Find the perfect skill
ICT/SMC Trading Knowledge Skill
Data & AI
ICT/SMC/ChanLun trading knowledge base (726 concepts, 1282 cases) with four interaction modes: concept Q&A, chart analysis, annotation, and K-line analysis via Mobius Quant API.
TDA Figure Specification
Data & AI
Scaffolds matplotlib figure code following established TDA visualization conventions. Uses predefined style constants for consistent publication-ready figures.
Playwright Web Navigator
Data & AI
Navigate dynamic JavaScript-heavy sites, take screenshots, and extract content using a headless browser. Use when standard requests fail.
Itô Training
Data & AI
Inspect ML training availability on completed Itô compute bookings and hand off a confirmed training manifest when the canonical backend is available. Chains off ito-compute and never runs its own training stack.
inference.sh CLI for AI Apps
Data & AI
Run 150+ AI apps including image generation, video creation, LLMs, search, and more via the inference.sh CLI. No GPU required.
Distill an ex into an AI Skill
Data & AI
Import chats, photos, generate memories and an evolving persona from an ex.
Vector Index Tuning
Data & AI
Guide to optimizing vector index performance for latency, recall, and memory. Covers HNSW tuning, quantization strategies, and scaling.
TaxHacker Receipt Analysis
Data & AI
Extracts structured financial data from receipt/invoice images using AI.
Kaggle Competition Automation
Data & AI
End-to-end workflow for Kaggle competitions, from validation to submission, optimized for Codex. Includes rule analysis, modeling, multi-notebook architectures, and score optimization.
GitHub Quantitative Trading Project Analysis
Data & AI
Search and analyze popular quantitative trading projects on GitHub, providing recommendations on tech stacks, architectures, and best practices.
Predictive Persona Performance
Data & AI
Predict persona performance in specific task contexts.
NotebookLM Knowledge Base Organizer
Data & AI
Prepares files for NotebookLM by scoring and merging sources, converting formats, and optimizing RAG retrieval within the 50-source limit.