molecular-optimization

A step-by-step method for improving drug candidate molecules while checking that proposed chemical structures are valid. It uses repeated analysis, candidate generation, checking, evaluation, and ranking.

In plain words
What is it for?
It is for lead optimization, finding alternative molecular scaffolds, and designing molecule variants aimed at specific property changes such as lower fat solubility or reduced hERG risk.
Why use it?
Drug discovery often requires improving several properties at once without losing the useful effects of a molecule. It helps reduce invalid or unsuitable molecule suggestions.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/synthetic-sciences/openscience/molecular-optimization
Any agent
npx skills add synthetic-sciences/openscience --skill molecular-optimization
Clone the repo
git clone --depth 1 https://github.com/synthetic-sciences/openscience

Made for: Claude Code, Codex.

Per session 34 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,389 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00034 $0.01389
Opus 5 $0.00017 $0.00694
Sonnet 5 $0.00007 $0.00278
Haiku 4.5 $0.00003 $0.00139

Measured 2d ago against content hash 28af9bc27fb4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

molecular-optimization scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

The scan reads SKILL.md. This mod also ships 3 executable files (scripts/compare_candidates.py, scripts/optimize.py, scripts/verify_smiles.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

backend/cli/skills/chemistry/molecular-optimization/SKILL.md · 161 lines

How it starts

The opening of the file, as written. The whole thing — 161 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Molecular Optimization

Overview

Lead optimization is the bottleneck of drug discovery — modifying a hit compound to improve potency, selectivity, and ADMET properties without breaking what already works. LLMs frequently generate invalid SMILES or propose modifications that don't appear in the actual structure.

This skill implements an iterative optimization protocol based on three peer-reviewed approaches:

  • MT-Mol (Kim et al., 2025): Multi-agent tool-based reasoning with verification — SOTA on 17/23 PMO benchmark tasks
  • DrugR (Liu et al., 2026): Explicit liability reasoning before generation — 18× improvement over blind generation
  • MultiMol (Yu et al., 2025): Generate-then-rank with scaffold preservation — 82.3% multi-objective success rate

The core loop: Analyze → Identify Liabilities → Generate Candidates → Verify → Evaluate & Rank → Iterate.

When to Use This Skill

  • Lead optimization: Improve ADMET properties of a hit while preserving potency
  • Scaffold hopping: Find new scaffolds that maintain key pharmacophoric features
  • Property-driven design: Generate analogs targeting specific property improvements (lower LogP, reduce hERG, improve solubility)
  • Multi-objective optimization: Balance multiple properties simultaneously

Do NOT use this skill for:

  • De novo design from scratch (use denovo-design instead)
  • Simple property prediction without optimization (use admet-prediction)
  • Docking or binding affinity estimation (use molecular-docking, binding-affinity)

Related Skills

  • admet-prediction: Compute ADMET properties (this skill uses it internally)
  • admet-reasoning: Interpretable ADMET analysis with mechanistic explanations
  • smiles-validation: Strict SMILES parsing and structural verification
  • rdkit: Core molecular operations
  • medchem: Medicinal chemistry filters and transformations

Installation

Required dependencies

pip install rdkit-pypi numpy pandas

Read the full file on GitHub · 161 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 161 lines · 34 tokens per session scan A 28af9bc27fb4

Subscribe to this mod's changes

molecular-optimization is a skill published in the GitHub repository synthetic-sciences/openscience (3,385 stars, last pushed today), licensed Apache-2.0. It adds 34 tokens to every session and 1,389 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

drug-lead-analysis

Analyze drug candidate molecules for drug-likeness, ADMET properties, and safety profiles. Use this skill when: (1) Evaluating a molecule's potential as a drug candidate, (2) Checking drug-likeness scores (QED, Lipinski), (3) Predicting blood-brain barrier penetration, (4) Assessing side effects and ADMET properties…

PharMolix/OpenBioMed · 91 tokens

admet_genetic

ADMET-guided genetic molecule optimization workflow from seed SMILES; use when the agent needs to build or run an RDKit/SA-Score/ADMET-AI GA pipeline for molecule optimization, enforce molecule lineage logs, render optimization-history HTML dashboards, and write candidate triage reports.

PKU-YuanGroup/OpenAI4S · 63 tokens

target-based-lead-design

Generate diverse lead compounds for a specific protein target using structure-based drug design with MolCraft. Use this skill when: (1) Designing drug candidates for a known protein target (PDB ID or disease name), (2) Generating structurally diverse molecules with optimized binding affinity, (3) Filtering candidates…

PharMolix/OpenBioMed · 99 tokens

admet-prediction

Predict comprehensive ADMET (Absorption, Distribution, Metabolism, Excretion, Toxicity) properties for drug candidate molecules using GraphMVP ensemble models. Use this skill when: (1) Predicting blood-brain barrier penetration, (2) Assessing side effect profiles, (3) Estimating Caco-2 permeability, half-life, or LD50…

PharMolix/OpenBioMed · 96 tokens

chem-spectrum-matcher

Match an experimental spectrum (1H NMR, 13C NMR, IR) against predicted or database reference spectra for candidate ranking and structure confirmation. Supports local catalog lookup, public database fallback, and pluggable similarity metrics.

learningmatter-mit/AtomisticSkills · 52 tokens

chem-msms-predict

Predict LC-MS/MS (MS2, tandem mass spectra) from SMILES via ICEBERG, a two-stage deep neural network. Outputs predicted m/z vs intensity spectrum, fragment ion SMILES, and a spectrum plot.

learningmatter-mit/AtomisticSkills · 51 tokens