A curated catalog of 5,806 quantitative trading strategy specifications, source code algorithms, and indicators across crypto spot, perpetual futures, grid systems, statistical arbitrage, and market making.
A strategy specification, Pine Script snippet, or backtest dump is not execution software. Turning a research spec into a production trading system requires order execution logic, risk controls, idempotency, WebSocket reconnection handling, and test coverage.
AgenKit acts as an AI engineering team harness inside your coding tool—using the strategy file's math, parameters, and indicators as ground truth to build a test-driven, risk-managed trading system without re-deriving the logic from scratch.
/agenkit build a production-grade trading bot from strategies/Python-Version-Multi-Asset-Momentum-Strategy-Tutorial.md
Index the vault locally, then generate a production-ready trading system in one command:
# 1. Build the local codebase map (scans strategies at zero token cost)
npx agenkit memory buildThen in your AI harness chat (Claude Code, Cursor, Codex, Antigravity, etc.):
/agenkit build a production-grade trading bot from strategies/Python-Version-Multi-Asset-Momentum-Strategy-Tutorial.md with risk controls, live Binance/OKX execution, and a backtest engine
(For Codex CLI / Astra, use @agenkit instead of /agenkit)
The strategies in this vault are organized across five programming languages:
| Language | Specs Count | Primary Focus |
|---|---|---|
| Pine Script | ~5,286 | TradingView indicators, trend breakout, multi-timeframe overlays |
| JavaScript | ~362 | Node.js execution scripts, WebSocket market data feeds, exchange utilities |
| Python | ~132 | Backtesting frameworks, quantitative research models, machine learning |
| MyLanguage | ~27 | CTA futures trend & grid formulas |
| C++ | ~3 | Low-latency execution templates |
Full searchable index categorized by language: strategies/README.md.
Here is how AgenKit processes a spec such as strategies/Python-Version-Multi-Asset-Momentum-Strategy-Tutorial.md:
- Spec & Parameter Extraction: Parses the markdown file to extract signal math, momentum thresholds (
arrRatio), rebalancing frequencies, and asset lists directly from the spec ground truth. - Architecture & State Management: Designs non-custodial order routing, position state persistence, WebSocket feeds, and order deduplication/idempotency.
- Test-Driven Implementation: Writes unit tests for indicator math, mock exchange backtest execution, and paper-trading simulators before writing production implementation files.
- Safety & Risk Gating: Integrates hard stop-loss checks, position size limits, API rate-limit controls, and secret management.
- Gated Deployment: Configures paper trading by default, requiring explicit user activation before live order routing.
| Phase / Concern | Manual Strategy Porting | AgenKit-Assisted Build |
|---|---|---|
| Logic Transcription | Manual re-writing of formulas; prone to transcription errors | Uses spec parameters & indicator math directly as ground truth |
| Order Execution | Ad-hoc exchange API calls, vulnerable to dropped webhooks | Structured order lifecycle with idempotency and retry handlers |
| Risk Controls | Hardcoded or easily omitted position limits | Built-in risk gates (max drawdown limits, emergency halt, inventory caps) |
| Test Coverage | Manually written mock tests (frequently skipped) | Automated test suite (unit tests, mock backtest simulators, execution checks) |
| Execution Default | Often tested directly against live APIs | Paper trading gated by default; live routing requires explicit opt-in |
Below are example commands for common quantitative models in this vault:
/agenkit build a production research system from strategies/Python-Version-Multi-Asset-Momentum-Strategy-Tutorial.md
/agenkit implement strategies/Digital-Currency-Futures-Multi-Variety-ATR-Strategy-Teaching.md as a research backtest with delay-1 execution and fee modeling
/agenkit implement strategies/Adaptive-Intelligent-Grid-Trading-Strategy.md as a production grid bot for Hyperliquid perps with order deduplication and inventory controls
/agenkit build a spread arbitrage engine from strategies/High-Frequency-Intertemporal-Arbitrage-Strategy.md with WebSocket order routing
/agenkit build an Avellaneda-Stoikov market making bot for perps from strategies/Dynamic-Spread-Market-Making-Strategy.md with inventory skew management
Access the complete index of all 5,806 strategy specs in strategies/README.md.
Each strategy file contains:
- Name & Author
- Strategy Description & Math Logic
- Argument & Parameter Table
- Source Block
- Detail URL
- Educational & Research Purposes Only: The strategy specifications, algorithms, and source dumps contained in this repository are for educational and quantitative research purposes only. Nothing here constitutes financial, investment, legal, or tax advice.
- Unvalidated Research Specs: Published strategy logic and backtests are unvalidated dumps and do not guarantee future performance.
- Paper Trading Required: Always execute thorough paper-trading and backtesting in simulated environments before considering live capital deployment.
- Capital Risk: Quantitative trading involves substantial risk of financial loss. Never route live orders without independent code review, risk limit enforcement, and capital management controls.
Contributions to fix parameters, improve documentation, or add new research strategy specs are welcome. Please open a pull request or issue following the standard repository guidelines.
This repository is licensed under the MIT License.