Cheaper on a smaller model
Does a skill let a smaller model match the default model's results for less money?
- docs/experiments/DETECTION-2026-09-19.md · 80e4a13 · 2026-09-20
Headline figures
Skills that beat the base model, measured on the smaller model
measured-local · 2026-09-20
Held quality on every comparable task and cost less
measured-local · n=34 · 2026-09-20
Median cost on the smaller model as a share of the default model's, downgradeable skills
measured-local · 2026-09-20
Lowest cost ratio among downgradeable skills
measured-local · 2026-09-20
Highest cost ratio among downgradeable skills
measured-local · 2026-09-20
Tasks with a comparable default-model run
measured-local · 2026-09-20
Comparable tasks where quality held
measured-local · n=41 · 2026-09-20
Downgradeable skills where the skill also lifted the smaller model above no skill
measured-local · n=16 · 2026-09-20
Skills whose quality dropped on the smaller model
measured-local · 2026-09-20
Skills that held quality but were not cheaper
measured-local · 2026-09-20
Notes
- A cost result on its own axis: it does not reduce token count and is never added to a token figure.
- Smaller model: Haiku 4.5 with the skill. Default model: Sonnet 5 with the skill, from the lane L1 runs. A skill held when the smaller model matched within 0.05 on every comparable task, which says nothing about unseen tasks.
- A skill's cost ratio is the upper median of its tasks' ratios, as compiler-site.mjs computes it.
Superseded figures
A figure this project published and later corrected, kept beside what replaced it (CLAUDE.md, "Correcting a measurement instrument").
median_cost_ratio in the study's summary.json
0.42 → 0.657, the median over downgradeable skills
The summary's median runs over every task row, and a task with no comparable default-model run records a cost ratio of 0. The reported figure is per skill, over skills that held quality and cost less; cost_ratio_recorded keeps the raw value.
docs/experiments/DETECTION-2026-09-19.md
Split of the 18 skills that are not downgradeable
14 dropped quality, 4 held quality but were not cheaper → 13 dropped quality, 5 held quality but were not cheaper
Recounted from the per-task records with the document's own rule, held means quality held on every comparable task. The document's split does not reproduce from those records, and it names terracotta as a skill that held quality, while terracotta held on 1 of its 2 comparable tasks. The 16 downgradeable skills and their cost ratios are unchanged.
docs/experiments/DETECTION-2026-09-19.md
Tables
n: Runs on the smaller model behind the row: comparable tasks times three repetitions. 34 rows.
| skill | downgradeable | quality_held_all | cheaper | skill_lifts_haiku | cost_ratio | tasks | tasks_comparable | tasks_held | n | evidence_class | date | commit |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| web-quality-audit | Yes | Yes | Yes | Yes | 0.408 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| unity-script-roles | Yes | Yes | Yes | Yes | 0.433 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| typescript-server | Yes | Yes | Yes | Yes | 0.536 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| napkin | Yes | Yes | Yes | Yes | 0.541 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| landing-page-conversion-audit | Yes | Yes | Yes | Yes | 0.573 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| compose-modifier-and-layout-style | Yes | Yes | Yes | Yes | 0.6 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| llm-council | Yes | Yes | Yes | No | 0.623 | 2 | 2 | 2 | 6 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| hermes-agent-skill-authoring | Yes | Yes | Yes | Yes | 0.656 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| catchup | Yes | Yes | Yes | Yes | 0.658 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| cost-efficiency-analyzer | Yes | Yes | Yes | Yes | 0.748 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| eol-message | Yes | Yes | Yes | Yes | 0.751 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| unity-primetween | Yes | Yes | Yes | Yes | 0.753 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| optimize-simplicite-logs | Yes | Yes | Yes | Yes | 0.776 | 2 | 2 | 2 | 6 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| geo-fundamentals | Yes | Yes | Yes | Yes | 0.799 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| caveman-commit | Yes | Yes | Yes | Yes | 0.842 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| research-agent | Yes | Yes | Yes | Yes | 0.94 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| offensive-reporting | No | No | Yes | No | 0.216 | 2 | 1 | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| debugging-network-issues | No | No | Yes | No | 0.376 | 2 | 2 | 0 | 6 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| feedback | No | No | Yes | Yes | 0.407 | 2 | 1 | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| product-vision | No | No | Yes | Yes | 0.544 | 2 | 2 | 1 | 6 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| wiki-lint | No | No | Yes | No | 0.587 | 2 | 1 | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| modeling-product-usage-metrics | No | No | Yes | Yes | 0.652 | 2 | 1 | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| data-analyst | No | No | Yes | No | 0.714 | 2 | 2 | 1 | 6 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| ads-plan | No | No | Yes | Yes | 0.859 | 2 | 1 | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| n8n-error-handling | No | No | Yes | Yes | 0.979 | 2 | 2 | 1 | 6 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| triton-kernel | No | No | No | Yes | 1.007 | 2 | 1 | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| event-store-design | No | Yes | No | Yes | 1.02 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| vision-sft | No | Yes | No | Yes | 1.036 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| latent-briefing | No | Yes | No | Yes | 1.094 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| dotnet-local-tools | No | Yes | No | Yes | 1.256 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| alphagbm-tepper-signal | No | No | No | Yes | 1.466 | 2 | 1 | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| architecture-designer | No | Yes | No | Yes | 1.524 | 2 | 1 | 1 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| friendly | No | No | No | Yes | 2.181 | 2 | 1 | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| terracotta | No | No | No | Yes | 2.545 | 2 | 2 | 1 | 6 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
n: Repetitions per arm. 68 rows.
