skills.rest
qusong0627
qusong0627COMMUNITY
Explore all skills by this author →
GitHub

quantmind-operations

Operates QuantMind platform APIs for model training, data sync, inference, and RSS news analysis.

1.5k|337|Updated Apr 15, 2026
Search Tags:#model training#data sync#quantitative trading#api operations#batch inference#rss news#qlib
One-click install
npx skills add https://github.com/qusong0627/QuantMind --skill quantmind-operations-qusong0627
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: quantmind-operations
Source: https://github.com/qusong0627/QuantMind/tree/main/skills/quantmind-operations
Command: npx skills add https://github.com/qusong0627/QuantMind --skill quantmind-operations-qusong0627

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Operating the QuantMind quantitative trading platform requires knowing dozens of REST API endpoints for model training, data synchronization, inference, and news analysis. This Skill consolidates the full operational playbook—including authentication, environment detection between QuantBot containers and local hosts, and correct API call sequences—so users can run platform operations through natural language without memorizing endpoints. ## Core Features & Use Cases - Model Training & Management: Runs the 5-step training flow (feature selection, target definition, hyperparameters, execution, registry) across 13 model types including LightGBM, XGBoost, Transformer, and TabNet, plus rolling backtests and default-model management. - Multi-Market Data Operations: Triggers and monitors daily data sync for A-shares, HK, US, crypto, and futures markets, manages Qlib cache rebuilds, feature snapshots, and data freshness/quality checks. - Inference & RSS News Analysis: Executes single-day and batch inference (range/lookback modes), retrieves stock score history, and queries enriched RSS news with sentiment, event tags, and ticker filters. - Use Case: Ask the assistant to train a LightGBM model on momentum features for the A-share market, then check data freshness and pull bullish news for the top-scored stocks—all handled through the correct authenticated API calls. ## Quick Start Ask the assistant to check the current default model and data freshness on the QuantMind platform, then submit an incremental daily sync for the A-share market.

Frequently Asked Questions about quantmind-operations

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

FAQPage Schema
How do I train a model on the QuantMind platform?▼

Model training follows a 5-step flow: fetch the feature catalog, define the T+N target label, configure hyperparameters and time splits, submit a run-training request, then poll the training run status. Supported model types include lightgbm, xgboost, catboost, gru, lstm, transformer, and others.

How do I sync market data for A-shares, US, or crypto?▼

Submit a daily-sync task to the admin data-platform endpoint with the market code (A/CN, US, HK, BC, or FUTURES), then poll the task status by task_id. Each market uses its own data source such as QuantDB SDK, Yahoo Finance, or Binance.

What is the difference between range and lookback batch inference modes?▼

Range mode runs single-day inference for every trading day in a specified date interval. Lookback mode anchors on one date and infers backward over a window of N days, defaulting to the model's prediction horizon.

Why does the feature catalog return different categories over time?▼

Feature categories are dynamically generated from the QuantDB l1_factors dataset and change with each data version. Always read the categories from the feature-catalog API response instead of hardcoding a category list.

Can I run data scripts outside the quantmind Docker container?▼

Only pure standard-library scripts run locally. Scripts importing pandas, duckdb, psycopg2, numpy, sqlalchemy, or backend packages must be copied into the quantmind container and executed there via docker exec.

How do I filter RSS news by stock sentiment?▼

Query the news articles endpoint with tickers, sentiment (bullish, bearish, or neutral), event_tags, and a since timestamp. Use sort=sentiment_bullish to rank by positive signal strength or strong_only for high-confidence signals.

Related Skills

Explore More Data & Analytics Skills→
noahsilve123noahsilve123
1

huggingface-datasets

Validate Hugging Face datasets and retrieve metadata via the Dataset Viewer API.

Community
Advanced
maminul007maminul007

backtest

Execute backtesting simulations for trading strategies and generate performance reports.

Community
Intermediate
aliyunaliyun
606

alternative-factor-mining

Mines alpha factors from alternative data and validates them with IC and group-return tests.

Official
Advanced
xfs96192xfs96192

mx_select_stock

Filter global stocks by natural-language criteria into CSV output.

Community
Advanced