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

83%

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

Quality

Content

85%

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 with clear sequenced workflows and excellent progressive disclosure pointing to verified bundle files. Its main weakness is conciseness — the body is long with some prose padding and redundancy that could be tightened.

Suggestions

Trim promotional/contextual prose (e.g. 'represents the state-of-the-art... crucial for...') and consolidate the duplicate confidence-score table, keeping the detailed version only in references/confidence_and_limitations.md.

Reduce the Resources section's per-file bullet re-descriptions since the scripts and references are already described where they are used inline; keep it as a concise file index instead.

Dimension	Reasoning	Score
Conciseness	Mostly efficient with concrete commands and code, but ~570 lines include prose padding ('represents the state-of-the-art in computational docking, crucial for...'), a duplicated confidence table, and a Resources section that re-describes scripts already covered inline; not a 3 due to this redundancy and verbosity, not a 1 because it largely assumes domain competence and avoids explaining basics.	2 / 3
Actionability	Provides fully executable, copy-paste-ready guidance throughout — the Modal Python wrapper, `python -m inference` bash commands, batch CSV format, `analyze_results.py` invocations, and GNINA scoring loop — matching the score-3 anchor.	3 / 3
Workflow Clarity	Workflows 1–3 are clearly sequenced with explicit validation checkpoints (`setup_check.py`, `prepare_batch_csv.py --validate`, confidence-threshold filtering) and troubleshooting feedback loops; the batch-operation validation requirement is satisfied, so it is not capped at 2.	3 / 3
Progressive Disclosure	The body is an overview that clearly signals one-level-deep references ('Read `references/parameters_reference.md` using the Read tool'), with content appropriately split across verified-real references/scripts/assets and easy navigation; not a 2 because references are well-signaled and shallow rather than nested or poorly organized.	3 / 3
	Total	11 / 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 strong, specific description with good natural trigger terms and clear distinctiveness, including an explicit out-of-scope note. Its main weakness is the absence of an explicit 'Use when...' trigger clause, which caps completeness at 2.

Suggestions

Add an explicit trigger clause such as 'Use when the user asks to dock a ligand to a protein, predict a binding pose, run virtual screening, or perform structure-based drug design' to raise completeness to 3.

Expand natural trigger variations (e.g. 'dock this molecule', 'where does this ligand bind', 'protein-ligand docking') to broaden the terms users would naturally say.

Dimension	Reasoning	Score
Specificity	Quotes 'Predict protein-ligand binding poses from PDB/SMILES, confidence scores, virtual screening' — multiple specific concrete actions, matching the score-3 anchor of listing several concrete actions; not a 2 because actions are comprehensive rather than partial.	3 / 3
Completeness	Clearly states what ('predict protein-ligand binding poses... confidence scores, virtual screening') but lacks an explicit 'Use when...' trigger clause; 'for structure-based drug design' is a domain context, not trigger guidance, so per the guidelines completeness is capped at 2.	2 / 3
Trigger Term Quality	Natural domain terms a user would say — 'molecular docking', 'protein-ligand', 'binding poses', 'virtual screening', 'PDB/SMILES' — give good coverage; not a 2 because common variations are present rather than missing.	3 / 3
Distinctiveness Conflict Risk	'Diffusion-based molecular docking' is a clear niche and 'Not for affinity prediction' actively disambiguates from affinity-scoring skills, matching the score-3 anchor of a distinct niche unlikely to conflict.	3 / 3
	Total	11 / 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 (586 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/diffdock/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.