unstructured-medical-text-miner

Mine unstructured clinical text from MIMIC-IV to extract diagnostic logic and treatment details

2.00x

Quality

31%

Does it follow best practices?

Impact

76%

2.00x

Average score across 3 eval scenarios

Securityby

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./scientific-skills/Data analysis/unstructured-medical-text-miner/SKILL.md

Evaluation results

65%

15%

Clinical Note Analysis for Quality Assurance Review

Entity extraction and negation detection

Criteria

Without context

With context

Correct import path

Extract insights call

100%

Disease entity type

100%

Medication entity type

100%

Negation detection enabled

100%

66%

Negated field in output

100%

Position spans in output

100%

Medical NLP model specified

Export to JSON

100%

Summary report generated

84%

23%

Without context: $1.0511 · 4m 10s · 41 turns · 2,723 in / 13,319 out tokens

With context: $0.7957 · 2m 18s · 30 turns · 196 in / 7,524 out tokens

91%

35%

Patient Cohort Analysis from Clinical Note Records

CSV note loading and patient processing

Criteria

Without context

With context

Calls load_notes()

100%

hadm_id in output

100%

50%

extract_insights called

100%

Per-note breakdown

50%

100%

Note type filtering

100%

Patient-level retrieval

100%

Aggregated entities

100%

Results exported as JSON

100%

Handles charttime field

100%

Correct import path

28%

Without context: $0.2939 · 1m 17s · 16 turns · 22 in / 4,650 out tokens

With context: $0.5315 · 1m 32s · 23 turns · 24 in / 4,361 out tokens

73%

63%

Diagnostic Reasoning Trace Extraction for Clinical AI Training

Clinical logic parsing and relation extraction

Criteria

Without context

With context

Relation extraction enabled

100%

Timeline extraction enabled

100%

Clinical logic extraction

100%

TREATS relation in output

100%

presenting_complaint field

50%

differential_diagnoses field

50%

Timeline event structure

33%

Results saved to file

100%

Correct import path

Entities also extracted

100%

Without context: $0.5633 · 2m 23s · 24 turns · 31 in / 8,742 out tokens

With context: $0.8612 · 2m 36s · 32 turns · 3,064 in / 8,559 out tokens

Repository: aipoch/medical-research-skills
Commit: ca9aaa4

Evaluated: 17 days ago
Agent: Claude Code
Model: Claude Sonnet 4.6

Table of Contents

Clinical Note Analysis for Quality Assurance Review Patient Cohort Analysis from Clinical Note Records Diagnostic Reasoning Trace Extraction for Clinical AI Training

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.