Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →24 metadata & bibliometrics skills. Trigger: DOI resolution, citation metrics, author disambiguation, bibliometrics. Design: metadata APIs and bibliometric analysis tools for scholarly records.
.claude/skills/brycewang-stanford-metadata-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 3% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -50% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -3% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | academic-paper-summarizer | Summarize academic papers with structured extraction of key elements | | altmetrics-guide | Guide to altmetrics and research impact beyond traditional citations | | bibliometrix-guide | Perform science mapping and bibliometric analysis with R bibliometrix | | citation-network-guide | Analyze citation networks, impact metrics, and bibliometric patterns | | crossref-api | Resolve DOIs and retrieve publication metadata from CrossRef registry | | crossref-event-data-api | Track scholarly mentions across the web via Crossref Event Data | | datacite-api | Resolve dataset DOIs and query research data metadata via DataCite | | doi-content-negotiation | Retrieve structured metadata from any DOI via HTTP content negotiation | | doi-resolution-guide | DOI content negotiation and metadata retrieval techniques | | h-index-guide | Understanding and calculating research impact metrics | | journal-metrics-guide | Understand journal impact factors, h5-index, CiteScore, and SJR | | opencitations-api | Query open citation data and reference networks via OpenCitations | | orcid-api | Look up researcher profiles and academic identities via the ORCID registry | | orcid-integration-guide | Set up and leverage ORCID for researcher identification and profiles | | orkg-api | Query the Open Research Knowledge Graph for structured research data | | plumx-metrics-api | Track research impact beyond citations via PlumX altmetrics API | | ror-organization-api | Identify and link research organizations via the ROR registry API | | sophosia-reference-guide | Reference manager with PDF viewer and Markdown note support | | viaf-authority-api | Disambiguate author identities via the VIAF authority file API | | wikidata-api-guide | Query Wikidata SPARQL for scholarly metadata, authors, and entities | | zoplicate-dedup-guide | Detect and manage duplicate items in Zotero libraries | | zotero-actions-tags-guide | Zotero workflow automation with custom actions and tags | | zotmoov-guide | Zotero plugin for automatic attachment file organization | | zutilo-guide | Zotero utility plugin for keyboard shortcuts and batch editing |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 8,121 | 4,819 | -41% | 1 | 1 | 0% | 1,284 | 1,324 | +3% | 0 | 0 | — |
case-02 | fail→pass | 7,423 | 4,964 | -33% | 1 | 1 | 0% | 1,082 | 1,789 | +65% | 0 | 0 | — |
case-03 | fail→pass | 13,655 | 2,084 | -85% | 1 | 1 | 0% | 2,478 | 1,236 | -50% | 0 | 0 | — |
case-04 | fail→pass | 13,663 | 5,634 | -59% | 1 | 1 | 0% | 2,302 | 1,901 | -17% | 0 | 0 | — |
case-05 | fail→pass | 8,499 | 3,256 | -62% | 1 | 1 | 0% | 1,513 | 1,463 | -3% | 0 | 0 | — |
case-06 | fail→pass | 12,011 | 2,304 | -81% | 1 | 1 | 0% | 1,966 | 1,303 | -34% | 0 | 0 | — |
case-07 | fail→pass | 10,414 | 3,960 | -62% | 1 | 1 | 0% | 1,428 | 1,639 | +15% | 0 | 0 | — |
case-08 | fail→fail | 13,697 | 5,366 | -61% | 1 | 1 | 0% | 2,466 | 1,120 | -55% | 0 | 0 | — |
case-09 | fail→pass | 15,886 | 2,460 | -85% | 1 | 1 | 0% | 2,717 | 1,303 | -52% | 0 | 0 | — |
case-10 | fail→pass | 8,530 | 2,326 | -73% | 1 | 1 | 0% | 1,520 | 1,332 | -12% | 0 | 0 | — |
case-11 | fail→pass | 9,599 | 1,703 | -82% | 1 | 1 | 0% | 1,544 | 1,101 | -29% | 0 | 0 | — |
case-12 | fail→pass | 8,490 | 1,873 | -78% | 1 | 1 | 0% | 1,600 | 1,173 | -27% | 0 | 0 | — |
case-13 | fail→pass | 10,529 | 2,699 | -74% | 1 | 1 | 0% | 1,984 | 1,211 | -39% | 0 | 0 | — |
case-14 | fail→pass | 10,717 | 4,189 | -61% | 1 | 1 | 0% | 1,705 | 1,632 | -4% | 0 | 0 | — |
case-15 | fail→pass | 9,070 | 2,929 | -68% | 1 | 1 | 0% | 1,544 | 1,422 | -8% | 0 | 0 | — |
case-16 | fail→pass | 10,182 | 11,494 | +13% | 1 | 1 | 0% | 1,777 | 1,535 | -14% | 0 | 0 | — |
case-17 | fail→pass | 12,408 | 1,708 | -86% | 1 | 1 | 0% | 2,005 | 1,156 | -42% | 0 | 0 | — |
case-18 | fail→pass | 15,078 | 2,892 | -81% | 1 | 1 | 0% | 2,449 | 1,482 | -39% | 0 | 0 | — |
case-19 | fail→pass | 9,398 | 2,636 | -72% | 1 | 1 | 0% | 1,488 | 1,355 | -9% | 0 | 0 | — |
case-20 | fail→pass | 11,951 | 1,660 | -86% | 1 | 1 | 0% | 2,087 | 1,098 | -47% | 0 | 0 | — |
case-21 | fail→pass | 13,915 | 3,065 | -78% | 1 | 1 | 0% | 2,235 | 1,456 | -35% | 0 | 0 | — |
case-22 | fail→pass | 15,086 | 3,036 | -80% | 1 | 1 | 0% | 2,375 | 1,445 | -39% | 0 | 0 | — |
case-23 | fail→pass | 11,718 | 3,861 | -67% | 1 | 1 | 0% | 1,980 | 1,557 | -21% | 0 | 0 | — |
case-24 | fail→pass | 16,574 | 3,249 | -80% | 1 | 1 | 0% | 2,750 | 1,523 | -45% | 0 | 0 | — |
case-25 | pass→pass | 13,849 | 9,665 | -30% | 1 | 1 | 0% | 2,545 | 2,807 | +10% | 0 | 0 | — |
case-26 | pass→pass | 13,103 | 11,719 | -11% | 1 | 1 | 0% | 2,231 | 3,067 | +37% | 0 | 0 | — |
case-27 | pass→pass | 15,518 | 13,071 | -16% | 1 | 1 | 0% | 2,466 | 3,206 | +30% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 27 cases were attempted, and 26 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +85 percentage points is the difference between those two pass rates over the 26 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.