xtquant
WarnAudited by ClawScan on May 10, 2026.
Overview
This is a coherent XtQuant/QMT SDK guide, but it enables live brokerage trading and account access without clear approval or safety boundaries.
Install only if you intend to let an agent help with XtQuant/QMT workflows. Treat this as a live-finance integration: verify the package source, keep QMT logged out unless needed, start with read-only market/account queries, and require explicit confirmation before any order, cancellation, fund transfer, or automated trading action.
Findings (4)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If the local QMT client is connected, an agent following the skill could submit real securities orders that affect the user's brokerage account.
This documents direct live order placement through the broker-facing trading API, with no accompanying requirement for explicit user approval, dry-run mode, or trade limits.
order_id = xt_trader.order_stock(account, '600000.SH', xtconstant.STOCK_BUY, 1000, xtconstant.FIX_PRICE, 10.5, 'strategy1', 'test_order')
Require explicit per-order user confirmation, account/stock/quantity/price limits, and a read-only or paper-trading mode unless the user deliberately enables live trading.
The agent may gain access to brokerage account state, assets, positions, orders, and trading permissions through the user's local QMT session.
The skill uses a local QMT trading profile and brokerage account object, which may inherit the user's logged-in trading session even though the registry metadata declares no primary credential or required config path.
path points to miniQMT's userdata_mini directory ... account = StockAccount('your_account_id') ... xt_trader.subscribe(account)Use a dedicated limited account or sandbox where possible, keep QMT logged out when not intentionally trading, and require explicit user authorization before any account-specific access.
A future or unexpected package version could behave differently from the documentation, including in trading workflows.
The skill depends on external Python packages using lower-bound, unpinned versions; this is normal for an SDK guide but important because the package can interact with financial accounts.
xtquant>=1.0.0; pandas>=1.5.0
Install only from the official provider, pin known-good versions, and review package provenance before using it with a live brokerage session.
A script using these examples may continue receiving data or trading callbacks until manually stopped.
The skill documents long-running callback loops for real-time data and trading notifications; this is expected for market data/trading but can keep an agent-controlled process active.
xtdata.run() # Block the current thread, continuously receiving callback data
Run these workflows only in supervised sessions, add explicit shutdown handling, and avoid leaving live trading callbacks running unattended.