[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6AFCB4F3RNWA7173A4N4E8J":3,"skill-same:01M6AFCB4F3RNWA7173A4N4E8J":91},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":23,"publishedAt":38,"linkOnly":39,"markdownContent":40,"markdownBody":41,"frontMatter":42,"refName":43,"commitSha":44,"skillDirectory":45,"forksCount":46,"repositoryHtmlUrl":47,"repositoryHomepageUrl":48,"ownerHtmlUrl":49,"ownerBlogUrl":50,"ownerDescription":51,"updatedAt":52,"install":53,"resources":66},"01M6AFCB4F3RNWA7173A4N4E8J","deepchem",null,"Molecular machine learning toolkit. Property prediction (ADMET, toxicity), GNNs (GCN, MPNN), MoleculeNet benchmarks, pretrained models, featurization, for drug discovery ML.","davila7\u002Fclaude-code-templates","davila7","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F6216945?v=4","01M6ADEWWFPS191NFM0Y1EX1JH","Daniel Avila","01M6ADJHGM78PHQRD3XZTFDVKK",33481,"MIT","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fdeepchem\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[5,24,25,26,27,28,29,30,31,32,33,34,35,36,37],"machine learning","ml","toolkit","discovery","learning","models","benchmarks","molecular","drug discovery","machine","for","prediction","ADMET","toxicity","2026-11-10 10:12:30",false,"---\nname: deepchem\ndescription: \"Molecular machine learning toolkit. Property prediction (ADMET, toxicity), GNNs (GCN, MPNN), MoleculeNet benchmarks, pretrained models, featurization, for drug discovery ML.\"\n---\n\n# DeepChem\n\n## Overview\n\nDeepChem is a comprehensive Python library for applying machine learning to chemistry, materials science, and biology. Enable molecular property prediction, drug discovery, materials design, and biomolecule analysis through specialized neural networks, molecular featurization methods, and pretrained models.\n\n## When to Use This Skill\n\nThis skill should be used when:\n- Loading and processing molecular data (SMILES strings, SDF files, protein sequences)\n- Predicting molecular properties (solubility, toxicity, binding affinity, ADMET properties)\n- Training models on chemical\u002Fbiological datasets\n- Using MoleculeNet benchmark datasets (Tox21, BBBP, Delaney, etc.)\n- Converting molecules to ML-ready features (fingerprints, graph representations, descriptors)\n- Implementing graph neural networks for molecules (GCN, GAT, MPNN, AttentiveFP)\n- Applying transfer learning with pretrained models (ChemBERTa, GROVER, MolFormer)\n- Predicting crystal\u002Fmaterials properties (bandgap, formation energy)\n- Analyzing protein or DNA sequences\n\n## Core Capabilities\n\n### 1. Molecular Data Loading and Processing\n\nDeepChem provides specialized loaders for various chemical data formats:\n\n```python\nimport deepchem as dc\n\n# Load CSV with SMILES\nfeaturizer = dc.feat.CircularFingerprint(radius=2, size=2048)\nloader = dc.data.CSVLoader(\n    tasks=['solubility', 'toxicity'],\n    feature_field='smiles',\n    featurizer=featurizer\n)\ndataset = loader.create_dataset('molecules.csv')\n\n# Load SDF files\nloader = dc.data.SDFLoader(tasks=['activity'], featurizer=featurizer)\ndataset = loader.create_dataset('compounds.sdf')\n\n# Load protein sequences\nloader = dc.data.FASTALoader()\ndataset = loader.create_dataset('proteins.fasta')\n```\n\n**Key Loaders**:\n- `CSVLoader`: Tabular data with molecular identifiers\n- `SDFLoader`: Molecular structure files\n- `FASTALoader`: Protein\u002FDNA sequences\n- `ImageLoader`: Molecular images\n- `JsonLoader`: JSON-formatted datasets\n\n### 2. Molecular Featurization\n\nConvert molecules into numerical representations for ML models.\n\n#### Decision Tree for Featurizer Selection\n\n```\nIs the model a graph neural network?\n├─ YES → Use graph featurizers\n│   ├─ Standard GNN → MolGraphConvFeaturizer\n│   ├─ Message passing → DMPNNFeaturizer\n│   └─ Pretrained → GroverFeaturizer\n│\n└─ NO → What type of model?\n    ├─ Traditional ML (RF, XGBoost, SVM)\n    │   ├─ Fast baseline → CircularFingerprint (ECFP)\n    │   ├─ Interpretable → RDKitDescriptors\n    │   └─ Maximum coverage → MordredDescriptors\n    │\n    ├─ Deep learning (non-graph)\n    │   ├─ Dense networks → CircularFingerprint\n    │   └─ CNN → SmilesToImage\n    │\n    ├─ Sequence models (LSTM, Transformer)\n    │   └─ SmilesToSeq\n    │\n    └─ 3D structure analysis\n        └─ CoulombMatrix\n```\n\n#### Example Featurization\n\n```python\n# Fingerprints (for traditional ML)\nfp = dc.feat.CircularFingerprint(radius=2, size=2048)\n\n# Descriptors (for interpretable models)\ndesc = dc.feat.RDKitDescriptors()\n\n# Graph features (for GNNs)\ngraph_feat = dc.feat.MolGraphConvFeaturizer()\n\n# Apply featurization\nfeatures = fp.featurize(['CCO', 'c1ccccc1'])\n```\n\n**Selection Guide**:\n- **Small datasets (\u003C1K)**: CircularFingerprint or RDKitDescriptors\n- **Medium datasets (1K-100K)**: CircularFingerprint or graph featurizers\n- **Large datasets (>100K)**: Graph featurizers (MolGraphConvFeaturizer, DMPNNFeaturizer)\n- **Transfer learning**: Pretrained model featurizers (GroverFeaturizer)\n\nSee `references\u002Fapi_reference.md` for complete featurizer documentation.\n\n### 3. Data Splitting\n\n**Critical**: For drug discovery tasks, use `ScaffoldSplitter` to prevent data leakage from similar molecular structures appearing in both training and test sets.