref	Runs (W/T/L)
▼ Establish build performance baseline and recommend optimizations	Eligible	-100.0%	-40.0%	0/0/1

Illustrative judge evidence:

• Establish build performance baseline and recommend optimizations: B has the substantially better three-scenario baseline and useful CI guidance, but its false claim that the existing Directory.Build.props is absent is a serious analysis error that causes it to miss the solution's central, directly evidenced configuration opportunities. A is ...

This is one example, not the aggregate verdict. Open Full Results for every judgment.

⚠️ Underpowered — build-perf-baseline (gpt-5.6-luna)

Why: Net win +100.0% (1W/0T/0L over 1 preference-eligible stimulus vote(s), sign test p=0.500), mean preference +40.0% across 1 paired run(s) — underpowered (1 preference-eligible stimulus vote(s); a credible verdict needs at least 5, and this eval won every one of them) — add distinct, discriminating stimuli; repeated runs do not increase task breadth

Next action: Predeclare more independent, discriminating stimuli; repeated runs do not add power.

State: INVALID_INCONCLUSIVE (underpowered)

Gate evidence: n=1; 1W/0T/0L; d=1; p=0.500; net +100.0%

Overfit: Moderate (score 0.35)

Repeated-run reliability (not used by the gate): 1 paired run (1W/0T/0L).

🔍 Full Results - all metrics and investigation details

To investigate non-passing or warning results, paste this to your AI coding agent:

For PR 1123 in dotnet/skills, download eval artifacts with gh run download 33885875962 --repo dotnet/skills --pattern "vally-results-*" --dir ./eval-results, then fetch https://raw.githubusercontent.com/dotnet/skills/6e275edf39c033643f2afc2d4f2c6d5f085e227b/eng/vally-adapter/InvestigatingResults.md and follow it. Classify each result as measurement-invalid, underpowered, not-proven, preference-loss, or passing-with-warning. Use stateReason, result accounting, weak scenarios, and judge evidence to give the cause and exact next fix.

▶ Sessions Visualisation -- interactive replay of all evaluation sessions
📊 Session Analytics (preview) -- aggregated metrics across evaluation sessions

[github-actions]

✅ Evaluation passed for 6e275ed. cc @AbhitejJohn @JanKrivanek @dotnet/msbuild @YuliiaKovalova — please review.

[github-actions]

✅ Approved by @AbhitejJohn. cc @dotnet/skills-merge-approvers — ready to merge.