Install any skill in seconds. Free to start, no credit card required.
Get Started Free →AI-powered myeloproliferative neoplasm monitoring for disease progression prediction, treatment response tracking, and transformation risk assessment in PV, ET, and myelofibrosis.
.claude/skills/mpn-progression-monitor-agent/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-15 | ✗→✓ | ▲ Improved | 90% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 65% | 0% |
| case-12 | ✓→✓ | = Same ✓ | 290% | 0% |
| case-13 | ✓→✓ | = Same ✓ | 154% | 0% |
| case-14 | ✓→✓ | = Same ✓ | 64% | 0% |
<!--
#
#
-->
The MPN Progression Monitor Agent provides comprehensive monitoring of myeloproliferative neoplasms (PV, ET, MF) for disease progression, treatment response, and transformation risk. It integrates molecular profiling, clinical parameters, and AI-based risk models to guide management of chronic phase disease and predict blast transformation.
| MPN Type | Driver Mutations | Progression Risk | |----------|------------------|------------------| | PV | JAK2 V617F (95%), JAK2 exon 12 | Fibrosis 10-15%, AML 2-5% | | ET | JAK2 (55%), CALR (25%), MPL (5%) | Fibrosis 5-10%, AML 1-2% | | Pre-PMF | Same as PMF | Variable | | PMF | JAK2 (60%), CALR (25%), MPL (5%) | AML 10-20% |
| Mutation | Impact | MF Association | |----------|--------|----------------| | ASXL1 | Adverse | Strong | | SRSF2 | Adverse | Strong (PMF) | | EZH2 | Adverse | Moderate | | IDH1/2 | Adverse | Transformation | | RUNX1 | Very Adverse | Transformation | | TP53 | Very Adverse | Transformation | | U2AF1 | Adverse | Moderate |
| Score | Components | Application | |-------|------------|-------------| | IPSS | Age, Hb, WBC, blasts, symptoms | PMF at diagnosis | | DIPSS | Same, dynamic | PMF follow-up | | DIPSS+ | + karyotype, transfusion, platelets | PMF refined | | MIPSS70 | Molecular markers | Transplant-age PMF | | MIPSS70+ v2.0 | + U2AF1, karyotype | Most comprehensive | | MTSS | Transplant-specific | Allo-HSCT outcomes |
User: "Monitor this myelofibrosis patient's disease trajectory and update risk scores with new molecular data."
Agent Action:
bashpython3 Skills/Hematology/MPN_Progression_Monitor_Agent/mpn_monitor.py \ --patient_id MF_001 \ --molecular_data serial_mutations.csv \ --cbc_data serial_cbc.csv \ --clinical_data symptoms.json \ --mpn_type pmf \ --baseline_date 2024-01-15 \ --calculate_scores dipss,mipss70 \ --output mpn_monitoring/
| Data Type | Parameters | Frequency | |-----------|------------|-----------| | Molecular | JAK2/CALR/MPL VAF, NGS panel | q3-6 months | | CBC | Hb, WBC, platelets, blasts | Monthly | | Clinical | Symptoms, spleen size | q3 months | | Bone Marrow | Fibrosis grade, cytogenetics | q6-12 months |
| Output | Description | Format | |--------|-------------|--------| | Risk Scores | DIPSS, MIPSS70 over time | .csv | | VAF Trends | Mutation burden plots | .png | | Progression Alert | Warning if criteria met | .json | | Response Assessment | IWG-MRT criteria | .json | | Transplant Timing | Recommendation if indicated | .json | | Clone Evolution | New mutations, clonal shifts | .csv |
| Progression Type | Criteria | Action | |------------------|----------|--------| | Clinical | New symptoms, splenomegaly | Intensify therapy | | Hematologic | Cytopenias, increased blasts | BMB, cytogenetics | | Molecular | New high-risk mutations | Risk restaging | | Fibrotic | Increased fibrosis grade | Consider transplant | | Blast Phase | ≥20% blasts | Urgent intervention |