\n\n```python\n# Scaffold splitting (recommended for molecules)\nsplitter = dc.splits.ScaffoldSplitter()\ntrain, valid, test = splitter.train_valid_test_split(\n    dataset,\n    frac_train=0.8,\n    frac_valid=0.1,\n    frac_test=0.1\n)\n\n# Random splitting (for non-molecular data)\nsplitter = dc.splits.RandomSplitter()\ntrain, test = splitter.train_test_split(dataset)\n\n# Stratified splitting (for imbalanced classification)\nsplitter = dc.splits.RandomStratifiedSplitter()\ntrain, test = splitter.train_test_split(dataset)\n```\n\n**Available Splitters**:\n- `ScaffoldSplitter`: Split by molecular scaffolds (prevents leakage)\n- `ButinaSplitter`: Clustering-based molecular splitting\n- `MaxMinSplitter`: Maximize diversity between sets\n- `RandomSplitter`: Random splitting\n- `RandomStratifiedSplitter`: Preserves class distributions\n\n### 4. Model Selection and Training\n\n#### Quick Model Selection Guide\n\n| Dataset Size | Task | Recommended Model | Featurizer |\n|-------------|------|-------------------|------------|\n| \u003C 1K samples | Any | SklearnModel (RandomForest) | CircularFingerprint |\n| 1K-100K | Classification\u002FRegression | GBDTModel or MultitaskRegressor | CircularFingerprint |\n| > 100K | Molecular properties | GCNModel, AttentiveFPModel, DMPNNModel | MolGraphConvFeaturizer |\n| Any (small preferred) | Transfer learning | ChemBERTa, GROVER, MolFormer | Model-specific |\n| Crystal structures | Materials properties | CGCNNModel, MEGNetModel | Structure-based |\n| Protein sequences | Protein properties | ProtBERT | Sequence-based |\n\n#### Example: Traditional ML\n```python\nfrom sklearn.ensemble import RandomForestRegressor\n\n# Wrap scikit-learn model\nsklearn_model = RandomForestRegressor(n_estimators=100)\nmodel = dc.models.SklearnModel(model=sklearn_model)\nmodel.fit(train)\n```\n\n#### Example: Deep Learning\n```python\n# Multitask regressor (for fingerprints)\nmodel = dc.models.MultitaskRegressor(\n    n_tasks=2,\n    n_features=2048,\n    layer_sizes=[1000, 500],\n    dropouts=0.25,\n    learning_rate=0.001\n)\nmodel.fit(train, nb_epoch=50)\n```\n\n#### Example: Graph Neural Networks\n```python\n# Graph Convolutional Network\nmodel = dc.models.GCNModel(\n    n_tasks=1,\n    mode='regression',\n    batch_size=128,\n    learning_rate=0.001\n)\nmodel.fit(train, nb_epoch=50)\n\n# Graph Attention Network\nmodel = dc.models.GATModel(n_tasks=1, mode='classification')\nmodel.fit(train, nb_epoch=50)\n\n# Attentive Fingerprint\nmodel = dc.models.AttentiveFPModel(n_tasks=1, mode='regression')\nmodel.fit(train, nb_epoch=50)\n```\n\n### 5. MoleculeNet Benchmarks\n\nQuick access to 30+ curated benchmark datasets with standardized train\u002Fvalid\u002Ftest splits:\n\n```python\n# Load benchmark dataset\ntasks, datasets, transformers = dc.molnet.load_tox21(\n    featurizer='GraphConv',  # or 'ECFP', 'Weave', 'Raw'\n    splitter='scaffold',     # or 'random', 'stratified'\n    reload=False\n)\ntrain, valid, test = datasets\n\n# Train and evaluate\nmodel = dc.models.GCNModel(n_tasks=len(tasks), mode='classification')\nmodel.fit(train, nb_epoch=50)\n\nmetric = dc.metrics.Metric(dc.metrics.roc_auc_score)\ntest_score = model.evaluate(test, [metric])\n```\n\n**Common Datasets**:\n- **Classification**: `load_tox21()`, `load_bbbp()`, `load_hiv()`, `load_clintox()`\n- **Regression**: `load_delaney()`, `load_freesolv()`, `load_lipo()`\n- **Quantum properties**: `load_qm7()`, `load_qm8()`, `load_qm9()`\n- **Materials**: `load_perovskite()`, `load_bandgap()`, `load_mp_formation_energy()`\n\nSee `references\u002Fapi_reference.md` for complete dataset list.\n\n### 6. Transfer Learning\n\nLeverage pretrained models for improved performance, especially on small datasets:\n\n```python\n# ChemBERTa (BERT pretrained on 77M molecules)\nmodel = dc.models.HuggingFaceModel(\n    model='seyonec\u002FChemBERTa-zinc-base-v1',\n    task='classification',\n    n_tasks=1,\n    learning_rate=2e-5  # Lower LR for fine-tuning\n)\nmodel.fit(train, nb_epoch=10)\n\n# GROVER (graph transformer pretrained on 10M molecules)\nmodel = dc.models.GroverModel(\n    task='regression',\n    n_tasks=1\n)\nmodel.fit(train, nb_epoch=20)\n```\n\n**When to use transfer learning**:\n- Small datasets (\u003C 1000 samples)\n- Novel molecular scaffolds\n- Limited computational resources\n- Need for rapid prototyping\n\nUse the `scripts\u002Ftransfer_learning.py` script for guided transfer learning workflows.\n\n### 7. Model Evaluation\n\n```python\n# Define metrics\nclassification_metrics = [\n    dc.metrics.Metric(dc.metrics.roc_auc_score, name='ROC-AUC'),\n    dc.metrics.Metric(dc.metrics.accuracy_score, name='Accuracy'),\n    dc.metrics.Metric(dc.metrics.f1_score, name='F1')\n]\n\nregression_metrics = [\n    dc.metrics.Metric(dc.metrics.r2_score, name='R²'),\n    dc.metrics.Metric(dc.metrics.mean_absolute_error, name='MAE'),\n    dc.metrics.Metric(dc.metrics.root_mean_squared_error, name='RMSE')\n]\n\n# Evaluate\ntrain_scores = model.evaluate(train, classification_metrics)\ntest_scores = model.evaluate(test, classification_metrics)\n```\n\n### 8. Making Predictions\n\n```python\n# Predict on test set\npredictions = model.predict(test)\n\n# Predict on new molecules\nnew_smiles = ['CCO', 'c1ccccc1', 'CC(C)O']\nnew_features = featurizer.featurize(new_smiles)\nnew_dataset = dc.data.NumpyDataset(X=new_features)\n\n# Apply same transformations as training\nfor transformer in transformers:\n    new_dataset = transformer.transform(new_dataset)\n\npredictions = model.predict(new_dataset)\n```\n\n## Typical Workflows\n\n### Workflow A: Quick Benchmark Evaluation\n\nFor evaluating a model on standard benchmarks:\n\n```python\nimport deepchem as dc\n\n# 1. Load benchmark\ntasks, datasets, _ = dc.molnet.load_bbbp(\n    featurizer='GraphConv',\n    splitter='scaffold'\n)\ntrain, valid, test = datasets\n\n# 2. Train model\nmodel = dc.models.GCNModel(n_tasks=len(tasks), mode='classification')\nmodel.fit(train, nb_epoch=50)\n\n# 3. Evaluate\nmetric = dc.metrics.Metric(dc.metrics.roc_auc_score)\ntest_score = model.evaluate(test, [metric])\nprint(f\"Test ROC-AUC: {test_score}\")\n```\n\n### Workflow B: Custom Data Prediction\n\nFor training on custom molecular datasets:\n\n```python\nimport deepchem as dc\n\n# 1. Load and featurize data\nfeaturizer = dc.feat.CircularFingerprint(radius=2, size=2048)\nloader = dc.data.CSVLoader(\n    tasks=['activity'],\n    feature_field='smiles',\n    featurizer=featurizer\n)\ndataset = loader.create_dataset('my_molecules.csv')\n\n# 2. Split data (use ScaffoldSplitter for molecules!)