Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →AI-powered bone marrow morphology analysis, cell classification, and hematologic disorder diagnosis using deep learning on aspirate and biopsy images.
.claude/skills/bone-marrow-ai-agent/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -14% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 48% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 51% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -5% | 0% |
<!--
#
#
-->
The Bone Marrow AI Agent provides comprehensive AI-driven analysis of bone marrow aspirate and biopsy specimens. It performs automated cell identification, differential counting, morphological assessment, and pattern recognition for hematologic disease diagnosis.
| Lineage | Cell Types | Key Features | |---------|------------|--------------| | Erythroid | Pronormoblast, basophilic, polychromatic, orthochromatic | Size, chromatin, cytoplasm color | | Myeloid | Myeloblast, promyelocyte, myelocyte, metamyelocyte, band, seg | Granules, nuclear shape | | Monocytic | Monoblast, promonocyte, monocyte | Nuclear folding, cytoplasm | | Lymphoid | Lymphocyte, plasma cell | Size, chromatin density | | Megakaryocytic | Megakaryocytes (all stages) | Size, nuclear lobation | | Other | Mast cells, osteoblasts, osteoclasts | Distinctive morphology |
User: "Analyze this bone marrow aspirate smear and provide a differential count with morphological assessment."
Agent Action:
bashpython3 Skills/Hematology/Bone_Marrow_AI_Agent/bm_analyzer.py \ --image aspirate_smear.tiff \ --stain wright_giemsa \ --target_cells 500 \ --assess_dysplasia true \ --model coatnet_bm_v2 \ --output bm_report.json
CoAtNet Hybrid Model:
Training Data Sources:
| Pattern | Associated Conditions | AI Features | |---------|----------------------|-------------| | Increased blasts | AML, MDS, ALL | Blast%, CD34 correlation | | Dysplastic features | MDS, AML-MRC | Hypolobation, ring sideroblasts | | Left shift | Infection, CML, recovery | Myeloid maturation pyramid | | Plasma cell infiltration | Myeloma, MGUS | Plasma cell%, morphology | | Lymphoid aggregates | CLL, lymphoma | Pattern, location |
| System | Approval | Application | |--------|----------|-------------| | CellaVision | FDA cleared | Peripheral blood and BM | | Scopio Labs X100 | FDA cleared | Full-field digital morphology | | Techcyte | Research | AI-powered hematology | | Morphogo | Research | Deep learning cytology |
Performance Benchmarks:
Quality Flags:
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-02 | fail→pass | 11,363 | 6,086 | -46% | 1 | 1 | 0% | 2,376 | 2,668 | +12% | 0 | 0 | — |
case-03 | fail→pass | 15,892 | 6,986 | -56% | 1 | 1 | 0% | 3,311 | 2,844 | -14% | 0 | 0 | — |
case-22 | pass→pass | 18,167 | 11,209 | -38% | 1 | 1 | 0% | 3,352 | 3,382 | +1% | 0 | 0 | — |
case-01 | fail→pass | 10,140 | 8,419 | -17% | 1 | 1 | 0% | 2,130 | 3,153 | +48% | 0 | 0 | — |
case-04 | fail→pass | 10,436 | 6,282 | -40% | 1 | 1 | 0% | 1,705 | 2,583 | +51% | 0 | 0 | — |
case-05 | pass→pass | 18,480 | 9,798 | -47% | 1 | 1 | 0% | 3,315 | 3,200 | -3% | 0 | 0 | — |
case-06 | pass→pass | 9,932 | 2,571 | -74% | 1 | 1 | 0% | 1,477 | 1,716 | +16% | 0 | 0 | — |
case-07 | pass→pass | 6,049 | 2,149 | -64% | 1 | 1 | 0% | 998 | 1,683 | +69% | 0 | 0 | — |
case-08 | pass→pass | 10,250 | 2,457 | -76% | 1 | 1 | 0% | 1,819 | 1,674 | -8% | 0 | 0 | — |
case-09 | pass→pass | 18,103 | 14,735 | -19% | 1 | 1 | 0% | 2,804 | 3,921 | +40% | 0 | 0 | — |
case-10 | pass→pass | 7,517 | 5,040 | -33% | 1 | 1 | 0% | 1,213 | 2,228 | +84% | 0 | 0 | — |
case-11 | fail→pass | 19,495 | 1,897 | -90% | 1 | 1 | 0% | 1,667 | 1,590 | -5% | 0 | 0 | — |
case-12 | pass→pass | 13,092 | 4,781 | -63% | 1 | 1 | 0% | 2,088 | 2,054 | -2% | 0 | 0 | — |
case-13 | fail→pass | 8,014 | 1,933 | -76% | 1 | 1 | 0% | 1,305 | 1,628 | +25% | 0 | 0 | — |
case-14 | pass→pass | 14,014 | 4,584 | -67% | 1 | 1 | 0% | 2,466 | 2,089 | -15% | 0 | 0 | — |
case-15 | fail→pass | 11,250 | 3,811 | -66% | 1 | 1 | 0% | 1,887 | 1,937 | +3% | 0 | 0 | — |
case-16 | pass→pass | 14,560 | 10,577 | -27% | 1 | 1 | 0% | 2,757 | 3,265 | +18% | 0 | 0 | — |
case-17 | pass→pass | 13,155 | 10,553 | -20% | 1 | 1 | 0% | 2,358 | 3,001 | +27% | 0 | 0 | — |
case-18 | pass→pass | 13,600 | 1,895 | -86% | 1 | 1 | 0% | 2,351 | 1,656 | -30% | 0 | 0 | — |
case-19 | pass→pass | 18,634 | 18,258 | -2% | 1 | 1 | 0% | 3,205 | 4,612 | +44% | 0 | 0 | — |
case-20 | pass→pass | 16,453 | 17,938 | +9% | 1 | 1 | 0% | 3,333 | 5,012 | +50% | 0 | 0 | — |
case-21 | pass→pass | 9,914 | 18,192 | +83% | 1 | 1 | 0% | 1,763 | 4,951 | +181% | 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 +32 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 | 7/27/2026 | — |
Other measured skills in the registry, with their headline benchmark lift.