skills.rest
mims-harvard
mims-harvardOFFICIAL
Explore all skills by this author →
Website𝕏@marinkazitnikGitHub

tooluniverse-chemical-safety

Assess chemical toxicity by integrating ADMET-AI predictions, CTD toxicogenomics, and regulatory safety data.

1.7k|254|Updated Mar 3, 2025
Search Tags:#FDA#toxicology#admet#chemical safety#Risk Assessment#toxicogenomics#hazard classification
One-click install
npx skills add https://github.com/mims-harvard/ToolUniverse --skill tooluniverse-chemical-safety
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: tooluniverse-chemical-safety
Source: https://github.com/mims-harvard/ToolUniverse/tree/main/plugins/tooluniverse/skills/tooluniverse-chemical-safety
Command: npx skills add https://github.com/mims-harvard/ToolUniverse --skill tooluniverse-chemical-safety

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Evaluating whether a chemical or drug is safe requires pulling evidence from many disconnected sources — predictive models, toxicogenomics databases, FDA labels, and hazard classifications — and reconciling conflicting findings. This Skill orchestrates that entire workflow into a single structured toxicology report with graded evidence.

Core Features & Use Cases

  • Predictive Toxicology & ADMET Profiling: Runs ADMET-AI endpoints (AMES, DILI, hERG, LD50, carcinogenicity) plus absorption, distribution, metabolism, and excretion predictions from a SMILES string.
  • Experimental & Regulatory Evidence: Retrieves GHS/IARC hazard classifications and LD50 values from PubChemTox, chemical-gene-disease associations from CTD, FDA boxed warnings and contraindications, and DrugBank safety data.
  • Evidence-Graded Risk Report: Produces a structured markdown report classifying overall risk as Critical/High/Medium/Low, with every finding annotated by evidence tier [T1]-[T4] and explicit data gaps.
  • Use Case: Given a drug name like acetaminophen, the Skill resolves it to SMILES and PubChem CID, runs toxicity predictions, queries CTD and FDA labels, and delivers an integrated safety dossier flagging hepatotoxicity risk with supporting evidence tiers.

Quick Start

Assess the safety and toxicity profile of acetaminophen, including ADMET predictions, FDA warnings, and an overall risk classification.

Frequently Asked Questions about tooluniverse-chemical-safety

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I assess the toxicity of a chemical from its SMILES string?▼

Provide the SMILES string and the Skill runs ADMET-AI toxicity endpoints including AMES mutagenicity, DILI, hERG inhibition, carcinogenicity, and LD50, then adds ADMET property predictions. Results are graded as computational [T3] evidence and combined with database findings.

What databases does this chemical safety assessment query?▼

It queries ADMET-AI for predictions, CTD for chemical-gene-disease associations, PubChemTox for GHS/IARC classifications and LD50 values, FDA labels for boxed warnings and contraindications, DrugBank for drug safety, STITCH for chemical-protein interactions, and ChEMBL for structural alerts.

Can this Skill assess environmental chemicals that are not drugs?▼

Yes. For environmental chemicals it skips FDA and DrugBank phases and relies on CTD toxicogenomics, PubChemTox hazard data, AOP-Wiki adverse outcome pathways, and EPA facility data for exposure context instead.

What happens if ADMET-AI predictions are unavailable?▼

ADMET-AI tools require the tooluniverse[ml] extra; if unavailable, the Skill falls back to CTD and PubChemTox experimental data and notes the prediction gap. Failed individual endpoints are marked as prediction unavailable rather than failing the whole report.

How reliable are the toxicity predictions in the report?▼

Every finding carries an evidence tier: [T1] for human/regulatory evidence like FDA warnings, [T2] for animal and validated in vitro studies, [T3] for computational predictions, and [T4] for unvalidated annotations. When predictions conflict with experimental data, the experimental finding takes precedence.

Related Skills

Explore More Education & Research Skills→
biomaps-infrabiomaps-infra

db-pubchem

Query PubChem via PUG-REST API for compound properties and searches.

Official
Intermediate
xpert-aixpert-ai
8

idea-generation

Triage public equity idea candidates into research priorities with routing recommendations.

Official
Advanced
DrugClawDrugClaw
126

pharma-db-tools

Query public drug discovery and regulatory databases via Python HTTP APIs.

Official
Intermediate
LogauaEngstromLogauaEngstrom
18

fda-database

Query openFDA datasets for drug, device, food, and substance safety information.

Community
Advanced