Four Claude Code skills in one plugin that author SQX blocks, groups, templates, and projects from plain English. Each reads your install, generates buildable artifacts, and self-validates before import—no manual XML.
TL;DR
sqx-lab is a Claude Code plugin bundling four skills for StrategyQuant X / AlgoWizard: custom blocks → groups → templates → projects. Describe in plain English; each skill reads your SQX installation, catalogs what it supports, generates artifacts, and self-validates before import.
• 🧰 One plugin, four skills — full chain from atomic rule to scheduled build.
• 🗣️ Plain English in, importable XML /.sqx /.cfx out
• 🔍 Install-derived — uses only what's in your installation.
• 📁 One-command setup — /sqx-setup bootstraps all four skills.
• 🩺 Self-diagnosing — /sqx-doctor health-checks the whole chain.
• ✅ Self-validated — artifacts validated before completion.
• 🔗 Composable — each output feeds the next or stands alone.
Who this is for
• SQX users building custom blocks, curated pools, and thesis-driven templates.
• Anyone who has hand-edited XML and watched builds fail silently.
• Teams wanting repeatable workflows from paper idea to running build.
• Users of any standalone skill wanting the whole pipeline installed together.
Core principles
1. One chain, four layers
Block → group → template → project. Custom Block Builder authors rules; Random Group Builder pools them into menus; Strategy Template Builder wires pools into buildable scaffolds; Strategy Project Builder wraps templates into project.cfx. Each feeds the next or stands alone.
2. Install-derived, never guessed
Each skill bootstraps a catalog from your installation: indicators, custom blocks, clean groups, donor projects. Nothing from memory — if a key isn't in your catalog, the skill proposes an equivalent that is.
3. State your install folder once
/sqx-setup asks where StrategyQuant X is installed (the top-level folder with internal/ and user/) and bootstraps all four skills. The path is stored after validation. It survives plugin updates — run /sqx-setup again to point at a different install.
4. Proven structure, varied content
No hand-written SQX XML from scratch. Blocks use a grammar checked against your build; templates transplant into working skeletons; projects clone from donor projects in your install.
5. Validate before you import
Every skill validates—key resolution, type contracts, holes, structure—and won't declare success until it passes. SQX itself remains the final oracle, but the "silently wrong" class is caught first.
6. When something is off, the doctor says what and why
/sqx-doctor checks the whole setup: Python, install path, each catalog, and the block → group → template → project chain. It detects broken random groups referencing blocks not in your install. The doctor names the group, lists missing blocks, and Repair mode can rebuild them.
What's inside
sqx-custom-block: Atomic trading rules (Condition and Price-level blocks) as importable custom-block XML → feeds groups
sqx-random-group: Pools sampled by the builder (Condition and Value groups) from existing blocks or fresh atoms → feeds templates
sqx-strategy-template: Full .sqx strategy templates with typed holes bound to clean groups → feeds projects
sqx-strategy-project: Build projects (project.cfx) wiring templates as build tasks — data, timeframes, charts, analysis
Tutorial — install and first run
Step 1 · Download the plugin
Download the sqx-lab zip and unzip it. Requires Claude Code and Python 3.8+ (standard library only).
Step 2 · Install the plugin
In Claude Code, add the folder as a plugin and install:
/plugin marketplace add /absolute/path/to/sqx-lab
/plugin install sqx-lab@sqx-lab
All four skills are now available in every Claude Code session.
Step 3 · One-time setup: /sqx-setup
Run the setup command and point it at your StrategyQuant X folder (the top-level one containing internal/ and user/):
/sqx-setup
It validates the folder, stores the path for all four skills, bootstraps each catalog, and reports what it found. If anything behaves oddly later, run:
/sqx-doctor
It reports the Python interpreter, install, catalogs, and the block → group → template → project chain—including any broken random groups and how to repair them.
Step 4 · Ask for what you want, at any layer
Just describe the artifact:
"Build me a custom block: close above the upper Bollinger band, with its opposite."
"Group my breakout blocks into a Condition pool."
"Design a trend-filtered breakout template from my install."
"Make a build project that runs my gold templates on H1 with a daily regime chart."
Each skill shows a plain-English spec first, generates only after approval, validates, and hands you the file with a summary.
Step 5 · Import into SQX — that's the final oracle
Blocks and groups import into AlgoWizard; templates import and Build; projects deploy and run. If SQX rejects something, tell the skill what happened—it fixes the artifact. Most common cause: a referenced block not yet imported in that install.
Step 6 · Compose the chain
Author blocks from a paper → pool into themed groups → wire into thesis-driven templates → wrap into one project and let it build overnight. Every step stays inside what your install can do because every step reads the same catalog.
Updating
Download the latest zip and reinstall from the marketplace. Your stored path survives—run /sqx-setup once after to rebuild catalogs. Takes seconds.
Requirements
• StrategyQuant X / AlgoWizard — build 144
• Python 3.8+ — standard library only
• Claude Code — with sqx-lab plugin installed
• Your own SQX install — everything is derived from it
FAQ
Do I need all four skills?
No. Each works standalone. They get better together.
Will it work on my version of SQX?
It targets build 144. Everything is derived from your install, so it uses only what your build supports.
Does it send my strategies or data anywhere?
No. Local Python scripts read your local install. Catalogs and path are per-machine files that never leave your computer.
What if a generated file doesn't import or build?
Tell the skill what SQX said. Each fixes the artifact—usually a referenced block not yet imported—rather than starting over. If multiple things seem off, run /sqx-doctor first: it pinpoints the exact problem and the command that fixes it.
I'm Ivan Hudec—I built the free blocks you're using here and write about research at algotrading.space. Work with me.
