skills.rest
quantskills
quantskillsOFFICIAL
Explore all skills by this author →
GitHub

skill-quant-factor-skill-factory

Batch-generate and validate framework-neutral OHLCV quant factor Skills on real market data.

61|7|Updated Jun 11, 2026
Search Tags:#skill generation#ohlcv#alpha research#rank ic#quant factors#factor validation#batch pipeline
One-click install
npx skills add https://github.com/quantskills/skill-quant-factor-skill-factory --skill skill-quant-factor-skill-factory-quantskills
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: skill-quant-factor-skill-factory
Source: https://github.com/quantskills/skill-quant-factor-skill-factory
Command: npx skills add https://github.com/quantskills/skill-quant-factor-skill-factory --skill skill-quant-factor-skill-factory-quantskills

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pandas, and includes scripts (resource) and references (resource) components.

What problem does it solve? Turning a batch of OHLCV alpha ideas into installable, verifiable, agent-callable Skill packages requires repetitive scaffolding, deduplication against existing factor indexes, and real-data validation. This Skill automates that entire production pipeline so each new factor ships as a complete standalone Skill folder. ## Core Features & Use Cases - Batch Factor Skill Generation: Creates one standalone Skill folder per factor containing SKILL.md, bilingual README, scripts/factor.py, scripts/validate.py, references/formula.md, and agents/openai.yaml. - Real-Market Validation: Computes coverage, 5-day Rank IC, ICIR, Q5-Q1 spread, top-quintile turnover, and no-lookahead checks on cached AkShare A-share and Yahoo US panels, writing results to validation_real/result.json and report.md. - Index Management and Reporting: Deduplicates against existing factor indexes, maintains factor_index.json and merged combined indexes, and produces a batch evaluation report. - Use Case: You have 200 new factor ideas and an existing 1000-factor index. Run the batch script with --count 200 --start-id 1001 to generate, validate, and package all 200 factors while skipping any slugs already present in the prior index. ## Quick Start Ask the agent to use the skill-quant-factor-skill-factory to generate a new batch of OHLCV factor Skills starting from a given ID and validate them on the cached real market panel.

Frequently Asked Questions about skill-quant-factor-skill-factory

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I batch-generate quant factor Skills from OHLCV data?▼

Run scripts/generate_factor_skill_batch.py with --count, --start-id, --existing-index, --output-root, --combined-index, and --report-name arguments. It generates one standalone Skill folder per factor, validates each on the real market panel, and writes a merged index and evaluation report.

What validation metrics are computed for each generated factor?▼

Each factor is validated on the real OHLCV panel for usable-sample coverage, 5-day Rank IC mean, 5-day ICIR, quintile Q5-Q1 return spread, top-quintile turnover, and a no-lookahead check. Results are written to validation_real/result.json and validation_real/report.md.

What market data format does the factor factory require?▼

The target project must provide real_market_data/panels/ohlcv_panel.parquet plus a panel_manifest.json. The panel needs at least date, symbol, open, high, low, close, and volume columns, with an optional market field recommended.

Does the generated factor code depend on a specific trading framework?▼

No. Generated factor.py files are framework-neutral and use only pandas and numpy. The caller supplies the data vendor, universe, rebalance calendar, and cost model, so factors work with any backtesting or research stack.

How does the factory avoid generating duplicate factors?▼

The script loads every existing factor index passed via --existing-index and skips any candidate whose slug already appears. After generation it merges old and new rows into a combined index, which should contain no duplicate slugs.

Can I use this Skill without cached real market data?▼

No. The workflow requires a cached real OHLCV panel, preferring AkShare for China A-shares and Yahoo Finance for US data. The guardrails explicitly forbid describing synthetic validation as real validation.

Related Skills

Explore More Data & Analytics Skills→
SignmanalSignmanal
1

stocks

Retrieve stock and crypto quotes and OHLCV history via Yahoo Finance.

Community
Intermediate
santoosaraujosantoosaraujo

multi-factor

Calculate and combine financial factors to rank assets for portfolio construction.

Community
Advanced
HKUDSHKUDS
30.4k

factor-research

Compute IC/IR statistics and quantile backtest curves for factor selection.

Official
Advanced
varsity-tech-productvarsity-tech-product
12

factor-mining

Turn finance factor hypotheses into validated backtests and research artifacts.

Official
Advanced