Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add hannesill/m4 --skill mimic-sofa-24hgit clone --depth 1 https://github.com/hannesill/m4Wrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/hannesill/m4/mimic-sofa-24h)<a href="https://agentmods.dev/skills/hannesill/m4/mimic-sofa-24h"><img src="https://agentmods.dev/badge/skills/hannesill/m4/mimic-sofa-24h/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/hannesill/m4/mimic-sofa-24h"><img src="https://agentmods.dev/badge/skills/hannesill/m4/mimic-sofa-24h.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.02621 |
| Opus 5 | $0.00000 | $0.01311 |
| Sonnet 5 | $0.00000 | $0.00524 |
| Haiku 4.5 | $0.00000 | $0.00262 |
Grade A, and why
mimic-sofa-24h scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
100% identical to mimic-apsiii-24h-raw — 1,135 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 272 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Reference SQL (matched-content control)
The following block is the public reference SQL used to construct the ground truth for this task. It is provided verbatim, without procedural prose, to test whether matched task-relevant content alone explains the WITH-SKILL gain.
-- ------------------------------------------------------------------
-- Title: Sequential Organ Failure Assessment (SOFA) score
-- This query extracts the SOFA score for the first day of each ICU
-- patient's stay. SOFA quantifies organ dysfunction across 6 systems,
-- each scored 0-4, with a total range of 0-24.
-- ------------------------------------------------------------------
-- Reference for SOFA:
-- Vincent JL et al. "The SOFA (Sepsis-related Organ Failure
-- Assessment) score to describe organ dysfunction/failure."
-- Intensive Care Medicine. 1996;24(7):707-710.
-- Adapted from mimic-code first_day_sofa.sql
-- Uses data from 6 hours before to 24 hours after ICU admission.
-- Components:
-- Respiration: PaO2/FiO2 ratio (arterial only) with ventilation status
-- Coagulation: Platelet count
-- Liver: Bilirubin
-- Cardiovascular: MAP + vasopressor doses (mcg/kg/min)
-- CNS: Glasgow Coma Scale
-- Renal: Creatinine + urine output (mL/day)
--
-- DEVIATION from mimic-code: first_day_sofa does not filter blood gases by
-- specimen. The benchmark restricts PaO2/FiO2 to arterial specimens to match
-- the SOFA definition and task instructions.
WITH vaso_stg AS (
SELECT
ie.stay_id,
'norepinephrine' AS treatment,
vaso_rate AS rate
FROM mimiciv_icu.icustays AS ie
INNER JOIN mimiciv_derived.norepinephrine AS mv
ON ie.stay_id = mv.stay_id
AND mv.starttime >= ie.intime - INTERVAL '6' HOUR
AND mv.starttime <= ie.intime + INTERVAL '1' DAY
UNION ALL
SELECT
ie.stay_id,
'epinephrine' AS treatment,
vaso_rate AS rate
FROM mimiciv_icu.icustays AS ie
INNER JOIN mimiciv_derived.epinephrine AS mv
ON ie.stay_id = mv.stay_id
AND mv.starttime >= ie.intime - INTERVAL '6' HOUR
AND mv.starttime <= ie.intime + INTERVAL '1' DAY
UNION ALL
SELECT
ie.stay_id,
'dobutamine' AS treatment,
vaso_rate AS rate
FROM mimiciv_icu.icustays AS ie
INNER JOIN mimiciv_derived.dobutamine AS mv
ON ie.stay_id = mv.stay_id
AND mv.starttime >= ie.intime - INTERVAL '6' HOUR
AND mv.starttime <= ie.intime + INTERVAL '1' DAY
UNION ALL
SELECT
ie.stay_id,
'dopamine' AS treatment,
vaso_rate AS rate
FROM mimiciv_icu.icustays AS ie
INNER JOIN mimiciv_derived.dopamine AS mv
ON ie.stay_id = mv.stay_id
AND mv.starttime >= ie.intime - INTERVAL '6' HOUR
AND mv.starttime <= ie.intime + INTERVAL '1' DAY
)
, vaso_mv AS (
SELECT
ie.stay_id,
MAX(CASE WHEN treatment = 'norepinephrine' THEN rate ELSE NULL END) AS rate_norepinephrine,
MAX(CASE WHEN treatment = 'epinephrine' THEN rate ELSE NULL END) AS rate_epinephrine,
MAX(CASE WHEN treatment = 'dopamine' THEN rate ELSE NULL END) AS rate_dopamine,
MAX(CASE WHEN treatment = 'dobutamine' THEN rate ELSE NULL END) AS rate_dobutamine
FROM mimiciv_icu.icustays AS ie
LEFT JOIN vaso_stg AS v
ON ie.stay_id = v.stay_id
GROUP BY
ie.stay_id
)
, pafi1 AS (
SELECT
ie.stay_id,
bg.charttime,
bg.pao2fio2ratio,
CASE WHEN NOT vd.stay_id IS NULL THEN 1 ELSE 0 END AS isvent
FROM mimiciv_icu.icustays AS ie
LEFT JOIN mimiciv_derived.bg AS bg
ON ie.subject_id = bg.subject_id
AND bg.charttime >= ie.intime - INTERVAL '6' HOUR
AND bg.charttime <= ie.intime + INTERVAL '1' DAY
AND bg.specimen = 'ART.'
