deepchem

Molecular ML with diverse featurizers and pre-built datasets. Use for property prediction (ADMET, toxicity) with traditional ML or GNNs when you want extensive featurization options and MoleculeNet benchmarks. Best for quick experiments with pre-trained models, diverse molecular representations. For graph-first PyTorch workflows use torchdrug; for benchmark datasets use pytdc.

Quality

86%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Quality

Content

72%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A highly actionable, well-structured skill body that leans on executable code and correctly offloads deep API/workflow material to verified reference files. Its main weakness is verbosity from repeated patterns and missing inline validation checkpoints in the workflows.

Suggestions

Consolidate the scaffold-splitting guidance (sections 3, Workflow B, Pattern 1, Pitfall 1) into a single canonical treatment and cross-reference it to remove repetition.

Add explicit validation/feedback checkpoints inside Workflows A–C (e.g., 'check dataset loaded before training', 'if ROC-AUC < baseline, increase epochs/try AttentiveFP') rather than only a final evaluate step.

Trim the Overview paragraph's definitional framing ('DeepChem is a comprehensive Python library...') since Claude already knows what the library is.

Dimension	Reasoning	Score
Conciseness	Mostly code-driven and efficient, but the ~590-line body repeats scaffold splitting across sections 3, Workflow B, Pattern 1, and Pitfall 1, and opens with a 'what DeepChem is' overview that could be trimmed.	2 / 3
Actionability	Dense with executable, copy-paste-ready code (loaders, featurizers, models, MoleculeNet, transfer learning), concrete CLI invocations for the bundled scripts, and a featurizer decision tree — fully actionable.	3 / 3
Workflow Clarity	Workflows A/B/C are clearly numbered and end with evaluation, but they lack explicit in-line validation checkpoints or fix-and-retry feedback loops; only the separate 'Common Pitfalls' section provides error recovery.	2 / 3
Progressive Disclosure	Clear overview body with well-signaled, one-level-deep references to real bundle files (references/api_reference.md, references/workflows.md, scripts/*.py), each with 'When to reference' guidance; deep material is appropriately offloaded.	3 / 3
	Total	10 / 12 Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, well-scoped description that states concrete capabilities, gives explicit use-when triggers, and distinguishes itself from adjacent skills by routing edge cases to alternatives. It is concise and free of vague fluff.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete capabilities — 'property prediction (ADMET, toxicity)', 'traditional ML or GNNs', 'MoleculeNet benchmarks', 'pre-trained models', 'diverse featurizers' — matching the anchor for multiple specific concrete actions.	3 / 3
Completeness	Explicitly answers both what ('Molecular ML with diverse featurizers and pre-built datasets') and when ('Use for property prediction... when you want extensive featurization options'), satisfying the explicit-trigger requirement.	3 / 3
Trigger Term Quality	Includes natural domain terms users would say — 'property prediction', 'ADMET', 'toxicity', 'GNNs', 'MoleculeNet', 'molecular' — with good coverage of common variations rather than opaque jargon.	3 / 3
Distinctiveness Conflict Risk	Carves a clear niche and explicitly steers away from neighbors ('For graph-first PyTorch workflows use torchdrug; for benchmark datasets use pytdc'), making wrong-skill triggering unlikely.	3 / 3
	Total	12 / 12 Passed

Validation

81%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (601 lines); consider splitting into references/ and linking	Warning
metadata_version	'metadata.version' is missing	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/chemistry/deepchem/SKILL.md
Commit: 52845c3

Reviewed: about 10 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.