Kraken Trading Agent

Kraken AI Trading Agent is an autonomous crypto trading system built to research, validate, and execute trading strategies in a more adaptive and trustworthy way. Instead of relying on one static strategy, it maintains a research pipeline that screens multiple Kraken trading pairs, evaluates pair-specific strategy families on standardized backtests, diagnoses which candidates hold up across older and recent market windows, and promotes only the strongest ideas into a live research registry. This lets the system behave more like an evolving trading desk than a single hardcoded bot. The project is split into two major layers: research and execution. On the research side, Python modules run dense 15-minute backtests, compare performance across full 120-day, older-60-day, and recent-60-day windows, and generate proposal artifacts showing which pairs or strategies are worth promoting. On the execution side, an intraday agent runs on 15-minute intervals, reads the active registry, applies suppression logic during weak conditions, checks tradability and risk guardrails, and then paper-trades or live-trades the current validated book. n8n is used as the orchestration layer to schedule the daily research pipeline, trigger the intraday loop, and handle approval-style workflow automation, while the strategy logic itself stays in Python for transparency and reproducibility. A major focus of the system is trust and verification. The agent includes risk guardrails, portfolio-state suppression, real-time P&L tracking, research artifacts saved to disk, and ERC-8004 integration for trustless trade validation and reputation logging. The result is a trading agent that is not just trying to maximize returns, but also to make its decisions inspectable, testable, and safer to operate. In short, this project combines algorithmic trading, continuous strategy research, workflow automation, and on-chain accountability into a single hackathon-ready platform.