CtrlK
BlogDocsLog inGet started
Tessl Logo

ecrf-designer

Design clinical trial CRFs with proper validation rules

Install with Tessl CLI

npx tessl i github:aipoch/medical-research-skills --skill ecrf-designer
What are skills?

44

Does it follow best practices?

Validation for skill structure

SKILL.md
Review
Evals

Evaluation results

100%

54%

Demographics CRF for Phase II Oncology Trial

Script-based CRF generation

Criteria
Without context
With context

Script invoked

0%

100%

Template argument used

0%

100%

Output argument used

0%

100%

JSON output produced

100%

100%

Standard fields present

0%

100%

ECOG field type correct

100%

100%

ECOG validation range

100%

100%

Diagnosis field type correct

0%

100%

Diagnosis max_length validation

100%

100%

No undefined field types

0%

100%

No external calls

100%

100%

Without context: $0.1794 · 56s · 11 turns · 16 in / 3,468 out tokens

With context: $0.4729 · 1m 34s · 26 turns · 2,125 in / 4,970 out tokens

100%

26%

Adverse Event CRF Design for Regulatory Submission

CDISC SDTM compliance and complete CRF documentation

Criteria
Without context
With context

CDISC domain specified

100%

100%

Visit schedule documented

100%

100%

Data dictionary produced

100%

100%

Skip logic documented

100%

100%

Validation rules file

100%

100%

Severity field type

0%

100%

Severity options defined

100%

100%

Date field type

100%

100%

Yes/No field type

0%

100%

ae_crf JSON valid

100%

100%

No undefined field types

0%

100%

No external packages

100%

100%

Without context: $0.6953 · 3m 49s · 15 turns · 61 in / 18,201 out tokens

With context: $1.0148 · 4m 15s · 34 turns · 41 in / 17,168 out tokens

92%

40%

Efficacy Assessment CRF for Cardiovascular Trial

Custom field types and validation rules

Criteria
Without context
With context

Script used for generation

0%

100%

Efficacy template argument

0%

0%

Systolic BP field type

0%

100%

Systolic BP min/max validation

100%

100%

Heart rate field type

100%

100%

Heart rate min/max validation

100%

100%

Response field type

0%

100%

Response options present

100%

100%

Calculated field type

0%

100%

Formula in calculated field

100%

100%

Data dictionary produced

100%

100%

No undefined field types

0%

100%

JSON output valid

100%

100%

Without context: $0.3171 · 1m 36s · 13 turns · 16 in / 6,122 out tokens

With context: $0.5507 · 2m 13s · 21 turns · 2,058 in / 8,125 out tokens

Evaluated
Agent
Claude Code

Table of Contents

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.