Taylor 2001 ISTH SSC overt DIC (four lab items; max 8). Platelets >=100=0, 50-<100=1, <50=2. fibrin_related_marker none/moderate/strong = 0/2/3 (caller-assigned; magent does not grade D-dimer). PT sec above ULN <3=0, >=3 and <6=1, >=6=2. Fibrinogen >=1.0 g/L=0, <1.0=1 (XOR g/L or mg/dL÷100). Band <5 not compatible with overt DIC; >=5 compatible. Not a diagnosis. Not the non-overt serial score.

DRAGON (Strbian 2012 Neurology) for 3-month outcome after IV thrombolysis. CT hyperdense artery or early infarct 0/1/2, prestroke mRS >1 (1), age <65/65-79/>=80 (0/1/2), glucose >8 mmol/L (mg/dL /18) (1), onset-to-treatment >90 min (1), NIHSS 0-4/5-9/10-15/>15 (0/1/2/3); max 10. Tabulated mRS 0-2 and mRS 5-6 percents for scores 0-1, 2, 3, 8, and 9-10 only. Not a tPA decision. Not studied in basilar occlusion. magent does not read CT or assign NIHSS.

McMillan 2007 modified Glasgow Prognostic Score (mGPS) from CRP in mg/L and albumin. Elevated CRP is >10 mg/L; hypoalbuminaemia is albumin <35 g/L. mGPS 0 when CRP ≤10 (albumin ignored, including isolated hypoalbuminaemia); mGPS 1 when CRP >10 and albumin ≥35; mGPS 2 when CRP >10 and albumin <35. Not original GPS, not hs-mGPS, not a cancer diagnosis or treatment order. magent does not assay CRP or albumin.

HAT (Lou 2008 Neurology) for sICH after stroke thrombolysis. Diabetes or glucose >200 mg/dL (1; not additive). NIHSS <15 0, 15-20 1, >20 2. CT hypodensity none 0, <1/3 MCA 1, >1/3 MCA 2 (no equals-1/3). Max 5. Collective sICH 2/5/10/15/44% for scores 0/1/2/3/>3. Glucose in mg/dL (mmol × 18). Not a tPA withhold. magent does not assign NIHSS or read CT.

Hoster 2008 simplified MIPI for advanced-stage mantle cell lymphoma: age (<50 0, 50–59 1, 60–69 2, ≥70 3), ECOG (0–1 0, 2–4 2), LDH/ULN (<0.67 0, 0.67–<1 1, 1–<1.5 2, ≥1.5 3), WBC ×10^9/L (<6.7 0, 6.7–<10 1, 10–<15 2, ≥15 3); max 11. Low 0–3, intermediate 4–5, high 6–11. Deterministic published score with citation; not a lymphoma diagnosis; not continuous MIPI; not MIPI-c. magent does not assay LDH or WBC.

Guglielmelli 2018 MIPSS70 for transplant-age PMF: Hb <10 g/dL (1), leukocytes >25×10^9/L (2), platelets <100×10^9/L (2), blasts ≥2% (1), fibrosis ≥2 (1), symptoms (1), no type 1/like CALR (1), HMR category (1), ≥2 HMR (2); max 12. Low 0–1, intermediate 2–4, high ≥5. Italian-cohort median (5-year) OS 27.7 y (95%), 7.1 y (70%), 2.3 y (29%). Not MIPSS70-plus or version 2.0. magent does not sequence or grade marrow.

SEDAN (Strbian 2012 Ann Neurol) for sICH after stroke thrombolysis. Glucose 8.1-12.0 mmol/L (1) or >12.0 (2); early infarct signs (1); hyperdense cerebral artery (1); age >75 (1); admission NIHSS ≥10 (1); max 6. Derivation absolute sICH 1.4%, 2.9%, 8.5%, 12.2%, 21.7%, 33.3% for scores 0-5; no published percent for 6. Not a tPA withhold decision. magent does not assign NIHSS or read CT.

Jawhar 2019 mutation-adjusted risk score (MARS) for advanced systemic mastocytosis: age >60 (1), hemoglobin <10 g/dL (1), platelets <100×10^9/L (1), exactly one S/A/R mutation among SRSF2, ASXL1, RUNX1 (1) or two or more (2, not both). Max 5. Low 0–1 (training-set median OS not reached), intermediate 2 (3.9 y), high 3–5 (1.9 y). Not WHO SM criteria. magent does not sequence or assay labs.