| Response | Definition | Implications | |----------|------------|--------------| | Complete Remission | No disease manifestations | Excellent outcome | | Partial Remission | >50% improvement | Good response | | Clinical Improvement | Symptom/spleen improvement | Benefit | | Stable Disease | No change | Observe | | Progressive Disease | Progression criteria | Change therapy |
Progression Prediction:
Clone Tracking:
Transplant Decision:
| Therapy | Response Markers | Timeline | |---------|------------------|----------| | Ruxolitinib | Spleen, symptoms, JAK2 VAF | 12-24 weeks | | Fedratinib | Similar to ruxolitinib | 24 weeks | | Momelotinib | + anemia improvement | 24 weeks | | Interferon | Molecular response, JAK2 VAF | 12+ months |
| Factor | Risk Increase | Management | |--------|---------------|------------| | Age >60 | 2-3x | Cytoreduction | | Prior thrombosis | 3-5x | Anticoagulation | | JAK2 V617F | 2x | Higher for homozygous | | High WBC | 1.5-2x | Control counts | | CV risk factors | Additive | Aggressive management |
| Indication | Criteria | Timing | |------------|----------|--------| | High-Risk PMF | MIPSS70+ high/very high | Consider early | | Blast Phase | ≥20% blasts | Urgent if fit | | Refractory Disease | Failed JAKi | Evaluate | | Transfusion Dependence | RBC/platelet dependent | Factor in decision |
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-12 | pass→pass | 3,523 | 4,699 | +33% | 1 | 1 | 0% | 735 | 2,866 | +290% | 0 | 0 | — |
case-13 | pass→pass | 7,830 | 7,574 | -3% | 1 | 1 | 0% | 1,306 | 3,315 | +154% | 0 | 0 | — |
case-14 | pass→pass | 16,103 | 15,066 | -6% | 1 | 1 | 0% | 2,835 | 4,641 | +64% | 0 | 0 | — |
case-05 | pass→pass | 9,183 | 8,445 | -8% | 1 | 1 | 0% | 1,812 | 3,499 | +93% | 0 | 0 | — |
case-06 | pass→pass | 8,890 | 6,311 | -29% | 1 | 1 | 0% | 1,648 | 3,117 | +89% | 0 | 0 | — |
case-07 | pass→pass | 5,533 | 5,102 | -8% | 1 | 1 | 0% | 995 | 2,873 | +189% | 0 | 0 | — |
case-01 | pass→pass | 4,607 | 4,459 | -3% | 1 | 1 | 0% | 911 | 2,732 | +200% | 0 | 0 | — |
case-02 | pass→pass | 9,649 | 7,504 | -22% | 1 | 1 | 0% | 1,730 | 3,218 | +86% | 0 | 0 | — |
case-03 | pass→pass | 12,830 | 9,600 | -25% | 1 | 1 | 0% | 2,227 | 3,616 | +62% | 0 | 0 | — |
case-04 | pass→pass | 7,092 | 6,235 | -12% | 1 | 1 | 0% | 1,395 | 3,170 | +127% | 0 | 0 | — |
case-08 | pass→pass | 4,796 | 6,991 | +46% | 1 | 1 | 0% | 870 | 3,131 | +260% | 0 | 0 | — |
case-09 | pass→pass | 8,895 | 5,221 | -41% | 1 | 1 | 0% | 1,497 | 2,809 | +88% | 0 | 0 | — |
case-10 | pass→pass | 19,504 | 13,184 | -32% | 1 | 1 | 0% | 3,472 | 4,205 | +21% | 0 | 0 | — |
case-11 | pass→pass | 9,962 | 7,618 | -24% | 1 | 1 | 0% | 1,773 | 3,230 | +82% | 0 | 0 | — |
case-15 | fail→pass | 10,070 | 7,171 | -29% | 1 | 1 | 0% | 1,689 | 3,211 | +90% | 0 | 0 | — |
case-16 | pass→pass | 6,697 | 6,939 | +4% | 1 | 1 | 0% | 1,193 | 2,748 | +130% | 0 | 0 | — |
case-17 | fail→pass | 9,426 | 6,270 | -33% | 1 | 1 | 0% | 1,813 | 2,999 | +65% | 0 | 0 | — |
case-18 | pass→pass | 8,489 | 6,585 | -22% | 1 | 1 | 0% | 1,474 | 3,030 | +106% | 0 | 0 | — |
case-19 | pass→pass | 6,104 | 5,995 | -2% | 1 | 1 | 0% | 1,023 | 2,992 | +192% | 0 | 0 | — |
case-20 | pass→pass | 6,124 | 5,258 | -14% | 1 | 1 | 0% | 1,197 | 2,951 | +147% | 0 | 0 | — |
case-21 | pass→pass | 4,271 | 6,865 | +61% | 1 | 1 | 0% | 848 | 3,211 | +279% | 0 | 0 | — |
case-22 | pass→pass | 6,698 | 6,235 | -7% | 1 | 1 | 0% | 1,491 | 3,375 | +126% | 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 +9 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 | 7/26/2026 | — |
Other measured skills in the registry, with their headline benchmark lift.