Custom Prompts¶

Customize the agent's optimization goal with custom user prompts.

Default Prompt¶

Without customization, the agent uses:

Develop a machine learning model that generalizes well to new unseen data.

Using Custom Prompts¶

Command Line¶

agentomics-run --user-prompt "Your custom instructions here"

Examples¶

Simple models only:

agentomics-run --user-prompt "Only create simple ML models like logistic regression and shallow decision trees"

Focus on interpretability:

agentomics-run --user-prompt "Prioritize model interpretability over performance. Use models where feature importance can be easily explained."

Specific model type:

agentomics-run --user-prompt "Use gradient boosting models like XGBoost or LightGBM"

Handle imbalanced data:

agentomics-run --user-prompt "The dataset is highly imbalanced. Use appropriate techniques like SMOTE, class weights, or focal loss."

Neural networks:

agentomics-run --user-prompt "Focus on deep learning approaches. Design custom neural network architectures."

Quick iterations:

agentomics-run --user-prompt "Keep models simple and training fast. Avoid complex architectures that take long to train."

What Custom Prompts Affect¶

The user prompt influences the agentic steps:

Step	How It's Used
Iteration Planning	Cross-step guidance for the current iteration
Data Exploration	What to look for in the data
Data Split	Split strategy considerations
Data Representation	Feature encoding choices
Model Architecture	Model selection and design
Model Training	Training approach and hyperparameters
Model Inference	Prediction pipeline design

Validation evaluation itself is deterministic: it runs the generated inference script on train/validation data and scores the configured --val-metric.

Combining with Other Options¶

Custom prompts work with all other options:

agentomics-run \
  --user-prompt "Use only sklearn models, no neural networks" \
  --model openai/gpt-5.1-codex-max \
  --dataset my_data \
  --iterations 15 \
  --val-metric AUROC

Limitations¶

Custom prompts guide the agent but don't guarantee specific outcomes:

The agent may still try different approaches
Very restrictive prompts may limit performance
Some requests may not be feasible for certain datasets

Dataset Description vs User Prompt¶

Dataset Description	User Prompt
Domain information about the data	Instructions for the agent
Goes in `dataset_description.md`	Passed via `--user-prompt`
Describes what the data is	Describes what to do

Example dataset_description.md:

This dataset contains RNA-seq expression levels from tumor samples. Features are gene expression values.

Example user prompt:

Focus on gene signature discovery. Use feature selection to identify the most predictive genes.

Both can be used together - they complement each other.