\nsplitter = dc.splits.ScaffoldSplitter()\ntrain, valid, test = splitter.train_valid_test_split(dataset)\n\n# 3. Normalize (optional but recommended)\ntransformers = [dc.trans.NormalizationTransformer(\n    transform_y=True, dataset=train\n)]\nfor transformer in transformers:\n    train = transformer.transform(train)\n    valid = transformer.transform(valid)\n    test = transformer.transform(test)\n\n# 4. Train model\nmodel = dc.models.MultitaskRegressor(\n    n_tasks=1,\n    n_features=2048,\n    layer_sizes=[1000, 500],\n    dropouts=0.25\n)\nmodel.fit(train, nb_epoch=50)\n\n# 5. Evaluate\nmetric = dc.metrics.Metric(dc.metrics.r2_score)\ntest_score = model.evaluate(test, [metric])\n```\n\n### Workflow C: Transfer Learning on Small Dataset\n\nFor leveraging pretrained models:\n\n```python\nimport deepchem as dc\n\n# 1. Load data (pretrained models often need raw SMILES)\nloader = dc.data.CSVLoader(\n    tasks=['activity'],\n    feature_field='smiles',\n    featurizer=dc.feat.DummyFeaturizer()  # Model handles featurization\n)\ndataset = loader.create_dataset('small_dataset.csv')\n\n# 2. Split data\nsplitter = dc.splits.ScaffoldSplitter()\ntrain, test = splitter.train_test_split(dataset)\n\n# 3. Load pretrained model\nmodel = dc.models.HuggingFaceModel(\n    model='seyonec\u002FChemBERTa-zinc-base-v1',\n    task='classification',\n    n_tasks=1,\n    learning_rate=2e-5\n)\n\n# 4. Fine-tune\nmodel.fit(train, nb_epoch=10)\n\n# 5. Evaluate\npredictions = model.predict(test)\n```\n\nSee `references\u002Fworkflows.md` for 8 detailed workflow examples covering molecular generation, materials science, protein analysis, and more.\n\n## Example Scripts\n\nThis skill includes three production-ready scripts in the `scripts\u002F` directory:\n\n### 1. `predict_solubility.py`\nTrain and evaluate solubility prediction models. Works with Delaney benchmark or custom CSV data.\n\n```bash\n# Use Delaney benchmark\npython scripts\u002Fpredict_solubility.py\n\n# Use custom data\npython scripts\u002Fpredict_solubility.py \\\n    --data my_data.csv \\\n    --smiles-col smiles \\\n    --target-col solubility \\\n    --predict \"CCO\" \"c1ccccc1\"\n```\n\n### 2. `graph_neural_network.py`\nTrain various graph neural network architectures on molecular data.\n\n```bash\n# Train GCN on Tox21\npython scripts\u002Fgraph_neural_network.py --model gcn --dataset tox21\n\n# Train AttentiveFP on custom data\npython scripts\u002Fgraph_neural_network.py \\\n    --model attentivefp \\\n    --data molecules.csv \\\n    --task-type regression \\\n    --targets activity \\\n    --epochs 100\n```\n\n### 3. `transfer_learning.py`\nFine-tune pretrained models (ChemBERTa, GROVER) on molecular property prediction tasks.\n\n```bash\n# Fine-tune ChemBERTa on BBBP\npython scripts\u002Ftransfer_learning.py --model chemberta --dataset bbbp\n\n# Fine-tune GROVER on custom data\npython scripts\u002Ftransfer_learning.py \\\n    --model grover \\\n    --data small_dataset.csv \\\n    --target activity \\\n    --task-type classification \\\n    --epochs 20\n```\n\n## Common Patterns and Best Practices\n\n### Pattern 1: Always Use Scaffold Splitting for Molecules\n```python\n# GOOD: Prevents data leakage\nsplitter = dc.splits.ScaffoldSplitter()\ntrain, test = splitter.train_test_split(dataset)\n\n# BAD: Similar molecules in train and test\nsplitter = dc.splits.RandomSplitter()\ntrain, test = splitter.train_test_split(dataset)\n```\n\n### Pattern 2: Normalize Features and Targets\n```python\ntransformers = [\n    dc.trans.NormalizationTransformer(\n        transform_y=True,  # Also normalize target values\n        dataset=train\n    )\n]\nfor transformer in transformers:\n    train = transformer.transform(train)\n    test = transformer.transform(test)\n```\n\n### Pattern 3: Start Simple, Then Scale\n1. Start with Random Forest + CircularFingerprint (fast baseline)\n2. Try XGBoost\u002FLightGBM if RF works well\n3. Move to deep learning (MultitaskRegressor) if you have >5K samples\n4. Try GNNs if you have >10K samples\n5. Use transfer learning for small datasets or novel scaffolds\n\n### Pattern 4: Handle Imbalanced Data\n```python\n# Option 1: Balancing transformer\ntransformer = dc.trans.BalancingTransformer(dataset=train)\ntrain = transformer.transform(train)\n\n# Option 2: Use balanced metrics\nmetric = dc.metrics.Metric(dc.metrics.balanced_accuracy_score)\n```\n\n### Pattern 5: Avoid Memory Issues\n```python\n# Use DiskDataset for large datasets\ndataset = dc.data.DiskDataset.from_numpy(X, y, w, ids)\n\n# Use smaller batch sizes\nmodel = dc.models.GCNModel(batch_size=32)  # Instead of 128\n```\n\n## Common Pitfalls\n\n### Issue 1: Data Leakage in Drug Discovery\n**Problem**: Using random splitting allows similar molecules in train\u002Ftest sets.\n**Solution**: Always use `ScaffoldSplitter` for molecular datasets.\n\n### Issue 2: GNN Underperforming vs Fingerprints\n**Problem**: Graph neural networks perform worse than simple fingerprints.\n**Solutions**:\n- Ensure dataset is large enough (>10K samples typically)\n- Increase training epochs (50-100)\n- Try different architectures (AttentiveFP, DMPNN instead of GCN)\n- Use pretrained models (GROVER)\n\n### Issue 3: Overfitting on Small Datasets\n**Problem**: Model memorizes training data.\n**Solutions**:\n- Use stronger regularization (increase dropout to 0.5)\n- Use simpler models (Random Forest instead of deep learning)\n- Apply transfer learning (ChemBERTa, GROVER)\n- Collect more data\n\n### Issue 4: Import Errors\n**Problem**: Module not found errors.