molfeat
Build Molecular ML Features with Molfeat
Molecular datasets need consistent numerical representations before modeling. This skill guides Molfeat selection, configuration, featurization, caching, and integration with common machine-learning workflows.
Stop for confirmation before installing.
Review the plan and obtain explicit user consent before changing files.
Install with my Agent
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Review the Skillstore skill "molfeat" from https://skillstore.io/skills/davila7-molfeat.md and its manifest at https://skillstore.io/api/skills/davila7-molfeat/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.Your Agent should still show its plan and request any confirmation required by the security policy.
Agent-readable resources
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Test it
Using "molfeat". Which featurizer should I try first for a small QSAR regression dataset?
Expected outcome:
- Start with ECFP at radius 2 or 3 and 2,048 bits as a strong, efficient baseline.
- Compare it with RDKit 2D descriptors when interpretability matters.
- Use scaffold-aware validation and save the exact featurizer configuration.
Using "molfeat". How should I process one million SMILES without exhausting memory?
Expected outcome:
- Validate inputs and process fixed-size chunks instead of creating one full feature matrix.
- Write each completed chunk to a safe array format and record failed molecule indices.
- Benchmark worker count because maximum parallelism can increase memory pressure.
Using "molfeat". Should I use ChemBERTa or ECFP for an initial experiment?
Expected outcome:
Use ECFP for the first baseline because it is fast and inexpensive. Add ChemBERTa when transfer learning may justify extra dependencies and compute.
Security Audit
High RiskAll 96 static findings are false positives caused by Markdown syntax, benign local cache writes, molecule-validation language, and passive documentation links. However, two examples use pickle.load without authenticity checks, creating a high-severity code-execution risk if a cache file is replaced or untrusted.
Confirmed security concerns (1)
Risk Factors
📁 Filesystem access (2)
⚙️ External commands (50)
🌐 Network access (4)
Share & cite this report
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
Copy report link
https://skillstore.io/skills/davila7-molfeat/audits/10?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown badge
[](https://skillstore.io/skills/davila7-molfeat?utm_source=security_passport_badge)HTML badge
<a href="https://skillstore.io/skills/davila7-molfeat?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/davila7-molfeat/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Embed card
<iframe src="https://skillstore.io/embed/skills/davila7-molfeat.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>Academic citations (APA · BibTeX · CFF)
APA citation
davila7. (2026). molfeat security audit report (audit version 10) [Author version unspecified]. Skillstore. https://skillstore.io/skills/davila7-molfeat/audits/10BibTeX citation
@techreport{davila7-davila7-molfeat-2026,
author = {davila7},
title = {molfeat security audit report (audit version 10)},
institution = {Skillstore},
year = {2026},
number = {10},
url = {https://skillstore.io/skills/davila7-molfeat/audits/10},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "molfeat security audit report (audit version 10)"
version: "unspecified"
type: report
authors:
- name: "davila7"
date-released: "2026-07-23"
url: "https://skillstore.io/skills/davila7-molfeat/audits/10"
identifiers:
- type: other
value: "skillstore:davila7-molfeat:audit:10"
description: "Skillstore immutable audit report identifier"
Compare variants
2 installable variantsEach author remains a separate installable skill. The recommended variant is ranked by Skillstore evidence.
Why this variant is first
k-dense-ai-molfeat
2026-08-21
davila7-molfeat
2026-08-21
Skillstore Score
Why this score Evidence Confidence: MediumWhat You Can Build
Prepare QSAR Features
Choose reproducible fingerprints or descriptors and connect them to a scikit-learn property prediction pipeline.
Screen Compound Libraries
Batch-featurize candidate molecules, score them with a classifier, and rank likely active compounds.
Prototype Molecular Embeddings
Compare ChemBERTa, graph models, and classical fingerprints before committing to an architecture.
Try These Prompts
Recommend a Molfeat featurizer for [task] using [dataset size] molecules and [model type]. Explain accuracy, speed, dependencies, and feature dimensions.
Design a Molfeat workflow for SMILES data in [input format]. Include validation, error handling, batching, parallel settings, and reproducible configuration.
Compare [featurizer A], [featurizer B], and [featurizer C] for [prediction task]. Define a fair evaluation plan and resource estimates.
Design a production Molfeat pipeline for [volume] molecules. Cover chunking, trusted caching, configuration versioning, monitoring, failure handling, and deployment constraints.
Best Practices
- Validate SMILES, preserve row mappings, and report failed transformations before model training.
- Save featurizer configuration, package versions, preprocessing rules, and random seeds with every experiment.
- Benchmark simple fingerprints first, then add pretrained embeddings only when measured performance supports their cost.
Avoid
- Do not compare featurizers with different data splits or inconsistent preprocessing.
- Do not load pickle caches from shared, downloaded, or otherwise untrusted locations.
- Do not use maximum parallelism without measuring memory consumption and throughput.
Frequently Asked Questions
What input formats does this skill cover?
Which featurizer is a practical starting point?
Does the skill install Molfeat automatically?
Can it use pretrained molecular models?
How are invalid molecules handled?
Is the documented pickle cache safe?
Developer Details
Author
davila7License
MIT
Skillstore revision
r2
Version notice
The author did not declare a version.
Ref
c43861a65bb95efcae259cd161c9d6f4dc7eec6f
Maintenance freshness
7/24/2026
Usage
6 downloads · 209 views
File structure