Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →24 biomedical research skills. Trigger: medical research, clinical trials, genomics, bioinformatics. Design: domain databases, wet-lab/dry-lab methods, and ethical compliance guides.
.claude/skills/brycewang-stanford-biomedical-skills/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-19 | ✗→✓ | ▲ Improved | -17% | 0% |
| case-06 | ✓→✗ | ▼ Worse | -66% | 0% |
| case-05 | ✓→✗ | ▼ Worse | -53% | 0% |
| case-09 | ✓→✗ | ▼ Worse | -42% | 0% |
| case-20 | ✓→✓ | ＝ Same ✓ | 51% | 0% |
Select the skill matching the user's need, then read its SKILL.md.
| Skill | Description | |-------|-------------| | alphafold-api | Query AlphaFold protein structure predictions by UniProt accession | | bioagents-guide | AI scientist framework for autonomous biological research workflows | | biothings-api | Query gene, variant, and drug annotations via BioThings APIs | | clawbio-guide | OpenClaw bioinformatics skill library for genomics pipelines | | clinical-dialogue-agents-guide | Papers on AI agents for clinical dialogue and medical QA | | clinical-research-guide | Design clinical studies and report using CONSORT, STROBE guidelines | | clinicaltrials-api | Clinical trial registry database search API | | clinicaltrials-api-v2 | Search and analyze clinical trials via the ClinicalTrials.gov v2 API | | ena-sequence-api | Access nucleotide sequence data from the European Nucleotide Archive | | enrichr-api | Perform gene set enrichment analysis using the Enrichr API | | ensembl-rest-api | Query gene, sequence, and variant data via the Ensembl REST API | | epidemiology-guide | Epidemiological study designs, measures of association, and public health ana... | | genomas-guide | Automate gene expression analysis with the GenoMAS multi-agent system | | genomics-analysis-guide | Workflows for RNA-seq, GWAS, and variant calling in genomic research | | genotex-benchmark-guide | Benchmark for LLM agents on gene expression data analysis | | med-researcher-guide | Multi-agent system for biomedical literature review and synthesis | | med-researcher-r1-guide | Medical deep research agent with reasoning chain analysis | | medgeclaw-guide | AI research assistant for biomedicine, RNA-seq, and drug discovery | | medical-data-api | Access FDA drug data and WHO global health statistics for research | | medical-imaging-guide | Medical image analysis with deep learning for research applications | | ncbi-blast-api | Run sequence similarity searches via the NCBI BLAST REST API | | ncbi-datasets-api | Access genomes, genes, and taxonomy data via NCBI Datasets v2 API | | pdb-structure-api | Search and retrieve 3D protein structures from the RCSB Protein Data Bank | | quickgo-api | Browse and search Gene Ontology annotations via the QuickGO API |
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-06 | pass→fail | 15,133 | 2,386 | -84% | 1 | 1 | 0% | 3,129 | 1,051 | -66% | 0 | 0 | — |
case-16 | fail→fail | 17,972 | 9,643 | -46% | 1 | 1 | 0% | 2,772 | 1,133 | -59% | 0 | 0 | — |
case-01 | fail→fail | 17,883 | 62,722 | +251% | 1 | 1 | 0% | 3,590 | 1,032 | -71% | 0 | 0 | — |
case-02 | fail→fail | 14,285 | 34,619 | +142% | 1 | 1 | 0% | 2,577 | 1,144 | -56% | 0 | 0 | — |
case-03 | fail→fail | 23,567 | 5,988 | -75% | 1 | 1 | 0% | 3,944 | 1,108 | -72% | 0 | 0 | — |
case-04 | fail→fail | 15,624 | 4,501 | -71% | 1 | 1 | 0% | 2,407 | 1,153 | -52% | 0 | 0 | — |
case-05 | pass→fail | 13,857 | 23,412 | +69% | 1 | 1 | 0% | 2,531 | 1,182 | -53% | 0 | 0 | — |
case-07 | fail→fail | 9,225 | 3,245 | -65% | 1 | 1 | 0% | 1,792 | 1,074 | -40% | 0 | 0 | — |
case-08 | fail→fail | 13,048 | 4,577 | -65% | 1 | 1 | 0% | 2,927 | 1,147 | -61% | 0 | 0 | — |
case-09 | pass→fail | 8,615 | 7,422 | -14% | 1 | 1 | 0% | 1,812 | 1,045 | -42% | 0 | 0 | — |
case-10 | fail→fail | 13,816 | 3,795 | -73% | 1 | 1 | 0% | 2,324 | 1,107 | -52% | 0 | 0 | — |
case-11 | fail→fail | 12,725 | 5,061 | -60% | 1 | 1 | 0% | 2,377 | 1,116 | -53% | 0 | 0 | — |
case-12 | fail→fail | 16,404 | 4,603 | -72% | 1 | 1 | 0% | 2,818 | 1,171 | -58% | 0 | 0 | — |
case-13 | fail→fail | 12,544 | 2,905 | -77% | 1 | 1 | 0% | 2,300 | 1,056 | -54% | 0 | 0 | — |
case-14 | fail→fail | 17,946 | 2,604 | -85% | 1 | 1 | 0% | 2,696 | 1,057 | -61% | 0 | 0 | — |
case-15 | fail→fail | 33,234 | 5,451 | -84% | 1 | 1 | 0% | 2,733 | 1,238 | -55% | 0 | 0 | — |
case-17 | fail→fail | 14,511 | 6,796 | -53% | 1 | 1 | 0% | 2,354 | 1,248 | -47% | 0 | 0 | — |
case-18 | fail→fail | 19,453 | 3,500 | -82% | 1 | 1 | 0% | 2,811 | 1,130 | -60% | 0 | 0 | — |
case-19 | fail→pass | 10,863 | 2,908 | -73% | 1 | 1 | 0% | 1,552 | 1,286 | -17% | 0 | 0 | — |
case-20 | pass→pass | 6,647 | 5,201 | -22% | 1 | 1 | 0% | 1,305 | 1,975 | +51% | 0 | 0 | — |
case-21 | pass→pass | 20,777 | 17,353 | -16% | 1 | 1 | 0% | 3,519 | 4,500 | +28% | 0 | 0 | — |
case-22 | pass→pass | 16,825 | 22,651 | +35% | 1 | 1 | 0% | 2,631 | 4,695 | +78% | 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. 22 cases were attempted, and 4 counted toward the lift figure. The other 18 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 -9 percentage points is the difference between those two pass rates over the 4 comparable cases. 12 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.