\n**Solution**: Ensure DeepChem is installed with required dependencies:\n```bash\nuv pip install deepchem\n# For PyTorch models\nuv pip install deepchem[torch]\n# For all features\nuv pip install deepchem[all]\n```\n\n## Reference Documentation\n\nThis skill includes comprehensive reference documentation:\n\n### `references\u002Fapi_reference.md`\nComplete API documentation including:\n- All data loaders and their use cases\n- Dataset classes and when to use each\n- Complete featurizer catalog with selection guide\n- Model catalog organized by category (50+ models)\n- MoleculeNet dataset descriptions\n- Metrics and evaluation functions\n- Common code patterns\n\n**When to reference**: Search this file when you need specific API details, parameter names, or want to explore available options.\n\n### `references\u002Fworkflows.md`\nEight detailed end-to-end workflows:\n1. Molecular property prediction from SMILES\n2. Using MoleculeNet benchmarks\n3. Hyperparameter optimization\n4. Transfer learning with pretrained models\n5. Molecular generation with GANs\n6. Materials property prediction\n7. Protein sequence analysis\n8. Custom model integration\n\n**When to reference**: Use these workflows as templates for implementing complete solutions.\n\n## Installation Notes\n\nBasic installation:\n```bash\nuv pip install deepchem\n```\n\nFor PyTorch models (GCN, GAT, etc.):\n```bash\nuv pip install deepchem[torch]\n```\n\nFor all features:\n```bash\nuv pip install deepchem[all]\n```\n\nIf import errors occur, the user may need specific dependencies. Check the DeepChem documentation for detailed installation instructions.\n\n## Additional Resources\n\n- Official documentation: https:\u002F\u002Fdeepchem.readthedocs.io\u002F\n- GitHub repository: https:\u002F\u002Fgithub.com\u002Fdeepchem\u002Fdeepchem\n- Tutorials: https:\u002F\u002Fdeepchem.readthedocs.io\u002Fen\u002Flatest\u002Fget_started\u002Ftutorials.html\n- Paper: \"MoleculeNet: A Benchmark for Molecular Machine Learning\"\n","# DeepChem\n\n## Overview\n\nDeepChem is a comprehensive Python library for applying machine learning to chemistry, materials science, and biology. Enable molecular property prediction, drug discovery, materials design, and biomolecule analysis through specialized neural networks, molecular featurization methods, and pretrained models.\n\n## When to Use This Skill\n\nThis skill should be used when:\n- Loading and processing molecular data (SMILES strings, SDF files, protein sequences)\n- Predicting molecular properties (solubility, toxicity, binding affinity, ADMET properties)\n- Training models on chemical\u002Fbiological datasets\n- Using MoleculeNet benchmark datasets (Tox21, BBBP, Delaney, etc.)\n- Converting molecules to ML-ready features (fingerprints, graph representations, descriptors)\n- Implementing graph neural networks for molecules (GCN, GAT, MPNN, AttentiveFP)\n- Applying transfer learning with pretrained models (ChemBERTa, GROVER, MolFormer)\n- Predicting crystal\u002Fmaterials properties (bandgap, formation energy)\n- Analyzing protein or DNA sequences\n\n## Core Capabilities\n\n### 1. Molecular Data Loading and Processing\n\nDeepChem provides specialized loaders for various chemical data formats:\n\n```python\nimport deepchem as dc\n\n# Load CSV with SMILES\nfeaturizer = dc.feat.CircularFingerprint(radius=2, size=2048)\nloader = dc.data.CSVLoader(\n    tasks=['solubility', 'toxicity'],\n    feature_field='smiles',\n    featurizer=featurizer\n)\ndataset = loader.create_dataset('molecules.csv')\n\n# Load SDF files\nloader = dc.data.SDFLoader(tasks=['activity'], featurizer=featurizer)\ndataset = loader.create_dataset('compounds.sdf')\n\n# Load protein sequences\nloader = dc.data.FASTALoader()\ndataset = loader.create_dataset('proteins.fasta')\n```\n\n**Key Loaders**:\n- `CSVLoader`: Tabular data with molecular identifiers\n- `SDFLoader`: Molecular structure files\n- `FASTALoader`: Protein\u002FDNA sequences\n- `ImageLoader`: Molecular images\n- `JsonLoader`: JSON-formatted datasets\n\n### 2. Molecular Featurization\n\nConvert molecules into numerical representations for ML models.\n\n#### Decision Tree for Featurizer Selection\n\n```\nIs the model a graph neural network?\n├─ YES → Use graph featurizers\n│   ├─ Standard GNN → MolGraphConvFeaturizer\n│   ├─ Message passing → DMPNNFeaturizer\n│   └─ Pretrained → GroverFeaturizer\n│\n└─ NO → What type of model?\n    ├─ Traditional ML (RF, XGBoost, SVM)\n    │   ├─ Fast baseline → CircularFingerprint (ECFP)\n    │   ├─ Interpretable → RDKitDescriptors\n    │   └─ Maximum coverage → MordredDescriptors\n    │\n    ├─ Deep learning (non-graph)\n    │   ├─ Dense networks → CircularFingerprint\n    │   └─ CNN → SmilesToImage\n    │\n    ├─ Sequence models (LSTM, Transformer)\n    │   └─ SmilesToSeq\n    │\n    └─ 3D structure analysis\n        └─ CoulombMatrix\n```\n\n#### Example Featurization\n\n```python\n# Fingerprints (for traditional ML)\nfp = dc.feat.CircularFingerprint(radius=2, size=2048)\n\n# Descriptors (for interpretable models)\ndesc = dc.feat.RDKitDescriptors()\n\n# Graph features (for GNNs)\ngraph_feat = dc.feat.MolGraphConvFeaturizer()\n\n# Apply featurization\nfeatures = fp.featurize(['CCO', 'c1ccccc1'])\n```\n\n**Selection Guide**:\n- **Small datasets (\u003C1K)**: CircularFingerprint or RDKitDescriptors\n- **Medium datasets (1K-100K)**: CircularFingerprint or graph featurizers\n- **Large datasets (>100K)**: Graph featurizers (MolGraphConvFeaturizer, DMPNNFeaturizer)\n- **Transfer learning**: Pretrained model featurizers (GroverFeaturizer)\n\nSee `references\u002Fapi_reference.md` for complete featurizer documentation.\n\n### 3. Data Splitting\n\n**Critical**: For drug discovery tasks, use `ScaffoldSplitter` to prevent data leakage from similar molecular structures appearing in both training and test sets.