Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →16 full-text access skills. Trigger: accessing paper PDFs, bulk downloading, open access, text mining. Design: legal full-text retrieval from open repositories, archives, and preprint servers.
.claude/skills/brycewang-stanford-fulltext-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-04 | ✗→✓ | ▲ Improved | -58% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -54% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -47% | 0% |
| case-12 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-13 | ✗→✓ | ▲ Improved | -5% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | arxiv-latex-source | Download and parse LaTeX source files from arXiv preprints | | bioc-pmc-api | Access PMC Open Access articles in BioC format for text mining | | core-api-guide | Search and retrieve open access research papers via CORE aggregator | | dataverse-api | Deposit and discover research datasets via Harvard Dataverse API | | doaj-api | Search open access journals and articles in the DOAJ directory | | hal-archive-api | Access French and European research via the HAL open archive API | | institutional-repository-guide | Access papers from institutional and subject repositories at scale | | interlibrary-loan-guide | Access papers through interlibrary loan and document delivery services | | open-access-guide | Navigate open access policies, repositories, and legal full-text retrieval me... | | open-access-mining-guide | Mine open access full-text repositories for research data extraction | | osf-api | Manage open science projects and preprints via the OSF REST API | | pmc-ftp-bulk-download | Bulk download PMC Open Access articles via FTP for large-scale mining | | pmc-oai-api | PubMed Central OAI-PMH metadata harvesting | | preprint-servers-guide | Guide to preprint servers across scientific disciplines | | unpaywall-api | Find free legal full-text versions of scholarly articles via Unpaywall | | zotero-ai-butler-guide | AI-powered paper summarization plugin for Zotero |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 9,145 | 11,955 | +31% | 1 | 1 | 0% | 1,565 | 1,281 | -18% | 0 | 0 | — |
case-02 | fail→fail | 16,063 | 5,714 | -64% | 1 | 1 | 0% | 2,659 | 1,251 | -53% | 0 | 0 | — |
case-03 | pass→pass | 10,828 | 4,400 | -59% | 1 | 1 | 0% | 1,970 | 1,002 | -49% | 0 | 0 | — |
case-04 | fail→pass | 9,887 | 2,723 | -72% | 1 | 1 | 0% | 1,813 | 762 | -58% | 0 | 0 | — |
case-05 | fail→fail | 9,646 | 2,747 | -72% | 1 | 1 | 0% | 1,419 | 845 | -40% | 0 | 0 | — |
case-06 | fail→fail | 11,856 | 2,415 | -80% | 1 | 1 | 0% | 2,178 | 760 | -65% | 0 | 0 | — |
case-07 | fail→pass | 12,093 | 2,841 | -77% | 1 | 1 | 0% | 2,071 | 955 | -54% | 0 | 0 | — |
case-08 | fail→fail | 15,912 | 4,282 | -73% | 1 | 1 | 0% | 2,701 | 1,007 | -63% | 0 | 0 | — |
case-09 | pass→fail | 6,974 | 6,565 | -6% | 1 | 1 | 0% | 1,120 | 1,073 | -4% | 0 | 0 | — |
case-10 | fail→fail | 16,467 | 4,125 | -75% | 1 | 1 | 0% | 2,670 | 857 | -68% | 0 | 0 | — |
case-11 | fail→pass | 15,661 | 7,468 | -52% | 1 | 1 | 0% | 2,323 | 1,225 | -47% | 0 | 0 | — |
case-12 | fail→pass | 11,061 | 4,174 | -62% | 1 | 1 | 0% | 1,896 | 1,313 | -31% | 0 | 0 | — |
case-13 | fail→pass | 7,647 | 3,513 | -54% | 1 | 1 | 0% | 1,354 | 1,289 | -5% | 0 | 0 | — |
case-14 | fail→fail | 15,837 | 5,960 | -62% | 1 | 1 | 0% | 2,552 | 774 | -70% | 0 | 0 | — |
case-15 | fail→fail | 12,884 | 2,510 | -81% | 1 | 1 | 0% | 2,402 | 809 | -66% | 0 | 0 | — |
case-16 | fail→fail | 9,414 | 5,207 | -45% | 1 | 1 | 0% | 1,576 | 1,018 | -35% | 0 | 0 | — |
case-17 | pass→fail | 9,501 | 4,206 | -56% | 1 | 1 | 0% | 1,492 | 945 | -37% | 0 | 0 | — |
case-18 | pass→fail | 8,175 | 5,161 | -37% | 1 | 1 | 0% | 1,425 | 915 | -36% | 0 | 0 | — |
case-19 | pass→fail | 13,297 | 8,570 | -36% | 1 | 1 | 0% | 2,245 | 789 | -65% | 0 | 0 | — |
case-20 | pass→fail | 4,519 | 3,876 | -14% | 1 | 1 | 0% | 748 | 872 | +17% | 0 | 0 | — |
case-21 | fail→pass | 17,380 | 13,465 | -23% | 1 | 1 | 0% | 671 | 1,865 | +178% | 0 | 0 | — |
case-22 | fail→pass | 13,014 | 6,524 | -50% | 1 | 1 | 0% | 2,394 | 1,716 | -28% | 0 | 0 | — |
case-23 | fail→pass | 10,343 | 5,985 | -42% | 1 | 1 | 0% | 1,471 | 1,669 | +13% | 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. 23 cases were attempted, and 10 counted toward the lift figure. The other 13 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 +13 percentage points is the difference between those two pass rates over the 10 comparable cases. 5 cases got worse with the skill loaded, and they are included in that figure.
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.