NDF-RT Query Skill

SkillFiles & storage

Query NDF-RT (National Drug File Reference Terminology) via the NCI EVS REST API. Use when looking up drug mechanisms of action, physiological effects, pharmacologic classes, chemical structures, or drug, disease relationships (may_treat / may_prevent) in NDF-RT. Accepts drug names or NDF-RT codes.

Instructions available. Your AI can read the instructions. Execution depends on the setup they require.

Account requirements not reviewed. Check the skill instructions before use; ahel provides instructions and does not run this skill.

Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

Then ask your AI: use the NDF-RT Query Skill skill

What this skill tells your AI

The instructions your AI receives, as published by qsong-github/drugclaw in skills/drug_ontology/ndfrt/SKILL.md and read by ahel’s review.

Search NDF-RT concepts by drug name or NDF-RT code via the NCI EVS REST API. Auto-detects entity type by pattern:

Input Pattern	Detected As	Lookup Logic
`N0000145918`	NDF-RT code (`N` + 10 digits)	direct concept GET
anything else	free text	full-text search endpoint

API

Function	Input	Returns
`search(entity, limit)`	single entity string	`dict` with `entity`, `type`, `results`
`search_batch(entities, limit)`	list of entity strings	`dict[str, dict]`
`summarize(result, entity)`	search result dict + label	compact multi-line text
`to_json(result)`	search result dict	`list[dict]`

Result dict structure

Each item in results contains:

Key	Description
`code`	NDF-RT concept code
`name`	Preferred concept name
`kind`	Concept kind (e.g. `DRUG_KIND`, `DISEASE_KIND`)
`properties`	flat `{type: value}` property dict
`roles`	list of `{type, relatedCode, relatedName}` role relationships
`parents`	list of `{code, name}` ISA parent concepts
`synonyms`	deduplicated list of synonym strings

Key NDF-RT relationship types (roles)

Role	Meaning
`has_MoA`	mechanism of action
`has_PE`	physiological effect
`has_EPC`	established pharmacologic class
`may_treat`	indicated disease/condition
`may_prevent`	preventable disease/condition
`CI_with`	contraindicated with
`has_Chemical_Structure`	chemical structure classification

Usage

See if __name__ == "__main__" block in 52_NDF_RT.py for runnable examples covering: drug name search, code lookup, batch search, JSON output, and root concept listing.

Data

Source: NCI EVS REST API (https://api-evsrest.nci.nih.gov/api/v1)
Terminology: ndfrt
Auth: none required
Rate limits: not documented; use reasonable request spacing
Maintainer: U.S. Department of Veterans Affairs / NCI Enterprise Vocabulary Services

Signals

GitHub stars: 116
Forks: 3
Last commit: Aug 2026

Advanced

Item type: skill
Key: x-52-ndf-rt
Source: github.com/qsong-github/drugclaw

github.com/qsong-github/drugclaw