\n\n```python\n# Scaffold splitting (recommended for molecules)\nsplitter = dc.splits.ScaffoldSplitter()\ntrain, valid, test = splitter.train_valid_test_split(\n    dataset,\n    frac_train=0.8,\n    frac_valid=0.1,\n    frac_test=0.1\n)\n\n# Random splitting (for non-molecular data)\nsplitter = dc.splits.RandomSplitter()\ntrain, test = splitter.train_test_split(dataset)\n\n# Stratified splitting (for imbalanced classification)\nsplitter = dc.splits.RandomStratifiedSplitter()\ntrain, test = splitter.train_test_split(dataset)\n```\n\n**Available Splitters**:\n- `ScaffoldSplitter`: Split by molecular scaffolds (prevents leakage)\n- `ButinaSplitter`: Clustering-based molecular splitting\n- `MaxMinSplitter`: Maximize diversity between sets\n- `RandomSplitter`: Random splitting\n- `RandomStratifiedSplitter`: Preserves class distributions\n\n### 4. Model Selection and Training\n\n#### Quick Model Selection Guide\n\n| Dataset Size | Task | Recommended Model | Featurizer |\n|-------------|------|-------------------|------------|\n| \u003C 1K samples | Any | SklearnModel (RandomForest) | CircularFingerprint |\n| 1K-100K | Classification\u002FRegression | GBDTModel or MultitaskRegressor | CircularFingerprint |\n| > 100K | Molecular properties | GCNModel, AttentiveFPModel, DMPNNModel | MolGraphConvFeaturizer |\n| Any (small preferred) | Transfer learning | ChemBERTa, GROVER, MolFormer | Model-specific |\n| Crystal structures | Materials properties | CGCNNModel, MEGNetModel | Structure-based |\n| Protein sequences | Protein properties | ProtBERT | Sequence-based |\n\n#### Example: Traditional ML\n```python\nfrom sklearn.ensemble import RandomForestRegressor\n\n# Wrap scikit-learn model\nsklearn_model = RandomForestRegressor(n_estimators=100)\nmodel = dc.models.SklearnModel(model=sklearn_model)\nmodel.fit(train)\n```\n\n#### Example: Deep Learning\n```python\n# Multitask regressor (for fingerprints)\nmodel = dc.models.MultitaskRegressor(\n    n_tasks=2,\n    n_features=2048,\n    layer_sizes=[1000, 500],\n    dropouts=0.25,\n    learning_rate=0.001\n)\nmodel.fit(train, nb_epoch=50)\n```\n\n#### Example: Graph Neural Networks\n```python\n# Graph Convolutional Network\nmodel = dc.models.GCNModel(\n    n_tasks=1,\n    mode='regression',\n    batch_size=128,\n    learning_rate=0.001\n)\nmodel.fit(train, nb_epoch=50)\n\n# Graph Attention Network\nmodel = dc.models.GATModel(n_tasks=1, mode='classification')\nmodel.fit(train, nb_epoch=50)\n\n# Attentive Fingerprint\nmodel = dc.models.AttentiveFPModel(n_tasks=1, mode='regression')\nmodel.fit(train, nb_epoch=50)\n```\n\n### 5. MoleculeNet Benchmarks\n\nQuick access to 30+ curated benchmark datasets with standardized train\u002Fvalid\u002Ftest splits:\n\n```python\n# Load benchmark dataset\ntasks, datasets, transformers = dc.molnet.load_tox21(\n    featurizer='GraphConv',  # or 'ECFP', 'Weave', 'Raw'\n    splitter='scaffold',     # or 'random', 'stratified'\n    reload=False\n)\ntrain, valid, test = datasets\n\n# Train and evaluate\nmodel = dc.models.GCNModel(n_tasks=len(tasks), mode='classification')\nmodel.fit(train, nb_epoch=50)\n\nmetric = dc.metrics.Metric(dc.metrics.roc_auc_score)\ntest_score = model.evaluate(test, [metric])\n```\n\n**Common Datasets**:\n- **Classification**: `load_tox21()`, `load_bbbp()`, `load_hiv()`, `load_clintox()`\n- **Regression**: `load_delaney()`, `load_freesolv()`, `load_lipo()`\n- **Quantum properties**: `load_qm7()`, `load_qm8()`, `load_qm9()`\n- **Materials**: `load_perovskite()`, `load_bandgap()`, `load_mp_formation_energy()`\n\nSee `references\u002Fapi_reference.md` for complete dataset list.\n\n### 6. Transfer Learning\n\nLeverage pretrained models for improved performance, especially on small datasets:\n\n```python\n# ChemBERTa (BERT pretrained on 77M molecules)\nmodel = dc.models.HuggingFaceModel(\n    model='seyonec\u002FChemBERTa-zinc-base-v1',\n    task='classification',\n    n_tasks=1,\n    learning_rate=2e-5  # Lower LR for fine-tuning\n)\nmodel.fit(train, nb_epoch=10)\n\n# GROVER (graph transformer pretrained on 10M molecules)\nmodel = dc.models.GroverModel(\n    task='regression',\n    n_tasks=1\n)\nmodel.fit(train, nb_epoch=20)\n```\n\n**When to use transfer learning**:\n- Small datasets (\u003C 1000 samples)\n- Novel molecular scaffolds\n- Limited computational resources\n- Need for rapid prototyping\n\nUse the `scripts\u002Ftransfer_learning.py` script for guided transfer learning workflows.\n\n### 7. Model Evaluation\n\n```python\n# Define metrics\nclassification_metrics = [\n    dc.metrics.Metric(dc.metrics.roc_auc_score, name='ROC-AUC'),\n    dc.metrics.Metric(dc.metrics.accuracy_score, name='Accuracy'),\n    dc.metrics.Metric(dc.metrics.f1_score, name='F1')\n]\n\nregression_metrics = [\n    dc.metrics.Metric(dc.metrics.r2_score, name='R²'),\n    dc.metrics.Metric(dc.metrics.mean_absolute_error, name='MAE'),\n    dc.metrics.Metric(dc.metrics.root_mean_squared_error, name='RMSE')\n]\n\n# Evaluate\ntrain_scores = model.evaluate(train, classification_metrics)\ntest_scores = model.evaluate(test, classification_metrics)\n```\n\n### 8. Making Predictions\n\n```python\n# Predict on test set\npredictions = model.predict(test)\n\n# Predict on new molecules\nnew_smiles = ['CCO', 'c1ccccc1', 'CC(C)O']\nnew_features = featurizer.featurize(new_smiles)\nnew_dataset = dc.data.NumpyDataset(X=new_features)\n\n# Apply same transformations as training\nfor transformer in transformers:\n    new_dataset = transformer.transform(new_dataset)\n\npredictions = model.predict(new_dataset)\n```\n\n## Typical Workflows\n\n### Workflow A: Quick Benchmark Evaluation\n\nFor evaluating a model on standard benchmarks:\n\n```python\nimport deepchem as dc\n\n# 1. Load benchmark\ntasks, datasets, _ = dc.molnet.load_bbbp(\n    featurizer='GraphConv',\n    splitter='scaffold'\n)\ntrain, valid, test = datasets\n\n# 2. Train model\nmodel = dc.models.GCNModel(n_tasks=len(tasks), mode='classification')\nmodel.fit(train, nb_epoch=50)\n\n# 3. Evaluate\nmetric = dc.metrics.Metric(dc.metrics.roc_auc_score)\ntest_score = model.evaluate(test, [metric])\nprint(f\"Test ROC-AUC: {test_score}\")\n```\n\n### Workflow B: Custom Data Prediction\n\nFor training on custom molecular datasets:\n\n```python\nimport deepchem as dc\n\n# 1. Load and featurize data\nfeaturizer = dc.feat.CircularFingerprint(radius=2, size=2048)\nloader = dc.data.CSVLoader(\n    tasks=['activity'],\n    feature_field='smiles',\n    featurizer=featurizer\n)\ndataset = loader.create_dataset('my_molecules.csv')\n\n# 2. Split data (use ScaffoldSplitter for molecules!)