skills.rest
Leiisawesome
LeiisawesomeCOMMUNITY
Explore all skills by this author →
GitHub

system-architect

Design a deterministic layered architecture for an intraday trading platform.

Updated Mar 2, 2026
Search Tags:#architecture#event-driven#determinism#layered-design#intraday-trading#alpha-modules
One-click install
npx skills add https://github.com/Leiisawesome/feelies --skill system-architect-leiisawesome
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: system-architect
Source: https://github.com/Leiisawesome/feelies/tree/main/.cursor/skills/system-architect
Command: npx skills add https://github.com/Leiisawesome/feelies --skill system-architect-leiisawesome

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provide a canonical, robust architecture for building a unified intraday trading platform that guarantees layer separation, determinism, and dual-mode behavior for research and live trading.

Core Features & Use Cases

  • Layered architecture with clearly defined responsibilities for data ingestion, event routing, feature computation, signal generation, risk management, execution, alpha module system, and portfolio tracking.
  • Deterministic replay via a kernel orchestrator and five state machines, ensuring end-to-end traceability and safe failure handling.
  • Dual-mode capability (research/backtest replay vs live trading) with mode-agnostic pipelines and deterministic behavior guarantees.
  • Alpha Module System supports multi-strategy deployments with lifecycle management and composable feature/signal engines.

Quick Start

Set up the Kernel orchestrator with defined layers and run a backtest to validate deterministic replay, then enable live mode using the same core engine.

Frequently Asked Questions about system-architect

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

FAQPage Schema
How do I design a deterministic architecture for an intraday trading platform?▼

A deterministic intraday trading architecture uses a kernel orchestrator with state machines to separate layers like data ingestion, signal generation, and execution, ensuring end-to-end traceability and safe failure handling.

What is the best way to guarantee layer separation in an event-driven trading system?▼

Layer separation in an event-driven trading system is guaranteed by defining explicit component boundaries and using a typed event catalog with canonical message formats across all cross-layer interactions.

How does deterministic replay work for backtesting and live trading workflows?▼

Deterministic replay works by using a kernel orchestrator with five state machines to process mode-agnostic pipelines, allowing the same core engine to execute both research backtests and live trading safely.

Can I manage multiple alpha modules within a unified intraday system architecture?▼

Yes, an alpha module system supports multi-strategy deployments by providing lifecycle management and composable feature and signal engines within the layered architecture.

Does this system architecture support explicit latency modeling for live trading?▼

Yes, the system architecture requires and satisfies explicit latency modeling alongside canonical message formats to orchestrate safe fail-safes during live intraday trading workflows.

What are the limitations of using a state machine kernel orchestrator for trading?▼

Using a five-state machine kernel orchestrator requires strict adherence to deterministic pipelines and canonical message formats; any deviation in cross-layer interactions can break end-to-end traceability and safe failure handling.

Related Skills

Explore More Software Engineering Skills→
JoelLewisJoelLewis
164

exchange-connectivity

Guide FIX session design for routing orders to the NYSE.

Community
Advanced
79yuuki79yuuki

strategy-validation-gate

Validate trading strategies and automated bots through structured testing and approval gates.

Community
Advanced
perfectuser21perfectuser21

repo-lead

Coordinate AI script and LLM employees to achieve OKRs via the Brain API.

Community
Advanced
LeiisawesomeLeiisawesome

backtest-engine

Replay L1 NBBO data with deterministic fills and latency modeling.

Community
Advanced