[skills-eval] dotnet-test-migration: 5 skills, 60% pass — 1 P0, 1 P1, 1 to strengthen, 2 keep

[AbhitejJohn]

🔄 EDITED 2026-07-19 — correction per #909

This finding has been revised. The errored trials that drove the FIX-RELIABILITY (P0) flag(s) below were shown in #909 to be judge-side infrastructure failures — a disabled/throttled PAT (CAPIError 400 organization disabled) and Vally's session.idle judge timeout — not fixture nondeterminism. Verdict scores already exclude errored trials, so re-running the classifier with those judge errors dropped reclassifies the affected skill(s):

Skill	Was	Now
migrate-xunit-to-mstest	FIX-RELIABILITY (P0)	TRIM-COST (P1)

Struck-through text below is the superseded finding — in particular the "pin fixture SDK/tool versions" remediation does not apply. The title's P0 count has been updated. No eval re-run is required to correct the scores; recovering the lost judgments (optional) means re-judging those trials, not re-running the skill.

Context — cross-family skill evaluation: dotnet-test-migration

This issue is self-contained: it captures everything a skill author needs to act on the dotnet-test-migration plugin without opening the full report.

What this measures. Every runnable skill in dotnet/skills was run through Vally (0.7) on a cross-family matrix: 5 executor model families — opus-4.8, gpt-5.5, sonnet-4.6, haiku-4.5, mai-flash — each judged by a different family (judge ≠ executor; default judge = latest Opus, or GPT when Opus is the executor). For every skill, a skilled run is compared against a baseline (no-skill) run and scored per executor. This removes single-model and self-judging bias, so a skill that only helps one model family — or only its own family's judge — is visible.

• Data source: cross-family CI grid (run 29228914412 + backfills) — 5 executors × 85 runnable skills, 419 scored cells over 84 skills.
• Row grain: one row per skill, aggregated across its (up to 5) executor cells. avgN is the mean trial count behind the cells (trials 1–17; higher = more statistically trustworthy). thin-N flags directional-only rows.

dotnet-test-migration at a glance (portfolio scorecard)

Plugin	Skills	Cells	Pass	Impact	Tie-trials	Err	avg ΔTok	avg ΔTurns	avg ΔTools	Headline
dotnet-test-migration	4	20	60%	0.365	39	3	−6,971	−1.77	−2.69	Efficient wins, but high ties + 3 errors

Insights. 5 skill(s); mean impact 0.36; 3/4 help ≥1 frontier model. Exemplars (win incl. frontier): migrate-vstest-to-mtp, migrate-xunit-to-xunit-v3.

Address first: migrate-xunit-to-mstest — reliability; migrate-mstest-v3-to-v4 — harness drop (no verdict).

Address first: migrate-mstest-v3-to-v4 — harness drop (no verdict); migrate-xunit-to-mstest — cost (P1). (migrate-xunit-to-mstest reclassified from reliability — see EDIT above.)

How to read the table

Each skill is scored skilled vs. baseline on these axes:

Signal	Column	What it means	Good
Breadth	Families ✓ (n/5)	How many of the 5 model families the skill helps (per-family pass count, max 5/5)	4–5 / 5
Where	Passed on	Which families passed. Frontier = latest Opus + latest GPT are bold; Sonnet 4.6 / Haiku 4.5 / MAI Flash are mid/low-weight	frontier ✓
Magnitude	Impact (−1…+1)	How strongly the judge prefers skilled over baseline	≥ 0.4
Decisiveness	Ties▵	Trials where the judge saw no difference → skill is inert	low
Safety	Loss▵	Trials where skilled was WORSE than baseline → skill misfires	~0
Reliability	Err	Trials that errored/crashed in setup or judging	0
Efficiency	ΔTok / ΔTurns / ΔTools	Extra tokens / agent turns / tool calls vs baseline	≤ 0
Confidence	avgN	Mean trials behind the verdict; low N = directional only	≥ 3
Invocation	Call%	Share of skilled trials where the model actually invoked the skill	~100%

Families ✓ is cell-level (max 5/5); Ties▵/Loss▵ are trial-level tallies summed across all families (including the ones where the skill failed). A high Families ✓ next to non-zero Loss▵ is not a contradiction — see Passed on and the Action text for where losses landed.