LEFT JOIN mimiciv_derived.ventilation AS vd
ON ie.stay_id = vd.stay_id
AND bg.charttime >= vd.starttime
AND bg.charttime <= vd.endtime
AND vd.ventilation_status = 'InvasiveVent'
)
, pafi2 AS (
SELECT
stay_id,
MIN(CASE WHEN isvent = 0 THEN pao2fio2ratio ELSE NULL END) AS pao2fio2_novent_min,
MIN(CASE WHEN isvent = 1 THEN pao2fio2ratio ELSE NULL END) AS pao2fio2_vent_min
FROM pafi1
GROUP BY
stay_id
)
, scorecomp AS (
SELECT
ie.stay_id,
v.mbp_min,
mv.rate_norepinephrine,
mv.rate_epinephrine,
mv.rate_dopamine,
mv.rate_dobutamine,
l.creatinine_max,
l.bilirubin_total_max AS bilirubin_max,
l.platelets_min AS platelet_min,
pf.pao2fio2_novent_min,
pf.pao2fio2_vent_min,
uo.urineoutput,
gcs.gcs_min
FROM mimiciv_icu.icustays AS ie
LEFT JOIN vaso_mv AS mv
ON ie.stay_id = mv.stay_id
LEFT JOIN pafi2 AS pf
ON ie.stay_id = pf.stay_id
LEFT JOIN mimiciv_derived.first_day_vitalsign AS v
ON ie.stay_id = v.stay_id
LEFT JOIN mimiciv_derived.first_day_lab AS l
ON ie.stay_id = l.stay_id
LEFT JOIN mimiciv_derived.first_day_urine_output AS uo
ON ie.stay_id = uo.stay_id
LEFT JOIN mimiciv_derived.first_day_gcs AS gcs
ON ie.stay_id = gcs.stay_id
)
, scorecalc AS (
SELECT
stay_id,
CASE
WHEN pao2fio2_vent_min < 100
THEN 4
WHEN pao2fio2_vent_min < 200
THEN 3
WHEN pao2fio2_novent_min < 300
THEN 2
WHEN pao2fio2_novent_min < 400
THEN 1
WHEN COALESCE(pao2fio2_vent_min, pao2fio2_novent_min) IS NULL
THEN NULL
ELSE 0
END AS respiration,
CASE
WHEN platelet_min < 20
THEN 4
WHEN platelet_min < 50
THEN 3
WHEN platelet_min < 100
THEN 2
WHEN platelet_min < 150
THEN 1
WHEN platelet_min IS NULL
THEN NULL
ELSE 0
END AS coagulation,
CASE
WHEN bilirubin_max >= 12.0
THEN 4
WHEN bilirubin_max >= 6.0
THEN 3
WHEN bilirubin_max >= 2.0
THEN 2
WHEN bilirubin_max >= 1.2
THEN 1
WHEN bilirubin_max IS NULL
THEN NULL
ELSE 0
END AS liver,
CASE
WHEN rate_dopamine > 15 OR rate_epinephrine > 0.1 OR rate_norepinephrine > 0.1
THEN 4
-- Deviates from mimic-code which uses <= 0.1 (true for any
-- non-NULL rate). We use the clinically correct > 0 AND <= 0.1.
-- No effect on derived values: vasopressor tables only contain
-- rows with positive rates, so the conditions are equivalent.
