Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →AI-powered extracellular vesicle and exosome analysis for cancer biomarker discovery, liquid biopsy applications, and intercellular communication profiling.
.claude/skills/exosome-ev-analysis-agent/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-19 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-20 | ✗→✓ | ▲ Improved | 11% | 0% |
| case-08 | ✓→✓ | ＝ Same ✓ | 24% | 0% |
| case-09 | ✓→✓ | ＝ Same ✓ | 30% | 0% |
| case-15 | ✓→✓ | ＝ Same ✓ | 91% | 0% |
<!--
#
#
-->
The Exosome/EV Analysis Agent provides comprehensive AI-driven analysis of extracellular vesicles for cancer biomarker discovery, liquid biopsy applications, and tumor-microenvironment communication profiling.
| Type | Size | Origin | Markers | |------|------|--------|---------| | Exosomes | 30-150 nm | MVB fusion | CD9, CD63, CD81 | | Microvesicles | 100-1000 nm | Membrane budding | Annexin V, ARF6 | | Apoptotic bodies | 500-5000 nm | Cell death | Annexin V, PS | | Large oncosomes | 1-10 μm | Tumor-specific | Variable |
User: "Analyze exosomal miRNA profiles from plasma samples to identify pancreatic cancer biomarkers."
Agent Action:
bashpython3 Skills/Oncology/Exosome_EV_Analysis_Agent/ev_analyzer.py \ --ev_mirna exosome_smallrna.tsv \ --ev_protein exosome_proteome.tsv \ --sample_groups pancreatic_cancer,healthy \ --normalization spike_in \ --biomarker_discovery true \ --output ev_biomarker_report/
| Cancer Type | Elevated miRNAs | Clinical Use | |-------------|-----------------|--------------| | Pancreatic | miR-21, miR-17-5p, miR-155 | Early detection | | Lung | miR-21, miR-126, miR-210 | Screening | | Colorectal | miR-21, miR-92a, miR-29a | Detection | | Prostate | miR-141, miR-375, miR-1290 | Prognosis | | Ovarian | miR-21, miR-141, miR-200 family | Detection | | Breast | miR-21, miR-155, miR-10b | Metastasis |
| Method | Principle | Purity | Yield | Scalability | |--------|-----------|--------|-------|-------------| | Ultracentrifugation | Density | Moderate | High | Low | | Size exclusion | Size | High | Moderate | Moderate | | Immunocapture | Surface markers | Very high | Low | Low | | Precipitation | Polymer | Low | Very high | High | | Microfluidics | Various | Variable | Low | Low |
Biomarker Discovery:
Source Deconvolution:
Functional Prediction:
MISEV Guidelines Requirements:
AI Group - Biomedical AI Platform
<!-- AUTHOR_SIGNATURE: 9a7f3c2e-MD-BABU-MIA-2026-MSSM-SECURE -->
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 29,935 | 31,703 | +6% | 1 | 1 | 0% | 5,506 | 7,502 | +36% | 0 | 0 | — |
case-08 | pass→pass | 16,525 | 13,022 | -21% | 1 | 1 | 0% | 2,941 | 3,645 | +24% | 0 | 0 | — |
case-09 | pass→pass | 16,757 | 14,061 | -16% | 1 | 1 | 0% | 2,914 | 3,788 | +30% | 0 | 0 | — |
case-15 | pass→pass | 7,081 | 5,173 | -27% | 1 | 1 | 0% | 1,184 | 2,264 | +91% | 0 | 0 | — |
case-16 | pass→pass | 11,711 | 8,812 | -25% | 1 | 1 | 0% | 2,235 | 3,037 | +36% | 0 | 0 | — |
case-17 | pass→pass | 7,845 | 7,700 | -2% | 1 | 1 | 0% | 1,466 | 2,849 | +94% | 0 | 0 | — |
case-18 | pass→pass | 13,524 | 10,677 | -21% | 1 | 1 | 0% | 2,483 | 3,191 | +29% | 0 | 0 | — |
case-02 | fail→fail | 14,829 | 5,250 | -65% | 1 | 1 | 0% | 2,815 | 1,670 | -41% | 0 | 0 | — |
case-03 | pass→pass | 19,788 | 26,517 | +34% | 1 | 1 | 0% | 3,656 | 6,528 | +79% | 0 | 0 | — |
case-04 | pass→pass | 20,632 | 17,608 | -15% | 1 | 1 | 0% | 3,326 | 4,420 | +33% | 0 | 0 | — |
case-05 | pass→pass | 20,465 | 22,992 | +12% | 1 | 1 | 0% | 3,457 | 5,453 | +58% | 0 | 0 | — |
case-06 | pass→pass | 6,261 | 5,430 | -13% | 1 | 1 | 0% | 1,181 | 2,361 | +100% | 0 | 0 | — |
case-07 | pass→pass | 12,005 | 12,499 | +4% | 1 | 1 | 0% | 2,342 | 3,532 | +51% | 0 | 0 | — |
case-10 | pass→pass | 11,361 | 7,646 | -33% | 1 | 1 | 0% | 2,043 | 2,701 | +32% | 0 | 0 | — |
case-11 | pass→pass | 8,967 | 11,108 | +24% | 1 | 1 | 0% | 1,531 | 3,281 | +114% | 0 | 0 | — |
case-12 | pass→pass | 13,015 | 13,291 | +2% | 1 | 1 | 0% | 2,388 | 3,365 | +41% | 0 | 0 | — |
case-13 | pass→pass | 14,090 | 9,464 | -33% | 1 | 1 | 0% | 2,682 | 3,173 | +18% | 0 | 0 | — |
case-14 | pass→pass | 18,192 | 5,870 | -68% | 1 | 1 | 0% | 1,238 | 2,392 | +93% | 0 | 0 | — |
case-19 | fail→pass | 11,846 | 3,314 | -72% | 1 | 1 | 0% | 1,790 | 2,008 | +12% | 0 | 0 | — |
case-20 | fail→pass | 11,157 | 4,703 | -58% | 1 | 1 | 0% | 2,037 | 2,257 | +11% | 0 | 0 | — |
case-21 | pass→pass | 7,214 | 6,153 | -15% | 1 | 1 | 0% | 1,319 | 2,414 | +83% | 0 | 0 | — |
case-22 | pass→pass | 10,021 | 6,234 | -38% | 1 | 1 | 0% | 1,629 | 2,439 | +50% | 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 21 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 +9 percentage points is the difference between those two pass rates over the 21 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.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 7/26/2026 | +9% |
Other measured skills in the registry, with their headline benchmark lift.