Kando Data
中ENContact us

Edoxaban Dose Check for an Atrial Fibrillation Patient

MED-AF-EDOXABAN-DOSE-002 · Cardiology progress note · anticoagulation in non-valvular AF

As a cardiology resident, find the applicable dosing provisions in 25 untitled full-text papers and, given renal function, body weight and co-medication, write the current daily dose with an item-by-item justification.

Clinical medicationAnticoagulationLiterature searchItem-by-item reasoning
2525 files
4.0Mcharacters of material
9scoring criteria
20agent turns in the longest reference rollout
1.28Mcumulative prompt tokens in one rollout
01

Highlights

25 untitled papers, 4 million characters

Files are named only “文献-01” to “文献-25” with no titles. The model has to work out what each paper is before it can find the provisions that apply.

A conclusion alone is not enough

The note must state which renal-function measure sets the dose band, whether each other reduction criterion is met, and why every other dose mentioned in the materials does not apply.

Distractors from real charts

A coronary stent history, discontinued antiplatelet drugs and two renal-function measures given side by side — each can push the model into the wrong dose band.

Penalties target clinical risk

3 of the 9 criteria are penalty items, and 2 positive criteria are rated Critically Important.

02

Instruction excerpt

An excerpt of the instruction the agent actually receives, showing the task setup and constraints.

instruction.md · 1,303 chars

Scenario and task

You are a cardiology resident. The handover date is 2026-09-15. The edoxaban dose of a patient with atrial fibrillation was not recorded clearly. Check the handover notes and the literature below, and add a progress note. The note must state the current daily dose and the basis for it.

Write out the basis item by item, not just the conclusion:

  • which renal-function measure sets the dose band, and which reduction range in the materials it falls into;
  • whether each other reduction criterion in the materials (body weight, co-medication) is met in this case, stated one by one;
  • for every other daily dose that appears in the materials, why it is not used here.

Handover notes

Diagnosis: non-valvular atrial fibrillation. Date of birth 1962-03-08. Echo on 2025-08-02: no mechanical valve, mild mitral regurgitation, no moderate or severe stenosis. One coronary stent placed on 2024-11-03. Aspirin and clopidogrel stopped on 2025-06-01. No bleeding history.

Current medication: edoxaban, dose not recorded. Also metoprolol succinate extended-release 47.5 mg once daily and atorvastatin calcium 20 mg every night.

Measured on 2026-09-14: body weight 70 kg. Creatinine clearance 22 ml/min. eGFR 19 ml/min/1.73m². Not on dialysis.

NOTEEnglish translation of the original Chinese instruction; the deliverable requirements are omitted.

03

Attachments

Everything visible inside the agent container, with statistics generated from the task package. Original files ship with the full sample package.

FileTypeSizeBytes
文献-01.txttxt72K chars114 KB
文献-02.txttxt30K chars57 KB
文献-03.txttxt35K chars63 KB
文献-04.txttxt89K chars130 KB
文献-05.txttxt37K chars61 KB
文献-06.txttxt60K chars106 KB
文献-07.txttxt148K chars209 KB
文献-08.txttxt20K chars42 KB
文献-09.txttxt228K chars330 KB
文献-10.txttxt594K chars583 KB
文献-11.txttxt63K chars152 KB
文献-12.txttxt150K chars213 KB
文献-13.txttxt37K chars62 KB
文献-14.txttxt30K chars57 KB
文献-15.txttxt90K chars132 KB
文献-16.txttxt61K chars106 KB
文献-17.txttxt92K chars150 KB
文献-18.txttxt20K chars42 KB
文献-19.txttxt230K chars333 KB
文献-20.txttxt39K chars66 KB
文献-21.txttxt722K chars709 KB
文献-22.txttxt131K chars129 KB
文献-23.txttxt228K chars330 KB
文献-24.txttxt752K chars738 KB
文献-25.txttxt61K chars152 KB
04

Scoring design

9independent scoring criteria
6 scoring items3 penalty items

Full criteria text, weights and reference answers ship with the complete data package.

9 binary criteria covering correctness of conclusion, numerical accuracy, factual faithfulness and professional standards.

  • 9 criteria: 6 positive scoring items and 3 penalty items (negate)
  • Each criterion is judged separately by an agent judge (individual mode), reproducible and auditable
  • Judge credentials are injected only into the verifier; the agent container has no scoring channel
  • Full criteria text and reference answers ship with the complete data package
Conclusion correctness4
Professional standards2
Instruction following1
Numerical accuracy1
Factual faithfulness1
05

Model results

Trajectory statistics from frontier models in real agent harnesses. One turn is one model call; the context curve shows how state accumulates over a long-horizon task.

ModelAgent turnsTool callsCumulative prompt tokensPeak contextTimeResult
GPT-5.6-sol (Codex CLI)Codex CLI20161.28M—1m 47sCompleted
06

Specification

TASK SPECIFICATION
Task IDMED-AF-EDOXABAN-DOSE-002
Task formatHarbor task.toml · schema 1.4
Package version1.0.1
Agent time limit20 h
Verifier time limit5 h
Agent container2 vCPU · 8 GB memory · 30 GB storage
Tool setfilesystem · shell
Scoring isolationVerifier runs as a separate stageJudge credentials are injected only into the verifier and are invisible to the agent
Instruction length1,303 characters
LiteratureParsed text from a medical literature library; file names carry no titles
Reference solutionoracle reward 1.000
Contact

Want to run this task on your own platform?

Leave your details and we will send the full sample package for this task (instruction, attachments, verifier and reference trajectories) and walk you through volume and customisation for similar tasks.

  • Full task package and verifier
  • More tasks in this subdomain
  • Integration and re-run support

REMED-AF-EDOXABAN-DOSE-002 · Edoxaban Dose Check for an Atrial Fibrillation Patient

Used only to follow up on this request.