diffdock
Diffusion-based molecular docking. Predict protein-ligand binding poses from PDB/SMILES, confidence scores, virtual screening, for structure-based drug design. Not for affinity prediction.
76
66%
Does it follow best practices?
Impact
94%
1.54xAverage score across 3 eval scenarios
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./scientific-skills/diffdock/SKILL.mdQuality
Discovery
82%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong, domain-specific description with excellent specificity and trigger term coverage for its target audience of computational chemists and drug designers. The explicit exclusion ('Not for affinity prediction') is a nice touch for disambiguation. The main weakness is the absence of an explicit 'Use when...' clause, which would help Claude know exactly when to select this skill.
Suggestions
Add an explicit 'Use when...' clause, e.g., 'Use when the user asks about docking ligands to proteins, predicting binding poses, or performing virtual screening with molecular structures.'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: predict protein-ligand binding poses, confidence scores, virtual screening. Also specifies input formats (PDB/SMILES) and the broader domain (structure-based drug design), plus an explicit exclusion (not for affinity prediction). | 3 / 3 |
Completeness | Clearly answers 'what does this do' (diffusion-based molecular docking, predict binding poses, virtual screening) and includes a negative boundary ('Not for affinity prediction'), but lacks an explicit 'Use when...' clause or equivalent trigger guidance, which caps this at 2 per the rubric guidelines. | 2 / 3 |
Trigger Term Quality | Includes strong natural keywords a computational chemist or drug designer would use: 'molecular docking', 'protein-ligand', 'binding poses', 'PDB', 'SMILES', 'virtual screening', 'structure-based drug design', 'confidence scores'. These are the exact terms domain users would naturally say. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche: diffusion-based molecular docking is a very specific computational chemistry task. The explicit exclusion of affinity prediction further sharpens the boundary. Unlikely to conflict with other skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
50%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill is highly actionable with executable commands and clear examples for all major DiffDock workflows, which is its primary strength. However, it is significantly over-verbose, including unnecessary explanations, extensive best practices lists, citation blocks, and detailed resource descriptions that consume tokens without proportional benefit. The workflow clarity would benefit from explicit validation checkpoints after docking runs, and much of the inline content should be moved to reference files for better progressive disclosure.
Suggestions
Cut the 'When to Use This Skill' section entirely—Claude can infer when to use docking from the overview. Remove the 'Overview' paragraph explaining what molecular docking is, the citations section, and the 'Additional Resources' links to reduce token count by ~30%.
Move troubleshooting, advanced techniques (ensemble docking, scoring integration), and the detailed resource/script descriptions into reference files, keeping only a brief pointer in the main skill.
Add an explicit validation checkpoint after docking completes in the core workflows, e.g., 'Verify output: check that rank_1.sdf exists and confidence_scores.txt contains expected number of entries before proceeding to analysis.'
Consolidate the 10-item best practices list into the 3-4 most critical items that Claude wouldn't already know, removing generic advice like 'validate results experimentally when possible.'
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose at ~400+ lines. It explains concepts Claude would know (what molecular docking is, what PDB files are), includes a 'When to Use This Skill' section that restates obvious triggers, provides extensive best practices lists, citation formatting, and resource descriptions that inflate token count significantly without adding proportional value. | 1 / 3 |
Actionability | The skill provides fully executable commands for all workflows (single docking, batch processing, virtual screening), complete with specific CLI flags, CSV format specifications, Python code for ensemble docking, and concrete bash commands for integration with scoring functions like GNINA. | 3 / 3 |
Workflow Clarity | Multi-step workflows are clearly sequenced (batch processing has Step 1/Step 2, virtual screening has pre-compute then run), but there are no explicit validation checkpoints or error-recovery feedback loops within the core docking workflows. The batch CSV validation is a separate pre-step, but there's no 'validate output before proceeding' checkpoint after docking completes. | 2 / 3 |
Progressive Disclosure | References to external files exist (references/confidence_and_limitations.md, references/parameters_reference.md, references/workflows_examples.md) and are clearly signaled, but the main SKILL.md contains enormous amounts of inline content that could be offloaded—the troubleshooting section, advanced techniques, detailed parameter customization, and extensive resource descriptions all bloat the overview when they could live in reference files. | 2 / 3 |
Total | 8 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
metadata_version | 'metadata.version' is missing | Warning |
Total | 10 / 11 Passed | |
- Repository
- K-Dense-AI/claude-scientific-skills
- Commit
b58ad7e
- 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.