\nsplitter = dc.splits.ScaffoldSplitter()\ntrain, valid, test = splitter.train_valid_test_split(dataset)\n\n# 3. Normalize (optional but recommended)\ntransformers = [dc.trans.NormalizationTransformer(\n    transform_y=True, dataset=train\n)]\nfor transformer in transformers:\n    train = transformer.transform(train)\n    valid = transformer.transform(valid)\n    test = transformer.transform(test)\n\n# 4. Train model\nmodel = dc.models.MultitaskRegressor(\n    n_tasks=1,\n    n_features=2048,\n    layer_sizes=[1000, 500],\n    dropouts=0.25\n)\nmodel.fit(train, nb_epoch=50)\n\n# 5. Evaluate\nmetric = dc.metrics.Metric(dc.metrics.r2_score)\ntest_score = model.evaluate(test, [metric])\n```\n\n### Workflow C: Transfer Learning on Small Dataset\n\nFor leveraging pretrained models:\n\n```python\nimport deepchem as dc\n\n# 1. Load data (pretrained models often need raw SMILES)\nloader = dc.data.CSVLoader(\n    tasks=['activity'],\n    feature_field='smiles',\n    featurizer=dc.feat.DummyFeaturizer()  # Model handles featurization\n)\ndataset = loader.create_dataset('small_dataset.csv')\n\n# 2. Split data\nsplitter = dc.splits.ScaffoldSplitter()\ntrain, test = splitter.train_test_split(dataset)\n\n# 3. Load pretrained model\nmodel = dc.models.HuggingFaceModel(\n    model='seyonec\u002FChemBERTa-zinc-base-v1',\n    task='classification',\n    n_tasks=1,\n    learning_rate=2e-5\n)\n\n# 4. Fine-tune\nmodel.fit(train, nb_epoch=10)\n\n# 5. Evaluate\npredictions = model.predict(test)\n```\n\nSee `references\u002Fworkflows.md` for 8 detailed workflow examples covering molecular generation, materials science, protein analysis, and more.\n\n## Example Scripts\n\nThis skill includes three production-ready scripts in the `scripts\u002F` directory:\n\n### 1. `predict_solubility.py`\nTrain and evaluate solubility prediction models. Works with Delaney benchmark or custom CSV data.\n\n```bash\n# Use Delaney benchmark\npython scripts\u002Fpredict_solubility.py\n\n# Use custom data\npython scripts\u002Fpredict_solubility.py \\\n    --data my_data.csv \\\n    --smiles-col smiles \\\n    --target-col solubility \\\n    --predict \"CCO\" \"c1ccccc1\"\n```\n\n### 2. `graph_neural_network.py`\nTrain various graph neural network architectures on molecular data.\n\n```bash\n# Train GCN on Tox21\npython scripts\u002Fgraph_neural_network.py --model gcn --dataset tox21\n\n# Train AttentiveFP on custom data\npython scripts\u002Fgraph_neural_network.py \\\n    --model attentivefp \\\n    --data molecules.csv \\\n    --task-type regression \\\n    --targets activity \\\n    --epochs 100\n```\n\n### 3. `transfer_learning.py`\nFine-tune pretrained models (ChemBERTa, GROVER) on molecular property prediction tasks.\n\n```bash\n# Fine-tune ChemBERTa on BBBP\npython scripts\u002Ftransfer_learning.py --model chemberta --dataset bbbp\n\n# Fine-tune GROVER on custom data\npython scripts\u002Ftransfer_learning.py \\\n    --model grover \\\n    --data small_dataset.csv \\\n    --target activity \\\n    --task-type classification \\\n    --epochs 20\n```\n\n## Common Patterns and Best Practices\n\n### Pattern 1: Always Use Scaffold Splitting for Molecules\n```python\n# GOOD: Prevents data leakage\nsplitter = dc.splits.ScaffoldSplitter()\ntrain, test = splitter.train_test_split(dataset)\n\n# BAD: Similar molecules in train and test\nsplitter = dc.splits.RandomSplitter()\ntrain, test = splitter.train_test_split(dataset)\n```\n\n### Pattern 2: Normalize Features and Targets\n```python\ntransformers = [\n    dc.trans.NormalizationTransformer(\n        transform_y=True,  # Also normalize target values\n        dataset=train\n    )\n]\nfor transformer in transformers:\n    train = transformer.transform(train)\n    test = transformer.transform(test)\n```\n\n### Pattern 3: Start Simple, Then Scale\n1. Start with Random Forest + CircularFingerprint (fast baseline)\n2. Try XGBoost\u002FLightGBM if RF works well\n3. Move to deep learning (MultitaskRegressor) if you have >5K samples\n4. Try GNNs if you have >10K samples\n5. Use transfer learning for small datasets or novel scaffolds\n\n### Pattern 4: Handle Imbalanced Data\n```python\n# Option 1: Balancing transformer\ntransformer = dc.trans.BalancingTransformer(dataset=train)\ntrain = transformer.transform(train)\n\n# Option 2: Use balanced metrics\nmetric = dc.metrics.Metric(dc.metrics.balanced_accuracy_score)\n```\n\n### Pattern 5: Avoid Memory Issues\n```python\n# Use DiskDataset for large datasets\ndataset = dc.data.DiskDataset.from_numpy(X, y, w, ids)\n\n# Use smaller batch sizes\nmodel = dc.models.GCNModel(batch_size=32)  # Instead of 128\n```\n\n## Common Pitfalls\n\n### Issue 1: Data Leakage in Drug Discovery\n**Problem**: Using random splitting allows similar molecules in train\u002Ftest sets.\n**Solution**: Always use `ScaffoldSplitter` for molecular datasets.\n\n### Issue 2: GNN Underperforming vs Fingerprints\n**Problem**: Graph neural networks perform worse than simple fingerprints.\n**Solutions**:\n- Ensure dataset is large enough (>10K samples typically)\n- Increase training epochs (50-100)\n- Try different architectures (AttentiveFP, DMPNN instead of GCN)\n- Use pretrained models (GROVER)\n\n### Issue 3: Overfitting on Small Datasets\n**Problem**: Model memorizes training data.\n**Solutions**:\n- Use stronger regularization (increase dropout to 0.5)\n- Use simpler models (Random Forest instead of deep learning)\n- Apply transfer learning (ChemBERTa, GROVER)\n- Collect more data\n\n### Issue 4: Import Errors\n**Problem**: Module not found errors.