Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →End-to-end TCR/BCR repertoire analysis from FASTQ to clonotype diversity metrics. Use when analyzing immune repertoire sequencing data from bulk or single-cell experiments.
.claude/skills/bio-workflows-tcr-pipeline/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -33% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -21% | 0% |
| case-22 | ✗→✓ | ▲ Improved | -52% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -49% | 0% |
<!--
#
#
-->
FASTQ → MiXCR align → Assemble → Export → VDJtools diversity → Visualizationbash# Align reads to V(D)J segments mixcr align -s hsa -p rna-seq \ R1.fastq.gz R2.fastq.gz \ aligned.vdjca # Assemble clonotypes mixcr assemble aligned.vdjca clones.clns # Export mixcr exportClones clones.clns clones.txt
bash# Convert to VDJtools format vdjtools Convert -S mixcr clones.txt vdjtools/ # Diversity metrics vdjtools CalcDiversityStats vdjtools/clones.txt diversity/ # Sample overlap vdjtools CalcPairwiseDistances vdjtools/*.txt overlap/
bash# Spectratype plot vdjtools PlotFancySpectratype vdjtools/clones.txt spectra/ # V usage vdjtools PlotFancyVJUsage vdjtools/clones.txt usage/
<!-- 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→pass | 13,484 | 6,392 | -53% | 1 | 1 | 0% | 2,776 | 1,865 | -33% | 0 | 0 | — |
case-02 | fail→pass | 12,378 | 8,710 | -30% | 1 | 1 | 0% | 2,506 | 2,012 | -20% | 0 | 0 | — |
case-03 | fail→pass | 18,501 | 10,383 | -44% | 1 | 1 | 0% | 3,340 | 2,646 | -21% | 0 | 0 | — |
case-22 | fail→pass | 9,614 | 2,143 | -78% | 1 | 1 | 0% | 1,811 | 869 | -52% | 0 | 0 | — |
case-08 | pass→pass | 3,515 | 2,031 | -42% | 1 | 1 | 0% | 614 | 862 | +40% | 0 | 0 | — |
case-09 | fail→pass | 9,401 | 2,197 | -77% | 1 | 1 | 0% | 1,654 | 850 | -49% | 0 | 0 | — |
case-04 | pass→pass | 7,782 | 4,534 | -42% | 1 | 1 | 0% | 1,447 | 1,322 | -9% | 0 | 0 | — |
case-05 | pass→pass | 12,545 | 9,676 | -23% | 1 | 1 | 0% | 2,407 | 2,447 | +2% | 0 | 0 | — |
case-06 | pass→pass | 6,411 | 4,735 | -26% | 1 | 1 | 0% | 1,144 | 1,284 | +12% | 0 | 0 | — |
case-07 | pass→pass | 8,860 | 2,196 | -75% | 1 | 1 | 0% | 1,571 | 936 | -40% | 0 | 0 | — |
case-10 | pass→pass | 8,573 | 1,965 | -77% | 1 | 1 | 0% | 1,528 | 881 | -42% | 0 | 0 | — |
case-11 | pass→pass | 7,216 | 2,130 | -70% | 1 | 1 | 0% | 1,049 | 850 | -19% | 0 | 0 | — |
case-12 | fail→pass | 3,766 | 1,685 | -55% | 1 | 1 | 0% | 616 | 749 | +22% | 0 | 0 | — |
case-13 | pass→pass | 9,054 | 2,021 | -78% | 1 | 1 | 0% | 1,647 | 825 | -50% | 0 | 0 | — |
case-14 | pass→pass | 8,451 | 3,414 | -60% | 1 | 1 | 0% | 1,528 | 1,017 | -33% | 0 | 0 | — |
case-15 | fail→pass | 8,163 | 1,501 | -82% | 1 | 1 | 0% | 1,346 | 726 | -46% | 0 | 0 | — |
case-16 | fail→pass | 13,484 | 1,761 | -87% | 1 | 1 | 0% | 2,367 | 745 | -69% | 0 | 0 | — |
case-17 | fail→pass | 7,506 | 1,683 | -78% | 1 | 1 | 0% | 1,188 | 742 | -38% | 0 | 0 | — |
case-18 | pass→pass | 4,056 | 2,325 | -43% | 1 | 1 | 0% | 718 | 870 | +21% | 0 | 0 | — |
case-19 | pass→pass | 6,419 | 2,409 | -62% | 1 | 1 | 0% | 1,037 | 881 | -15% | 0 | 0 | — |
case-20 | pass→pass | 3,524 | 1,616 | -54% | 1 | 1 | 0% | 663 | 784 | +18% | 0 | 0 | — |
case-21 | fail→pass | 15,375 | 3,358 | -78% | 1 | 1 | 0% | 2,818 | 1,179 | -58% | 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. The headline lift of +45 percentage points is the difference between those two pass rates over the 22 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 | 7/24/2026 | +32% |
Other measured skills in the registry, with their headline benchmark lift.