| skill | task_id | comparable | score_haiku_skill | score_haiku_none | score_sonnet_skill | quality_held | skill_lifts_haiku | cost_ratio | cost_ratio_recorded | n | evidence_class | date | commit |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ads-plan | fitness-app-media-plan | Yes | 0.8 | 0.6 | 1 | No | Yes | 0.859 | 0.859 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| ads-plan | healthdash-optimization | No | 1 | 0.8 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| alphagbm-tepper-signal | tepper-quality-filter-validation | No | 0.8 | 0 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| alphagbm-tepper-signal | tepper-signal-multi-scenario | Yes | 0.8 | 1 | 1 | No | No | 1.466 | 1.466 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| architecture-designer | ecommerce-microservices-design | Yes | 1 | 0.6 | 1 | Yes | Yes | 1.524 | 1.524 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| architecture-designer | monolith-to-microservices-review | No | 1 | 1 | — | — | No | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| catchup | context-restoration-after-break | Yes | 1 | 0.6 | 1 | Yes | Yes | 0.658 | 0.658 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| catchup | prioritize-work-from-state | No | 0.8 | 0.8 | — | — | No | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| caveman-commit | breaking-api-endpoint | No | 0.75 | 0.75 | — | — | No | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| caveman-commit | perf-query-opt | Yes | 0.75 | 0.25 | 0.75 | Yes | Yes | 0.842 | 0.842 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| compose-modifier-and-layout-style | fix-composable-modifier-parameter | Yes | 1 | 0.75 | 1 | Yes | Yes | 0.6 | 0.6 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| compose-modifier-and-layout-style | refactor-modifier-chain-style | No | 1 | 1 | — | — | No | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| cost-efficiency-analyzer | cost-efficiency-current-quarter | Yes | 1 | 0.4 | 1 | Yes | Yes | 0.748 | 0.748 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| cost-efficiency-analyzer | cost-efficiency-trend-analysis | No | 0.6 | 0.4 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| data-analyst | compare-campaign-performance | Yes | 0.6 | 0.6 | 0.8 | No | No | 0.55 | 0.55 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| data-analyst | explore-customer-dataset | Yes | 0.8 | 0.8 | 0.8 | Yes | No | 0.714 | 0.714 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| debugging-network-issues | isolate-slow-download-throughput | Yes | 0.2 | 0.4 | 0.8 | No | No | 0.314 | 0.314 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| debugging-network-issues | timeout-root-cause-multi-layer | Yes | 0 | 0.75 | 1 | No | No | 0.376 | 0.376 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| dotnet-local-tools | fix-broken-local-tools-config | No | 1 | 0.5 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| dotnet-local-tools | setup-local-tools-fresh | Yes | 1 | 0.5 | 1 | Yes | Yes | 1.256 | 1.256 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| eol-message | eol-message-feature-replacement | Yes | 0.8 | 0.6 | 0.6 | Yes | Yes | 0.751 | 0.751 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| eol-message | eol-message-graceful-exit | No | 0.8 | 0.2 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| event-store-design | event-store-implementation | No | 1 | 1 | — | — | No | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| event-store-design | event-store-system-design | Yes | 1 | 0.8 | 1 | Yes | Yes | 1.02 | 1.02 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| feedback | extract-quiet-feedback | No | 1 | 0.5 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| feedback | specific-feedback-sbi | Yes | 0 | 0 | 1 | No | No | 0.407 | 0.407 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| friendly | dropdown-design-guidelines | Yes | 0.6 | 0.4 | 0.8 | No | Yes | 2.181 | 2.181 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| friendly | improve-incomplete-button-guidelines | No | 1 | 0.4 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| geo-fundamentals | audit-blog-posts | Yes | 1 | 0.4 | 1 | Yes | Yes | 0.799 | 0.799 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| geo-fundamentals | improve-article-for-ai-citation | No | 0.8 | 0.2 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| hermes-agent-skill-authoring | create-bundled-skill | Yes | 0.8 | 0.6 | 0.8 | Yes | Yes | 0.656 | 0.656 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| hermes-agent-skill-authoring | fix-noncompliant-skill | No | 0.8 | 0.2 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| landing-page-conversion-audit | checkout-bottleneck-diagnosis | No | 0.8 | 0.6 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| landing-page-conversion-audit | landing-page-message-match-audit | Yes | 1 | 0.4 | 1 | Yes | Yes | 0.573 | 0.573 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| latent-briefing | latent-briefing-config-tuning | No | 1 | 1 | — | — | No | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| latent-briefing | latent-briefing-design | Yes | 1 | 0.6 | 1 | Yes | Yes | 1.094 | 1.094 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| llm-council | payment-error-refactor | Yes | 0.25 | 0.25 | 0.25 | Yes | No | 0.623 | 0.623 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| llm-council | search-filter-plan | Yes | 0.25 | 0.75 | 0 | Yes | No | 0.461 | 0.461 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| modeling-product-usage-metrics | metrics-map-product-questions | No | 0.8 | 0 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| modeling-product-usage-metrics | retention-mobile-app-config | Yes | 0.4 | 0.6 | 0.8 | No | No | 0.652 | 0.652 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| n8n-error-handling | scheduled-sync-retry-alert | Yes | 0.75 | 0 | 1 | No | Yes | 0.979 | 0.979 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| n8n-error-handling | webhook-silent-error | Yes | 1 | 1 | 1 | Yes | No | 0.352 | 0.352 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| napkin | activate-napkin-whiteboard | Yes | 0.8 | 0 | 0.4 | Yes | Yes | 0.541 | 0.541 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| napkin | check-unsaved-napkin | No | 0.8 | 0.4 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| offensive-reporting | format-raw-finding | Yes | 0 | 0 | 0.6 | No | No | 0.216 | 0.216 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| offensive-reporting | write-executive-summary | No | 0.8 | 0.8 | — | — | No | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| optimize-simplicite-logs | find-and-analyze-errors-in-log | Yes | 1 | 0.75 | 1 | Yes | Yes | 0.776 | 0.776 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| optimize-simplicite-logs | parse-and-filter-simplicite-log | Yes | 1 | 1 | 1 | Yes | No | 0.606 | 0.606 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| product-vision | create-vision-from-scratch | Yes | 0 | 0 | 0.8 | No | No | 0.387 | 0.387 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| product-vision | refine-flat-vision | Yes | 1 | 0.8 | 0.8 | Yes | Yes | 0.544 | 0.544 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| research-agent | author-competitor-research-agent | Yes | 1 | 0.4 | 1 | Yes | Yes | 0.94 | 0.94 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| research-agent | author-regulation-monitoring-agent | No | 1 | 0.4 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| terracotta | fix-input-guidelines | Yes | 0.6 | 0.4 | 0.6 | Yes | Yes | 0.872 | 0.872 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| terracotta | modal-guidelines | Yes | 0.6 | 0.6 | 0.9 | No | No | 2.545 | 2.545 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| triton-kernel | fix-softmax-masking-bug | No | 1 | 0.8 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| triton-kernel | fused-vector-op-kernel | Yes | 0.8 | 0.8 | 1 | No | No | 1.007 | 1.007 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| typescript-server | product-inventory | Yes | 1 | 0.2 | 0.8 | Yes | Yes | 0.536 | 0.536 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| typescript-server | scheduled-notifications | No | 0.6 | 0.2 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| unity-primetween | generate-sequence-script | No | 0.6 | 0.4 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| unity-primetween | generate-tween-script | Yes | 1 | 0.8 | 1 | Yes | Yes | 0.753 | 0.753 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| unity-script-roles | collectible-system-architecture | Yes | 1 | 0.6 | 1 | Yes | Yes | 0.433 | 0.433 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| unity-script-roles | quest-dialogue-system-architecture | No | 1 | 0.8 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| vision-sft | vlm-config-technical-diagrams | Yes | 0.8 | 0 | 0.4 | Yes | Yes | 1.036 | 1.036 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| vision-sft | vlm-debug-silent-killers | No | 1 | 1 | — | — | No | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| web-quality-audit | comprehensive-website-audit | Yes | 1 | 0.8 | 1 | Yes | Yes | 0.408 | 0.408 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| web-quality-audit | launch-blocker-audit | No | 1 | 0.8 | — | — | Yes | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| wiki-lint | vault-audit-with-metadata | No | 0.2 | 0.6 | — | — | No | — | 0 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |
| wiki-lint | vault-health-check | Yes | 0.5 | 1 | 1 | No | No | 0.587 | 0.587 | 3 | measured-local | 2026-09-20 | 80e4a13fb96ebb96d8efff8690a8c77215657590 |