Action buckets (each skill has one primary action; [flags] note secondary concerns):

Bucket	Priority	Meaning
FIX-RELIABILITY	🔴 P0	Errored trials / no verdict — stabilize the harness before trusting the score
FIX-DISCOVERY	🔴 P0	Model doesn't invoke it (Call% < 50%) — a triggering/description problem
FIX-REGRESSION	🔴 P0	Skilled is worse than baseline on many trials — the skill misfires
ADD-DECISIVENESS	🟠 P1	Called ~100% but ties dominate, ~0 impact — inert; needs sharper behavioral steps
TRIM-COST	🟠 P1	Passes but with heavy token/turn overhead — trim verbosity
EXEMPLAR	🟢 keep	Broad, strong, reliable win — use as a template
EFFICIENT-WIN	🟢 protect	Wins and cuts turns/tools — the ideal shape
KEEP-POLISH	🟢	Solid majority win; minor polish + more trials
STRENGTHEN	🟡 P2	Marginal/mixed lift — sharpen triggers & success criteria

Per-skill actions

Skill	Families ✓	Passed on (frontier bold)	Impact	Ties▵	Loss▵	Err	avgN	Call%	ΔTok	ΔTurns	ΔTools	Action
migrate-vstest-to-mtp	4/5	GPT, Sonnet, Haiku, MAI	0.55	3	7	0	11	95%	-25,612	-2.11	-3.71	EXEMPLAR · ✅ Template-worthy. What's good: broad (passes 4/5, incl. frontier GPT); cheaper (2.11 turns, 3.71 tools, 25,612 tok). Keep as-is; lift its structure (crisp triggers + imperative steps) into weaker siblings. ⚠️ Caveat: frontier miss on Opus — confirm it isn't only lifting weaker models on cases frontier already handles. ([frontier-miss], [efficient])
migrate-xunit-to-xunit-v3	4/5	GPT, Sonnet, Haiku, MAI	0.45	11	6	0	12	100%	-68,917	-3.22	-4.07	EXEMPLAR · ✅ Template-worthy. What's good: broad (passes 4/5, incl. frontier GPT); cheaper (3.22 turns, 4.07 tools, 68,917 tok). Keep as-is; lift its structure (crisp triggers + imperative steps) into weaker siblings. ⚠️ Caveat: frontier miss on Opus — confirm it isn't only lifting weaker models on cases frontier already handles. ([frontier-miss], [efficient])
migrate-xunit-to-mstest	3/5	Opus, Sonnet, Haiku	0.26	15	6	3	11.4	100%	+151,598	+2.28	+2.81	FIX-RELIABILITY · 🔴 3 errored trial(s) — verdict can't be trusted until setup is deterministic. This is an infra/harness fix, not a content one: (1) capture the failing trial's stderr; (2) pin tool/SDK versions in the fixture; (3) add a setup smoke-check before scoring. ([reliability], [frontier-miss], [costly]) 🔄 EDIT (see #909): TRIM-COST · 🟠 Passes 3/5 (Opus, Sonnet, Haiku) but heavy (+151,598 tok, +2.28 turns). The 3 errored trials were judge-side session.idle timeouts (gpt ×2, haiku), not fixture flakiness — the real action is cost-trimming, not harness reliability. ([frontier-miss], [costly])
migrate-mstest-v1v2-to-v3	1/5	Haiku	0.20	10	8	0	8.8	93%	-84,953	-4.04	-5.82	STRENGTHEN · 🟡 Marginal/mixed (passes 1/5 — Haiku). Diagnosis: mostly ties — too generic/non-prescriptive. Try: (1) sharpen the trigger so it fires only where it wins; (2) add 1–2 opinionated, concrete steps that change behaviour; (3) add trials to separate signal from noise. If frontier models never benefit, scope it explicitly to weaker models or reconsider its value. ([both-frontier-miss])
migrate-mstest-v3-to-v4	—	—	—	—	—	—	—	93%	—	—	—	FIX-RELIABILITY (infra) · called 93% of trials but no verdict emitted — judge/scoring-side drop.

