AI4Cosmetics
I built the QSAR Cosmetics Skill for Claude & Claude Science to accelerate the safety assessment of cosmetic ingredients.
Core trick
Fuses over 80 QSAR toxicity predictions through Bayesian integration over a curated ingredient database, returning per-endpoint hazard probabilities with structural alerts.
Description
I built the QSAR Cosmetics Skill for Claude & Claude Science to accelerate the safety assessment of cosmetic ingredients. The solution combines a curated toxicology database of nearly 5,000 chemical ingredients with an ontology linking adverse effects, tissues, organs, and biological systems, and integrates over 80 machine learning and deep learning toxicity predictions into a probabilistic Bayesian framework.
The skill enables toxicology teams to rapidly screen and prioritise ingredients, generate evidence-based risk reports, identify similar compounds when data are unavailable, and support early decision-making during product development. The workflow and functionality were designed based on interviews with industry toxicologists to address real-world R&D needs.
Early feedback from toxicologists in industry was highly positive. They highlighted the value of the Bayesian integration approach, the transparent data provenance provided by Claude Science, and the potential to support organisations without dedicated in-house toxicology expertise. Beyond cosmetics, reviewers also identified broader applications in chemical safety assessment.
This work matters because it can significantly reduce the time required for early safety screening, helping R&D teams prioritise experimental resources, identify potential safety concerns sooner, and make more informed decisions while maintaining transparency and scientific traceability.
README excerpt
# QSAR Cosmetics Toxicity Explorer **An in-silico toxicity screening assistant for regulatory toxicologists.** Ask about any cosmetic ingredient and get calibrated, per–adverse-effect hazard probabilities — fused from three established QSAR platforms — with the exact structural alerts and model documentation behind every prediction. Built for use as a **Claude Skill** (Claude / Claude Science): you ask in plain language, it runs the analysis and returns figures, a triage report, and the evidence trail.  --- ## How to use it — in Claude (no coding needed) This repository **is a Claude Skill** — its instructions live in [`SKILL.md`](SKILL.md). You don't run any software; you just talk to Claude. 1. **Add the skill to Claude.** In **Claude Science** or **Claude**, upload the packaged skill file — download **[`qsar-cosmetics-skill.skill`](https://github.com/nicospinu/qsar-cosmetics-skill/releases/latest/download/qsar-cosmetics-skill.skill)** from this repository's [latest release](https://github.com/nicospinu/qsar-cosmetics-skill/releases/latest). It's a single self-contained bundle (`SKILL.md`, scripts, and all data), so it's the easiest way to install: just drop that one file into your Claude / Claude Science Skills workspace. Claude reads `SKILL.md` and loads everything automatically. 2. **Start chatting.** On first use the skill welcomes you and shows the *toxicity landscape* of all