TacoSkill LAB
TacoSkill LAB
HomeSkillHubCreatePlaygroundSkillKit
© 2026 TacoSkill LAB
AboutPrivacyTerms
  1. Home
  2. /
  3. SkillHub
  4. /
  5. diffdock
Improve

diffdock

8.3

by K-Dense-AI

145Favorites
359Upvotes
0Downvotes

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.

molecular-docking

8.3

Rating

0

Installs

AI & LLM

Category

Quick Review

Exceptional molecular docking skill with comprehensive documentation. The SKILL.md provides crystal-clear invocation guidance with concrete commands for single and batch docking workflows. Task knowledge is excellent - includes setup verification scripts, batch preparation tools, result analysis utilities, and detailed parameter tuning guidance. Structure is very well organized with a logical flow from basics to advanced techniques, with appropriate delegation to reference files for deep-dive topics. The skill addresses a genuinely novel need: molecular docking via deep learning requires specialized environment setup (PyTorch Geometric, ESM, CUDA), domain-specific parameter tuning (temperature sampling, torsion angles), and careful interpretation of confidence scores vs binding affinity - all of which would consume significant tokens if handled by a CLI agent alone. Minor point: while excellently organized, the main SKILL.md is quite comprehensive and could theoretically be slightly more concise, though the current balance between overview and actionable detail is still very strong.

LLM Signals

Description coverage10
Task knowledge10
Structure9
Novelty8

GitHub Signals

6,871
818
49
3
Last commit 1 days ago

Publisher

K-Dense-AI

K-Dense-AI

Skill Author

Related Skills

rag-architectprompt-engineerfine-tuning-expert

Loading SKILL.md…

Try onlineView on GitHub

Publisher

K-Dense-AI avatar
K-Dense-AI

Skill Author

Related Skills

rag-architect

Jeffallan

7.0

prompt-engineer

Jeffallan

7.0

fine-tuning-expert

Jeffallan

6.4

mcp-developer

Jeffallan

6.4
Try online