CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

rdkit

Cheminformatics toolkit for fine-grained molecular control. SMILES/SDF parsing, descriptors (MW, LogP, TPSA), fingerprints, substructure search, 2D/3D generation, similarity, reactions. For standard workflows with simpler interface, use datamol (wrapper around RDKit). Use rdkit for advanced control, custom sanitization, specialized algorithms.

Quality

82%

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

72%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.

Well-structured content with excellent progressive disclosure and strong embedded validation guidance, kept lean by deferring worked code to references and scripts. The main weakness is that the body never sequences a multi-step workflow inline (validation knowledge is scattered across pitfalls), and it carries inline dated version material that would sit better in a dedicated version/legacy section.

Suggestions

Add a short inline quick-start workflow (parse SMILES → check for None → sanitize/compute descriptor) with explicit validation checkpoints so the most common path doesn't require opening a reference file.

Move the dated baseline material ("reviewed 2026-10-01", RDKit 2026.03.6, rdkit-pypi legacy note) into a dedicated version/legacy section rather than the Overview, tightening the body and isolating time-sensitive details.

Condense the Overview paragraph, which restates the frontmatter description nearly verbatim.

DimensionReasoningScore

Conciseness

The body is dense and assumes competence (capability table, terse pitfalls, pointer sections), but the inline "Tested stable baseline (reviewed 2026-10-01): RDKit 2026.03.6" paragraph embeds time-sensitive version/date detail outside a deprecated-style section, and the Overview partly restates the frontmatter description. These are minor trimmable instances (anchor 4), not pervasive padding (anchor 3).

4 / 5

Actionability

Copy-paste install commands ("uv pip install \"rdkit==2026.3.6\"", conda create) and concrete API-level pitfalls ("DetectChemistryProblems()", "EmbedParameters.trackFailures", "GetFailureCounts()", conformer ID -1) give mostly executable guidance; the gap is that no inline quick-start RDKit snippet covers the most common case (parse → validate → compute), deferring it entirely to references.

4 / 5

Workflow Clarity

Validation checkpoints are explicitly embedded ("Always validate molecules after parsing", "-1 means embedding failed", "do not send a failed embedding into force-field optimization"), so the batch-operation cap for missing validation does not apply. But the body is organized topically rather than as a sequenced workflow — multi-step sequences are deferred to references/workflows_and_best_practices.md — fitting anchor 3 (checkpoints present, explicit sequence absent) rather than 4.

3 / 5

Progressive Disclosure

The body is a genuine overview: a twelve-area capability table routing to references/core_capabilities.md, well-signaled one-level-deep references (core_capabilities, workflows_and_best_practices, api_reference, descriptors_reference, smarts_patterns, review — all verified present), scripts listed with purposes, and explicit local-file-vs-installable-package disambiguation. This matches the anchor-5 clear-overview pattern.

5 / 5

Total

16

/

20

Passed

Description

83%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: specific, comprehensive capability list in third person with explicit use-guidance and clear differentiation from the datamol wrapper. The only weakness is trigger-term breadth — a few natural synonyms and file extensions users would say are absent.

DimensionReasoningScore

Specificity

"SMILES/SDF parsing, descriptors (MW, LogP, TPSA), fingerprints, substructure search, 2D/3D generation, similarity, reactions" lists seven concrete capabilities with named descriptors, matching the comprehensive multi-action anchor; not 4, since coverage gaps are negligible. Voice is third person throughout, so no person-penalty applies.

5 / 5

Completeness

"What" is explicit ("Cheminformatics toolkit... SMILES/SDF parsing, descriptors...") and "when" guidance exists ("Use rdkit for advanced control, custom sanitization, specialized algorithms"), so it is not capped at 3. However the trigger guidance is framed as datamol disambiguation rather than concrete trigger phrases, fitting anchor 4 rather than 5.

4 / 5

Trigger Term Quality

Good natural keyword coverage (SMILES, SDF, descriptors, fingerprints, substructure search, similarity, reactions), but common user variations like "molecules", "chemical structures", "Tanimoto/ECFP", or file extensions such as ".smi" are missing, which is the anchor-4 pattern rather than comprehensive anchor 5.

4 / 5

Distinctiveness Conflict Risk

It explicitly differentiates from the closest sibling skill ("For standard workflows with simpler interface, use datamol (wrapper around RDKit). Use rdkit for advanced control..."), establishing a clear niche with distinct triggers and minimal conflict risk, matching the anchor-5 example's pattern.

5 / 5

Total

18

/

20

Passed

Validation

100%

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

K-Dense-AI/scientific-agent-skills

Path

skills/rdkit/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.