Skip to content

Roadmap

This page distinguishes delivered capability from planned work. It is not a date commitment.

✅ Delivered in v1.0

  • Five years of full-market A-share data with atomic updates and integrity checks.
  • Zero market-data I/O after preload and one TradingBus.place_order() path.
  • Point-in-time masking across daily, intraday, news, announcements, fundamentals, and sandbox outlets.
  • Relative dates and deterministic company pseudonyms that preserve board semantics.
  • Pre-market, progressive open-window, and progressive close-window agent phases.
  • Isolated multi-agent comparison and a reference read-only-advisor committee with one executor.
  • Full-fidelity exchanges, fail-closed fingerprint replay, and trajectory export.
  • Serialized-artifact leakage auditing.
  • Local FastAPI/React console, non-root container, PyPI packaging, and CI.

See the changelog for release-level details.

✅ Delivered in v1.1: daily paper trading

A daily paper session reuses the Agent loop, tool contract, conditional orders, and single execution path while a small live one-minute attention set drives the four decision windows. It supports Agent, virtual-capital, session-date, and live/accelerated-clock selection with persistent state and an event audit journal. See Daily paper trading.

It retains the same order path and risk checks and prevents tools or sandbox code from reading beyond the simulated clock.

📋 Planned: broker adapter

A future adapter boundary may connect a research agent to a real brokerage API after a credential, authorization, and order-risk threat model matches the project’s security policy. There is no broker integration today.

📋 Planned: hardened sandbox

The current sandbox protects one trusted researcher from accidentally reading canonical data or starting nested backtests. Future hardening may add resource isolation for third-party agent cards, narrower capability domains, and structured sandbox audit logs.

❌ Non-goals

  • A public multi-tenant hosted service in the current security model.
  • Market-impact modeling.
  • A prescribed trading methodology.
  • Agents interacting with each other or sharing fills in real time.

See extensions for contribution contracts.