lightning-rod-labsOfficial·10 Agent Skills Included

lightningrod-python-sdk

Calibrated probability forecasts for any future question

Generates calibrated probability forecasts for any question about the future through an OpenAI-compatible API. Supports binary, multiple-choice, numeric, and free-text predictions with optional live web research for context. Removes guesswork from forecasting by returning structured, parsed results instead of raw text. Includes a verification skill that tests documentation code snippets against the live API.

57|6

↗

npx skills add lightning-rod-labs/lightningrod-python-sdk --all -g -y

Available:

Explains the SDK architecture, public API surface, and test commands so the agent can navigate the codebase and run predictions or tests correctly.

All Skills in This Repository (10)

Pure Emerald Level Indicators

📦 In Repo

lightning-rod-labs

examples-guide

Guide selection of Lightning Rod training patterns and answer framing by dataset source and prediction goal.

transform-pipeline-verification

Verify Lightning Rod transform pipeline outputs by running small-stage jobs and inspecting dataset rows.

custom-dataset-seeds

Generate Lightning Rod forecasting seed samples from PDFs, CSVs, and text files.

content-learning-examples

Generate SFT question-and-answer training examples from documents or web search.

tabular-examples

Convert tabular data into forecasting training samples with labels and temporal metadata.

lightningrod-assistant

Generate forecasting datasets and fine-tune Lightning Rod models.

bigquery-seeds

Creates BigQuery-backed seed datasets from publicly queryable tables for forecasting and training.

experiment-tracking

Records training experiment configs into notebooks and an index.

public-dataset-exploration

Search Kaggle, Hugging Face, and GitHub for raw public datasets.

forward-looking-examples

Generate GRPO forecasting datasets with seed generation, linting, and temporal filtering.

Official

Advanced

Frequently Asked Questions

FAQPage Schema

How to install Lightning Rod Python SDK?▼

Run `npx skills add lightning-rod-labs/lightningrod-python-sdk --all -g -y` in your terminal to install all skills in this repository globally.

How to get probability forecasts with Python?▼

Install the SDK, then call `lr.predict()` with your question and an answer type like binary or multiple choice. It returns a parsed, calibrated probability you can use directly in code.

Does Foresight work with the OpenAI client?▼

Yes. The API is OpenAI-compatible, so any standard OpenAI client works by simply pointing the base URL at Lightning Rod's endpoint.

Can forecasts include live web research?▼

Yes. Set the research parameter to true and the model automatically gathers relevant web context before producing its probability estimate.

How to verify forecasting documentation accuracy?▼

The included verify-forecasting-docs skill runs every code snippet in the docs against the live API and flags any prose that contradicts actual behavior.