CtrlK
BlogDocsLog inGet started
Tessl Logo

nih-biosketch-builder

Generate NIH Biosketch documents compliant with the 2022 OMB-approved format

Install with Tessl CLI

npx tessl i github:aipoch/medical-research-skills --skill nih-biosketch-builder
What are skills?

61

Does it follow best practices?

Validation for skill structure

SKILL.md
Review
Evals

Evaluation results

100%

26%

NIH Grant Application Biosketch

NIH biosketch document generation from JSON input

Criteria
Without context
With context

Uses main.py script

0%

100%

Correct --input flag

0%

100%

Correct --output flag

0%

100%

DOCX output produced

100%

100%

Installs python-docx

100%

100%

Section A present

100%

100%

Section B present

100%

100%

Section C present

100%

100%

Section D present

50%

100%

Max 4 contributions

100%

100%

JSON schema: personal_info

100%

100%

JSON schema: contributions array

100%

100%

JSON schema: positions_and_honors

100%

100%

Without context: $0.3700 · 1m 49s · 17 turns · 19 in / 6,876 out tokens

With context: $0.7854 · 2m 24s · 31 turns · 238 in / 7,735 out tokens

100%

41%

Biosketch Preparation with Publication Records

PubMed publication auto-import into NIH biosketch

Criteria
Without context
With context

Uses --import-pubmed flag

0%

100%

PMIDs comma-separated

33%

100%

Publications JSON produced

100%

100%

NCBI E-utilities API used

100%

100%

Fetches by PMID

100%

100%

DOCX biosketch produced

100%

100%

Uses main.py script

0%

100%

Installs requests library

0%

100%

Input JSON includes PMIDs

100%

100%

Two-step workflow

100%

100%

Without context: $0.9106 · 3m 41s · 47 turns · 174 in / 13,572 out tokens

With context: $0.7382 · 2m 3s · 30 turns · 33 in / 7,004 out tokens

90%

-10%

NIH Biosketch Format Audit

NIH format compliance: fonts, margins, and section labels

Criteria
Without context
With context

Approved font family

100%

100%

Font size at least 11pt

100%

16%

Margins at least 0.5 inches

100%

100%

Section A label

100%

100%

Section B label

100%

100%

Section C label

100%

100%

Section D label

100%

100%

Audit script checks font

100%

100%

Audit script checks margins

100%

100%

Audit report produced

100%

100%

Installs python-docx

100%

100%

Without context: $0.4612 · 2m 16s · 17 turns · 22 in / 9,245 out tokens

With context: $0.7910 · 2m 28s · 28 turns · 153 in / 9,327 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.