Project Plan

Phase 0: Repo & Environment

• Confirm Python version target (3.11+ recommended).
• Add pyproject.toml and dev tooling (ruff, black, mypy).
• Define .env template and secrets handling.

Phase 1: Data Backbone (No Trading)

• Implement Kalshi + Polymarket clients (REST + WS).
• Normalize data into a shared schema.
• Publish signals to RabbitMQ.
• Emit UI updates over socket.io.
• Persist market mappings and snapshots to Postgres.
• Deliverable: live data feed and normalized signals with audit trail.

Phase 2: Execution Backbone (Dry Run)

• Implement order manager (Kalshi + Polymarket).
• Position tracking and P&L accounting in Postgres.
• Circuit breaker and risk limits.
• Dry-run execution using synthetic signals.
• Deliverable: end-to-end pipeline without live orders.

Phase 3: Live Execution (Controlled)

• Enable live trading behind feature flags.
• Add idempotency, retries, and failure recovery.
• Monitor latency, error rate, and P&L drift.
• Deliverable: live execution with guardrails.

Phase 4: Strategy Expansion (Inkwell)

• Integrate Inkwell for scoring/ranking.
• Keep execution rules separate and deterministic.
• Deliverable: configurable strategy with safe execution.

Open Decisions

• Final signal schema and message routing keys.
• Orderbook retention policy and storage cost bounds.