Palumbo 2015 R-ISS for multiple myeloma. ISS from beta-2 microglobulin and albumin (Greipp 2005). R-ISS I: ISS I, standard-risk iFISH CA, and normal LDH. R-ISS III: ISS III plus high-risk CA (del(17p) and/or t(4;14) and/or t(14;16)) or LDH > ULN. R-ISS II: all other combinations. Not a myeloma diagnosis or treatment choice; not R2-ISS; not Durie-Salmon. magent does not assay beta-2 microglobulin, albumin, LDH, or cytogenetics.

Gwynn 2019 MRC-ICU: 39 printed ICU medication and device items (1, 2, or 3 points). Multiplier classes are weight × the caller-assigned count of qualifying medications. Non-multiplier items add the printed weight when present. Output is the integer sum plus components. Not a pharmacist-staffing order; magent does not read the MAR.

Modified Finnegan NAS from Maguire 2013 Table 1 (operational modified FNAS; not a 31-item website variant; not Maguire's 7-item short form). Exclusive cry, sleep, Moro, tremor, fever, respiratory, and stool bands plus true/false CNS, vasomotor, and GI items. Sum 0-46. at_or_above_8 iff score >= 8. magent does not examine the infant. Not Eat Sleep Console. Not a morphine order, diagnosis, or medical device.

Forrest 2003 original Glasgow Prognostic Score (GPS) from CRP in mg/L and albumin. Elevated CRP is >10 mg/L; hypoalbuminaemia is albumin <35 g/L. Assign 1 for CRP >10 and 1 for albumin <35, then sum: GPS 0 when CRP ≤10 and albumin ≥35; GPS 1 when CRP ≤10 and albumin <35 (isolated hypoalbuminaemia) or CRP >10 and albumin ≥35; GPS 2 when CRP >10 and albumin <35. Not mGPS, not hs-mGPS, not a cancer diagnosis or treatment order. magent does not assay CRP or albumin.

Neonatal glucose infusion rate (GIR, mg/kg/min) from the unit conversion used with AAP 2011 postnatal glucose homeostasis: GIR = (dextrose_percent x rate_ml_h) / (6 x weight_kg), or (dextrose_percent x rate_ml_kg_h) / 6. Provide rate_ml_h plus weight_kg, or rate_ml_kg_h, not both rate styles. Dextrose 2.5-25% and weight 0.4-10 kg are magent clamps, not AAP input bands. Reports GIR only. Does not order dextrose or TPN. Not a medical device.

Douketis 2022 ACCP elective perioperative DOAC interruption (PAUSE-style, recommendations 22-25). Returns hold_days_before and resume_days_after from INN (apixaban, rivaroxaban, edoxaban, dabigatran) and bleed_risk. dabigatran also needs caller-assigned CrCl. heparin_bridge is always false (recommendation 26). Not warfarin, not urgent surgery, not a prescription. magent does not assay CrCl.

Space milestone probabilities — discover Predix event markets on Base Sepolia and place tiny testnet orders.

Use this when the question is whether an ST0x tokenized stock on Base trades above or below the real share: send a ticker (NVDA, wtAAPL, FGI) and get the premium/discount in basis points from the deepest USDC pool (live) and the last Raindex fill vs the exchange quote, corrected for the ERC-4626 wrapper rate, with market session, age of every leg, pool depth, any cross-issuer pool (wtNVDA vs Coinbase NVDAc), and a one-sentence read. Unknown ticker -> 404, no charge.

Use this when you want to know what is actually trading in tokenized stocks on Base right now: the last 24 hours on ST0x's own DEX (Raindex) per token — USDC volume, trade count, last on-chain price and time — plus each token's issued vault supply and ERC-4626 wrapper rate, sorted by volume, with platform totals. Read straight from the public subgraphs and RPC; refreshed every 5 minutes.

EU + UK VAT number validation. Use when you must confirm an EU (27 states + XI) or UK customer's VAT number before reverse-charge invoicing or B2B onboarding, and want the registered name/address plus an official VIES consultation number as audit proof (proof=true).

What crypto tickers are heating up across crypto social right now: a ranked cross-source attention index fusing Farcaster cashtag mentions across curated crypto channels (keyless open-protocol hub) with crypto-news mention counts and a GDELT macro coverage-trend. Each row carries an attention score, a recent-vs-prior velocity, the channels driving it and cast sentiment — signal, not a raw feed. ToS-clean (no Twitter/Reddit).

Crime-risk verdict for a US location from public FBI UCR and NIBRS data. Give an address, a latitude and longitude, an agency ORI code, or a state, and get a 0-100 crime risk score, a low, medium, high or critical rating, a violent and property breakdown, a multi-year trend and a comparison with the national baseline, citing the reporting agency and the data as-of date. Deterministic, no AI. Aggregate risk indicators, not a consumer report.