Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Generates a predictive pharmacology profile for a drug candidate from a SMILES string. Covers drug-likeness scoring (Lipinski Rule of Five, Veber rules, QED, SA Score), eight ADME properties, and PAINS interference alerts. All predictions use RDKit descriptors and rule-based heuristics.
Returns eight ADME predictions plus drug-likeness scores and PAINS alerts in a single structured JSON call, chaining directly from chemistry-query without manual tool switching.
When to use it
- Checking Lipinski compliance before committing to synthesis
- Estimating BBB permeability for a CNS drug candidate
- Screening a compound for PAINS interference before running assays
- Flagging CYP3A4 inhibition risk during lead optimization
- Getting a full ADME snapshot from a SMILES string in one step
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: What is the ADME profile of aspirin?
- 1Receives SMILES string (CC(=O)Oc1ccccc1C(=O)O) and computes RDKit molecular descriptors
- 2Evaluates Lipinski Rule of Five and Veber oral bioavailability rules
- 3Calculates QED and SA Score
- 4Runs ADME predictions for BBB permeability, aqueous solubility, GI absorption, CYP3A4 inhibition risk, P-gp substrate likelihood, and plasma protein binding
- 5Checks for PAINS alerts and assembles risk flags into the final report
JSON report: Lipinski pass (0 violations), high GI absorption, moderate solubility, low CYP3A4 risk, QED 0.55, SA Score 1.58, no PAINS alerts, recommended next agents: toxicology, ip-expansion