\n**Solution**: Ensure DeepChem is installed with required dependencies:\n```bash\nuv pip install deepchem\n# For PyTorch models\nuv pip install deepchem[torch]\n# For all features\nuv pip install deepchem[all]\n```\n\n## Reference Documentation\n\nThis skill includes comprehensive reference documentation:\n\n### `references\u002Fapi_reference.md`\nComplete API documentation including:\n- All data loaders and their use cases\n- Dataset classes and when to use each\n- Complete featurizer catalog with selection guide\n- Model catalog organized by category (50+ models)\n- MoleculeNet dataset descriptions\n- Metrics and evaluation functions\n- Common code patterns\n\n**When to reference**: Search this file when you need specific API details, parameter names, or want to explore available options.\n\n### `references\u002Fworkflows.md`\nEight detailed end-to-end workflows:\n1. Molecular property prediction from SMILES\n2. Using MoleculeNet benchmarks\n3. Hyperparameter optimization\n4. Transfer learning with pretrained models\n5. Molecular generation with GANs\n6. Materials property prediction\n7. Protein sequence analysis\n8. Custom model integration\n\n**When to reference**: Use these workflows as templates for implementing complete solutions.\n\n## Installation Notes\n\nBasic installation:\n```bash\nuv pip install deepchem\n```\n\nFor PyTorch models (GCN, GAT, etc.):\n```bash\nuv pip install deepchem[torch]\n```\n\nFor all features:\n```bash\nuv pip install deepchem[all]\n```\n\nIf import errors occur, the user may need specific dependencies. Check the DeepChem documentation for detailed installation instructions.\n\n## Additional Resources\n\n- Official documentation: https:\u002F\u002Fdeepchem.readthedocs.io\u002F\n- GitHub repository: https:\u002F\u002Fgithub.com\u002Fdeepchem\u002Fdeepchem\n- Tutorials: https:\u002F\u002Fdeepchem.readthedocs.io\u002Fen\u002Flatest\u002Fget_started\u002Ftutorials.html\n- Paper: \"MoleculeNet: A Benchmark for Molecular Machine Learning\"\n",{"name":5,"description":7},"main","2b749c82b5d11fac39d9a6e021ce3f5b8a541bd1","cli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fdeepchem",3931,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates","https:\u002F\u002Faitmpl.com","https:\u002F\u002Fgithub.com\u002Fdavila7","https:\u002F\u002Fdanielavila.me","Building AI dev tools with LLMs","2026-11-14 06:14:46",{"source":8,"skillName":5,"commands":54,"prompt":64,"zipAvailable":65},[55,58,61],{"tool":56,"command":57},"npx","npx skills add davila7\u002Fclaude-code-templates --skill deepchem",{"tool":59,"command":60},"bunx","bunx skills add davila7\u002Fclaude-code-templates --skill deepchem",{"tool":62,"command":63},"pnpm","pnpm dlx skills add davila7\u002Fclaude-code-templates --skill deepchem","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fdeepchem\u002FSKILL.md",true,[67,73,77,83,87],{"path":68,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":71,"sourceHtmlUrl":72},"references\u002Fapi_reference.md","REFERENCE","参考文档",11597,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fdeepchem\u002Freferences\u002Fapi_reference.md",{"path":74,"resourceType":69,"resourceTypeDesc":70,"sizeBytes":75,"sourceHtmlUrl":76},"references\u002Fworkflows.md",11800,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fdeepchem\u002Freferences\u002Fworkflows.md",{"path":78,"resourceType":79,"resourceTypeDesc":80,"sizeBytes":81,"sourceHtmlUrl":82},"scripts\u002Fgraph_neural_network.py","SCRIPT","脚本",9742,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fdeepchem\u002Fscripts\u002Fgraph_neural_network.py",{"path":84,"resourceType":79,"resourceTypeDesc":80,"sizeBytes":85,"sourceHtmlUrl":86},"scripts\u002Fpredict_solubility.py",6546,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fdeepchem\u002Fscripts\u002Fpredict_solubility.py",{"path":88,"resourceType":79,"resourceTypeDesc":80,"sizeBytes":89,"sourceHtmlUrl":90},"scripts\u002Ftransfer_learning.py",11353,"https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fscientific\u002Fdeepchem\u002Fscripts\u002Ftransfer_learning.py",[92,119,146,179,214,237],{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":99,"publishedAt":38},"01M70NHBTQ8BV0A2SCAZXEPHCX","wealthville-pool-scoring","Score DeFi liquidity pools before providing liquidity. Returns a 0-100 pool score and an Enter\u002FHold\u002FExit\u002FReduce\u002FAvoid verdict with calibrated confidence for ~68,800 Solana pools and 575 EVM pools, plus a public miss-inclusive track record. Use when asked whether a pool is worth LPing into, which pools look best right now, or how accurate the scores have actually been.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fweb-data\u002Fwealthville-pool-scoring\u002FSKILL.md",[98],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[100,101,102,103,104,34,105,106,107,108,109,110,111,112,113,114,115,116,117,118],"scoring","pool","wealthville pool scoring","liquidity","which","exit","now","use","when","reduce","with","record","track","verdict","evm","solana","how","public","look",{"skillNo":120,"name":121,"nameZh":6,"description":122,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":123,"categories":124,"keywords":126,"publishedAt":38},"01M70MB6VR8ERJ4EM8D5KHBMEV","technical-debt-auditor","Scans a codebase and produces a quantified technical debt report with a prioritized remediation plan. Analyzes dead code, TODO\u002FFIXME\u002FHACK markers, dependency staleness, complexity hotspots, code duplication signals, test coverage gaps, unused exports, and outdated patterns. Use this skill when the user asks to audit their codebase for tech debt, wants a health check of their code quality, says things like \"where should I focus refactoring\", \"find code smells\", \"audit my codebase\", \"what technical debt do we have\", \"prioritize cleanup work\", or asks about code quality metrics. Also trigger when the user mentions \"code rot\", \"legacy cleanup\", \"refactoring priorities\", or wants to understand the overall health of a project before a major effort. Trigger even if the user doesn't say \"technical debt\" explicitly — any request to systematically assess code quality and identify improvement priorities belongs here.