WHEN rate_dopamine > 5 OR (rate_epinephrine > 0 AND rate_epinephrine <= 0.1) OR (rate_norepinephrine > 0 AND rate_norepinephrine <= 0.1)
THEN 3
WHEN rate_dopamine > 0 OR rate_dobutamine > 0
THEN 2
WHEN mbp_min < 70
THEN 1
WHEN COALESCE(mbp_min, rate_dopamine, rate_dobutamine, rate_epinephrine, rate_norepinephrine) IS NULL
THEN NULL
ELSE 0
END AS cardiovascular,
CASE
WHEN (gcs_min >= 13 AND gcs_min <= 14)
THEN 1
WHEN (gcs_min >= 10 AND gcs_min <= 12)
THEN 2
WHEN (gcs_min >= 6 AND gcs_min <= 9)
THEN 3
WHEN gcs_min < 6
THEN 4
WHEN gcs_min IS NULL
THEN NULL
ELSE 0
END AS cns,
CASE
WHEN (creatinine_max >= 5.0)
THEN 4
WHEN urineoutput < 200
THEN 4
WHEN (creatinine_max >= 3.5 AND creatinine_max < 5.0)
THEN 3
WHEN urineoutput < 500
THEN 3
WHEN (creatinine_max >= 2.0 AND creatinine_max < 3.5)
THEN 2
WHEN (creatinine_max >= 1.2 AND creatinine_max < 2.0)
THEN 1
WHEN COALESCE(urineoutput, creatinine_max) IS NULL
THEN NULL
ELSE 0
END AS renal
FROM scorecomp
)
SELECT
ie.subject_id, ie.hadm_id, ie.stay_id
-- Combine all the scores to get SOFA
-- Impute 0 if the score is missing
, COALESCE(respiration, 0)
+ COALESCE(coagulation, 0)
+ COALESCE(liver, 0)
+ COALESCE(cardiovascular, 0)
+ COALESCE(cns, 0)
+ COALESCE(renal, 0)
AS sofa
-- DEVIATION from mimic-code: COALESCE component scores to 0.
-- The original SQL leaves them NULL when underlying data is missing.
-- We impute 0 here so the ground truth matches the task instruction
-- ("treat missing data as normal, score 0") and agents are not
-- penalised for following the instruction. The NULL→0 semantics are
-- already applied to the sofa total above; this extends it to the
-- individual components for consistency.
, COALESCE(respiration, 0) AS respiration
, COALESCE(coagulation, 0) AS coagulation
, COALESCE(liver, 0) AS liver
, COALESCE(cardiovascular, 0) AS cardiovascular
, COALESCE(cns, 0) AS cns
, COALESCE(renal, 0) AS renal
FROM mimiciv_icu.icustays ie
LEFT JOIN scorecalc s
ON ie.stay_id = s.stay_id
;
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 272 lines · 0 tokens per session scan A f83028155d98
mimic-sofa-24h is a skill published in the GitHub repository hannesill/m4 (43 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,621 tokens. A static security scan graded it A with 0 findings. It is 100% identical to mimic-apsiii-24h-raw, differing in 1,135 lines, and is treated as a copy.
Other skills, from other repositories
arbor-agent-orchestrator
Top-level controller for recreating the open-source AutoResearch workflow as a suite of skills. Use when the user asks to run, emulate, extract, validate, or refine Arbor/AutoResearch behavior, especially when a coordinator must load phase skills for setup, ideation, executors, merge evaluation, novelty search…
dfam-check
Measure mesh files against Design for Additive Manufacturing (DfAM) rules and report printability findings per process (FDM, SLS, SLA/DLP, metal PBF, MJF). Use when the user asks whether a part is printable, wants overhang/wall-thickness/support analysis of an .stl, .obj, .ply, or .3mf mesh, wants a build-orientation…
tooluniverse-gene-enrichment
Gene-set enrichment analysis — GO (Biological Process, Molecular Function, Cellular Component), KEGG, Reactome pathway enrichment via clusterProfiler, gseapy, ORA, GSEA. Use for interpreting DEG lists, screen hit lists, or any gene-list-to-pathways query. Includes simplify-cutoff handling and union-vs-total…
tooluniverse-phylogenetics
Phylogenetic analysis — de novo multiple sequence alignment (Clustal Omega/MUSCLE/MAFFT via EBImsaalign) and neighbour-joining/UPGMA tree building (EBIbuildphylogenetictree) from your own sequences, plus tree analysis, treeness, saturation (PhyKIT), parsimony-informative sites, alignment gap analysis, DVMC…
tooluniverse-cell-line-profiling
Cancer cell-line selection and profiling for experimental model choice. Cross-references DepMap, Cellosaurus, COSMIC, PharmacoDB to deliver identity verification, mutation/CNV profile, gene dependencies, drug sensitivities, and druggable targets. Use to answer 'which cell line should I use for studying gene X?' or 'is…
tooluniverse-pharmacogenomics
Pharmacogenomics (PGx) research — drug-gene interactions (CPIC, PharmGKB), CPIC dosing guidelines, variant-drug-response associations, ethnic-allele-frequency considerations, and metabolizer-status scoring. Use for PGx-informed dosing recommendations, CYP/HLA pharmacogenomic allele interpretation, and…