---

_{Generated from the cross-family Call-to-Action report (CALL-TO-ACTION.md §4–§5; companion IMPACT-ANALYSIS.md). Regenerate the underlying tables with node deep-metrics.mjs "$env:TEMP\cf-ci" agg-ci → node gen-cta-tables.mjs agg-ci. Numbers are directional where avgN is low; treat single-trial cells as hypotheses to confirm with more runs.}

[github-actions]

🎯 Issue Triage

Summary: This is a cross-family skill evaluation report for the dotnet-test-migration plugin covering 5 skills with a 60% pass rate. It documents 2 P0 reliability issues (migrate-xunit-to-mstest, migrate-mstest-v3-to-v4), 2 exemplar skills to keep (migrate-vstest-to-mtp, migrate-xunit-to-xunit-v3), and 1 skill to strengthen (migrate-mstest-v1v2-to-v3). The area is area-test (test framework migration skills).

Type: task — an evaluation work item with actionable follow-up items for skill authors.

Assigned owners: @AbhitejJohn assigned directly. The CODEOWNERS path /plugins/dotnet-test-migration/ maps to @dotnet/dotnet-testing (team — cannot be set as assignee directly; manual assignment recommended).

Confidence: High — the issue maps clearly to the dotnet-test-migration plugin, type is unambiguous, and CODEOWNERS match is exact.

Similar issues: No duplicate open issues found for this plugin's evaluation.

Triage details

• Area mapping: Issue concerns dotnet-test-migration plugin skills → area-test (already labeled; consistent with CODEOWNERS /plugins/dotnet-test-migration/)
• CODEOWNERS path match: /plugins/dotnet-test-migration/ → @dotnet/dotnet-testing
• Confidence rationale: The issue is auto-generated from CI evaluation data, clearly scoped to a single plugin, with well-defined action items per skill. No ambiguity in area or type.
• Suggested next steps:
  1. 🔴 Fix migrate-xunit-to-mstest harness reliability (3 errored trials)
  2. 🔴 Fix migrate-mstest-v3-to-v4 scoring-side drop (no verdict emitted)
  3. 🟡 Strengthen migrate-mstest-v1v2-to-v3 triggers and behavioral steps
  4. ✅ Keep migrate-vstest-to-mtp and migrate-xunit-to-xunit-v3 as exemplars

Generated by Issue Triage for issue #900 · 98.8 AIC · ⊞ 28.5K · ◷

[AbhitejJohn]

Verifying fixes for this plugin's skills

To re-run the cross-family evaluation and confirm a change moves the pass rate, cherry-pick the dispatch workflow from #920 and dispatch it scoped to this plugin:

git fetch origin abhitejjohn-cross-family-eval-workflow
git cherry-pick 7c97b7346          # single commit in #920
git push origin HEAD

gh workflow run evaluation-run.yml --ref <your-branch> \
  -f plugins="dotnet-test-migration" \
  -f executors="opus,gpt,sonnet46,haiku,mai" -f runs=3

To verify a single skill faster, scope with skills (the skill directory name, e.g. one of the skills listed above):

gh workflow run evaluation-run.yml --ref <your-branch> \
  -f skills="<skill-dir>" -f executors="opus,gpt" -f runs=3

Each skill is run skilled-vs-baseline across the 5-family executor matrix, with the judge never in the executor's family (default judge = latest Opus; GPT when Opus executes). Results assemble into a cross-family pass matrix you can compare against the numbers in this issue. Full inputs/usage: #920.