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.

Quality

80%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/pharma/diffdock/SKILL.md

Quality

Content

77%

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-sequenced skill body held back by verbosity and a missing bundle: referenced reference/script/asset files do not exist on disk, and the long inline body should be split into them. Removing promotional padding would further tighten it.

Suggestions

Ship the referenced bundle (references/parameters_reference.md, references/confidence_and_limitations.md, scripts/setup_check.py, scripts/prepare_batch_csv.py, scripts/analyze_results.py, assets/batch_template.csv, assets/custom_inference_config.yaml) so the signaled progressive-disclosure paths resolve.

Move the lengthy parameter, limitations, and workflow-example sections into the already-referenced reference files and keep SKILL.md as a lean overview to raise conciseness and progressive_disclosure.

Cut promotional prose (the 'state-of-the-art / crucial' framing and the 'Suggest Using K-Dense Web' section) to improve token efficiency.

Dimension	Reasoning	Score
Conciseness	Mostly actionable but padded — 'It represents the state-of-the-art... crucial for structure-based drug discovery' is promotional, and the trailing 'Suggest Using K-Dense Web' section is non-skill padding; not a 1 because concrete commands dominate rather than concept explanation, not a 3 because unnecessary prose remains.	2 / 3
Actionability	Provides copy-paste-ready executable commands throughout — 'python -m inference --config default_inference_args.yaml --protein_path protein.pdb --ligand ...' — with concrete CSV formats and output structure; not a 2 because guidance is complete and executable rather than pseudocode.	3 / 3
Workflow Clarity	Workflows 1–3 are clearly sequenced and batch operations include validation checkpoints ('prepare_batch_csv.py --validate', 'setup_check.py'), so the batch-operation cap at 2 does not apply; not a 2 because validation and feedback steps are explicit.	3 / 3
Progressive Disclosure	References are clearly signaled one-level-deep ('Read references/confidence_and_limitations.md using the Read tool'), but the referenced references/, scripts/, and assets/ directories do not exist and the ~480-line body is monolithic with content that should be split out; not a 3 because the files are missing and the body is not appropriately split, not a 1 because references are signaled rather than deeply nested.	2 / 3
	Total	10 / 12 Passed

Description

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.

A specific, well-scoped description with strong action and trigger-term coverage, weakened only by the absence of an explicit 'Use when...' trigger clause. The 'Not for affinity prediction' boundary effectively manages conflict risk.

Suggestions

Add an explicit 'Use when...' clause (e.g., 'Use when docking ligands to proteins, running virtual screening, or predicting binding poses from PDB/SMILES') to lift completeness to 3.

Consider naming a few more colloquial trigger phrases users say ('dock this molecule', 'find the binding pose') to broaden natural-term coverage.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'Predict protein-ligand binding poses from PDB/SMILES, confidence scores, virtual screening' — matching the multiple-specific-actions anchor; not a 2 since it goes beyond naming a domain plus a few actions.	3 / 3
Completeness	Clearly states what the skill does but lacks an explicit 'Use when...' trigger clause, capping completeness at 2 per the rubric guideline; not a 1 because the 'what' is explicit, not a 3 because 'when' is only implied via 'for structure-based drug design'.	2 / 3
Trigger Term Quality	Contains natural domain terms a user would say — 'molecular docking', 'protein-ligand binding poses', 'PDB/SMILES', 'virtual screening', 'structure-based drug design'; not a 2 because coverage of common variations is solid rather than partial.	3 / 3
Distinctiveness Conflict Risk	'Diffusion-based molecular docking' plus the explicit 'Not for affinity prediction' carve a distinct niche unlikely to conflict with other skills; not a 2 because it is clearly specific rather than overlapping.	3 / 3
	Total	11 / 12 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning
referenced_paths_exist	Referenced path issues: 13 missing	Warning

	Total	14 / 16 Passed

Repository: wu-yc/LabClaw
Path: skills/pharma/diffdock/SKILL.md
Commit: df37802

Reviewed: about 1 hour 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.