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fdevelopment\u002Ftechnical-debt-auditor\u002FSKILL.md",[125],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[127,128,129,130,131,132,133,134,135,136,137,108,138,139,140,141,142,143,144,145],"technical debt","debt","technical","technical debt auditor","auditor","coverage","complexity","markers","project","cleanup","unused","code","todo","work","plan","focus","signals","audit","dependency",{"skillNo":147,"name":148,"nameZh":6,"description":149,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":150,"categories":151,"keywords":161,"publishedAt":38},"01M70KPYH9QNHCNME5GJQGE22K","readonly-sql-guard","Make an agent's database access provably read-only. Use when an agent runs caller-supplied SQL against a production\u002FERP\u002Fanalytics database (a \"query\" MCP tool, text-to-SQL, an admin console), or when reviewing DB-access code for write\u002FDDL\u002Fside-effect\u002Ftransaction-escape risks. Checks what a statement CALLS, not just its shape.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fdatabase\u002Freadonly-sql-guard\u002FSKILL.md",[152,155,156],{"categoryNo":153,"categoryName":154,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000010","后端与接口",{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":157,"categoryName":158,"taxonomyType":159,"taxonomyTypeDesc":160},"01JBSEEDCAT000000000000033","数据分析师","PROFESSION","适用职业",[162,163,164,165,166,167,168,34,169,170,171,172,173,174,175,107,176,177,108,178],"readonly","guard","sql guard","sql","readonly sql guard","erp","effect","read-only","db","just","tool","calls","only","MCP","access","database","query",{"skillNo":180,"name":181,"nameZh":6,"description":182,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":183,"licenseFileUrl":184,"sourceHtmlUrl":185,"categories":186,"keywords":190,"publishedAt":38},"01M6AFY6F7CC2BBZP1QD6VTNVC","web-artifacts-builder","Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn\u002Fui). Use for complex artifacts requiring state management, routing, or shadcn\u002Fui components - not for simple single-file HTML\u002FJSX artifacts.","Apache-2.0","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Futilities\u002Fweb-artifacts-builder\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Futilities\u002Fweb-artifacts-builder\u002FSKILL.md",[187],{"categoryNo":188,"categoryName":189,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000009","前端开发",[191,192,193,194,195,196,197,198,199,200,201,202,203,204,205,206,207,208,209,210,211,212,34,213],"artifacts builder","web","web artifacts builder","artifacts","builder","state","routing","state management","component","css","frontend","ui","Tools","file","components","shadcn","react","Tailwind","suite","jsx","html","single-file","ui components",{"skillNo":215,"name":216,"nameZh":6,"description":217,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":183,"licenseFileUrl":218,"sourceHtmlUrl":219,"categories":220,"keywords":225,"publishedAt":38},"01M6AFWF33SFV9Z3SF4PEEE4BF","webapp-testing","Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fsecurity\u002Fwebapp-testing\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fsecurity\u002Fwebapp-testing\u002FSKILL.md",[221,224],{"categoryNo":222,"categoryName":223,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":188,"categoryName":189,"taxonomyType":21,"taxonomyTypeDesc":22},[226,227,228,229,230,192,231,201,202,232,26,233,234,34,235,110,236],"webapp testing","testing","webapp","browser","playwright","applications","debugging","logs","screenshots","local","behavior",{"skillNo":238,"name":239,"nameZh":6,"description":240,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":183,"licenseFileUrl":241,"sourceHtmlUrl":242,"categories":243,"keywords":251,"publishedAt":38},"01M6AF67Q67JJJ5V51JR2P52KA","discernment-nudge","After you give a substantive answer or draft that the user may act on — advice or recommendations, drafted artifacts such as goals, plans, pitches, proposals, or emails, estimates or projections, analysis or interpretation of data, factual claims they may rely on, or a multi-step argument — invoke this skill BEFORE finalizing your reply and then, if it applies, append 2-3 short follow-up questions, each tied to something specific in what you just produced, that help the user check key facts, probe the reasoning or assumptions, and notice missing context. Do this at most once per conversation. Skip it when the user asked a trivial how-to or simple lookup, wants a purely educational explanation, asked you only to format, convert, or assemble a file from content they provided, is writing code they will run, is doing creative writing or casual chat, or already asked you to double-check, cite, or review — the skill file explains these boundaries and the exact output format.","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fproductivity\u002Fdiscernment-nudge\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Fdavila7\u002Fclaude-code-templates\u002Fblob\u002Fmain\u002Fcli-tool\u002Fcomponents\u002Fskills\u002Fproductivity\u002Fdiscernment-nudge\u002FSKILL.md",[244,247,248],{"categoryNo":245,"categoryName":246,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},{"categoryNo":249,"categoryName":250,"taxonomyType":159,"taxonomyTypeDesc":160},"01JBSEEDCAT000000000000036","内容创作者",[252,253,254,255,256,257,258,259,194,260,204,138,261,262,263,264,265,174,266,108,267,268,269,270,271,116,272],"discernment nudge","nudge","interpretation","help","run","reasoning","data","writing","lookup","content","review","conversation","short","up","invoke","you","analysis","claims","argument","assemble","recommendations"]