Nexus Scalp Engine — Documentation

Research-driven quantitative trading platform — a hexagonal, event-driven scalping engine for MetaTrader 5 (primary market: XAUUSD M1), built around causal feature engineering, artifact-first model governance, deterministic research tooling, and forensic observability.

مهم

This documentation describes a research and engineering platform. It is

not investment advice, and nothing here is a promise of profitability.

Leveraged scalping carries extreme financial risk. See

Project Status for what is certified vs experimental.

Start here

Question	Page
What is Nexus and why does it exist?	Vision
How do I run it?	Quickstart
How is the architecture structured?	Architecture Overview
How does data become a decision?	Data Flow
How is research validated?	Research Methodology
What is real vs experimental vs planned?	Project Status · Capability Matrix
Where is the project going?	Roadmap
How do I contribute?	Contribution Guide

Documentation map

text

docs/
├── getting-started/   Installation · Quickstart · First-run · Configuration
├── project/           Vision · Scope · Status · Roadmap · Milestones
├── architecture/      Overview · System map · Runtime · Research stack ·
│                      Data flow · Observability · Database
├── research/          Methodology · Backtesting · Walk-forward · Validation ·
│                      Replay · Counterfactuals · Reproducibility
├── engineering/       Quality · Testing · CI · Release process · Security
├── guides/            CLI · Troubleshooting · Common workflows
├── contributing/      Contribution guide · Docs authoring · Adding a language
└── reference/         CLI reference · Glossary · Terminology · FAQ

Languages

Language	Status
🇬🇧 English	Source of truth (complete)
🇮🇷 فارسی	Persian (partial — core pages)
🇪🇸 Español	Spanish (partial — core pages)
🇸🇦 العربية	Arabic (partial — core pages)
🇩🇪 Deutsch	German (partial — core pages)

Translation coverage and staleness are audited by scripts/docs/check_translations.py — numbers, not vibes. See Translation workflow.

Deep technical documentation (repository-native)

The repository also carries a large body of agent-generated forensic and contract documentation. The most important entry points:

agents/skill.md — authoritative architecture map
agents/bugs.md — forensic bug ledger (root causes, evidence, regression guards)
agents/runtime_invariants.md — runtime invariants (INV-001…)
docs/70D_DATA_CONTRACT.md — 70D canonical feature contract
docs/RELEASE.md — release, installer, CLI, update & rollback guide

These are internal engineering artifacts: preserved as-is, linked, and not rewritten by the documentation effort.