Therapeutics Data Commons. AI-ready drug discovery datasets (ADME, toxicity, DTI), benchmarks, scaffold splits, molecular oracles, for therapeutic ML and pharmacological prediction.
8.3
Rating
0
Installs
Machine Learning
Category
Excellent drug discovery skill providing comprehensive access to TDC's therapeutic datasets, benchmarks, and molecular oracles. The Description clearly conveys the skill's scope (ADME, toxicity, DTI, scaffolds, oracles) enabling proper invocation. Task knowledge is outstanding with complete examples for single/multi-instance prediction, generation tasks, benchmark evaluation, and proper data splitting strategies. Structure is well-organized with clear category breakdowns and appropriate delegation of detailed documentation to referenced files (datasets.md, oracles.md, utilities.md). Novelty is high: accessing curated drug discovery datasets with standardized splits, running proper 5-seed benchmarks, and using molecular oracles for optimization would require extensive setup and domain expertise from a CLI agent alone. This skill meaningfully reduces complexity and token costs for pharmaceutical ML workflows. Minor improvement possible: could add more explicit cross-references between SKILL.md sections and the scripts folder.
Loading SKILL.md…