CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

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

83%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

75%Weight 40%Scale 1-5

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

The body is highly actionable, with executable code, CLI commands, and decision aids throughout, and the bundle is well structured with real, clearly signaled one-level-deep references. Its weakness is token efficiency: the 606-line body duplicates several sections (scaffold splitting, installation, workflow examples) that should be consolidated or moved into the reference files.

Suggestions

Consolidate the three scaffold-splitting treatments (section 3 'Critical' note, Pattern 1, Workflow B) into a single authoritative statement and reference it from the other spots.

Remove the 'Installation Notes' section, which duplicates Pitfall Issue 4 verbatim, and trim the generic overview paragraph and 'Additional Resources' list, which restate knowledge Claude already has.

Move the model-selection table, transfer-learning section, and Patterns/Pitfalls detail into references/workflows.md (or a new reference file), keeping SKILL.md as a lean overview with pointers.

DimensionReasoningScore

Conciseness

Mostly efficient — the featurizer decision tree, model selection table, and pitfalls are operational knowledge Claude does not already know — but tightening is needed: scaffold-splitting guidance appears three times (section 3 "Critical", Pattern 1, Workflow B), the Typical Workflows restate Core Capabilities examples, "Installation Notes" duplicates Pitfall Issue 4 verbatim, and "Additional Resources" plus the generic overview paragraph add padding. Not a 2 because the padded sections are a minority of the body.

3 / 5

Actionability

Fully executable throughout: copy-paste-ready Python for loaders, splitters, models, evaluation, and prediction, plus concrete CLI invocations with flags for all three bundled scripts, a model-selection table, and a featurizer decision tree covering the common cases. Matches the anchor-5 example of code with specific usage guidance.

5 / 5

Workflow Clarity

Workflows A/B/C are clearly sequenced (numbered steps 1-5 with code per step and a final evaluation step), and the Common Pitfalls section provides error-recovery guidance. Minor validation gaps — no explicit data-sanity or split-verification checkpoints before training — keep it below anchor 5; the workflows are non-destructive so no cap applies.

4 / 5

Progressive Disclosure

Good structure with a verified two-file reference layer (references/api_reference.md and references/workflows.md, both real and one level deep with no nested references) plus documented scripts, each with explicit "When to reference" guidance. The main gap is that a 606-line body keeps model tables, patterns, and pitfalls inline that could move to the reference files, so it does not reach the lean-overview anchor 5.

4 / 5

Total

16

/

20

Passed

Description

87%Weight 40%Scale 1-5

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 description: it answers what and when explicitly, uses natural trigger vocabulary, and uniquely positions the skill against adjacent skills via explicit handoffs. Its only weakness is that capabilities are stated as terse noun phrases rather than concrete action verbs, which slightly blurs specificity.

DimensionReasoningScore

Specificity

Lists several specific capabilities — "property prediction (ADMET, toxicity)", "diverse featurizers", "GNNs", "pre-trained models", "MoleculeNet benchmarks" — but presents them as compressed noun phrases without concrete action verbs, so it falls between the 1-2-action anchor (3) and comprehensive action coverage (5).

4 / 5

Completeness

Clearly answers both: the what ("Molecular ML with diverse featurizers and pre-built datasets") and an explicit when with concrete triggers ("Use for property prediction (ADMET, toxicity) ... when you want extensive featurization options and MoleculeNet benchmarks"), matching the anchor-5 pattern exactly.

5 / 5

Trigger Term Quality

Contains natural phrases users would actually say — "property prediction", "ADMET", "toxicity", "GNN", "pre-trained models" — but misses common variations such as "solubility", "drug discovery", or "molecular fingerprints", placing it between anchors 3 and 5.

4 / 5

Distinctiveness Conflict Risk

Occupies a clear niche (DeepChem-style molecular ML with featurization breadth) and explicitly routes adjacent requests to other skills ("For graph-first PyTorch workflows use torchdrug; for benchmark datasets use pytdc"), minimizing wrong-skill triggering.

5 / 5

Total

18

/

20

Passed

Validation

81%

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

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (607 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

Reviewed

Table of Contents

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.