Add execution layer: order lifecycle management

[IamJasonBian]

Summary

• Adds trading_system/execution/ — an execution layer between strategy signals and the broker that handles multi-source price discovery (Twelve Data + Gamma + Alpaca + broker), fill monitoring with crossover detection, staggered repricing within a 0.3% slippage budget, and Slack escalation (@jason Bian) when budget is exhausted
• Wires ExecutionManager into main.py init, run_once(), and process_signal() — orders are registered for lifecycle tracking after placement
• Adds execution_summary to blob logger for state persistence, send_crossover_alert() to Slack utils, and alpaca-py dependency for Alpaca price feed

New files

File	Purpose
execution/__init__.py	Package exports
execution/price_discovery.py	Aggregates prices from 4 sources, median consensus, outlier flagging
execution/fill_monitor.py	Detects fills, cancellations, crossovers (0.1% noise threshold)
execution/slippage_controller.py	3-step repricing: 0.05% → 0.15% → 0.30% cumulative
execution/execution_manager.py	Orchestrates submit → monitor → reprice → escalate
execution/order_types.py	SmartOrderType enum + TWAP stub

Test plan

• python -m trading_system.main -v --ticker BTC dry run completes cleanly
• Price discovery cross-check fires divergence warning (1.24% for BTC stop-limit vs consensus)
• execution_summary key present in state log JSON with pending contexts
• Live run: verify register_order_id attaches broker IDs to contexts
• Multi-tick run: verify crossover detection triggers repricing and Slack escalation

🤖 Generated with Claude Code

[IamJasonBian]

Need to confirm fillrate strat

[IamJasonBian]

Code Review — PR #32

Overall: Ambitious execution layer adding order lifecycle management between strategy signals and the broker. The architecture (price discovery → fill monitoring → slippage control → escalation) is well-structured.

Concerns

1. Large surface area — 6 new files + integration into main.py. This PR has been open since Feb 27 and the base is stale. The longer it sits, the harder the merge. Consider whether it can be broken into smaller PRs (e.g., price discovery first, then fill monitoring, then slippage control).

2. alpaca-py is now a hard dependency — Moved from optional comment to required in requirements.txt. This means all environments (including publish-only Render deploys from PR Add --publish flag for Render publish-only deployment #52) must install it. Verify this is acceptable.

3. Test plan has unchecked items — The last two items (live run: register_order_id, multi-tick: crossover detection) are still unchecked. Are these blocking or nice-to-have?

4. Stale base — Base SHA is from Feb 27. Significant changes have landed since (drift comparator, PDT gate, etc.). Recommend rebasing before review to ensure compatibility.

---

Generated by Claude Code

[IamJasonBian]

Stale PR notice — This PR has been open for 44 days with no new commits since Feb 27. The prior review flagged the large surface area (6 new files + main.py integration), a stale base SHA, unchecked live-run test items, and alpaca-py becoming a hard dependency that affects publish-only deploys in #52. Given the size and age, consider splitting into smaller PRs (price discovery → fill monitor → slippage control) or rebasing and completing the live-run test plan.

---

Generated by Claude Code

[IamJasonBian]

Stale — 8 weeks, no activity since Feb 27. This execution layer PR has 6 new files and unchecked live-run test items. With PR #53 (Executor pattern) and PR #55 (IBKR client) also open, the execution architecture is fragmenting across multiple stale branches. Recommend deciding on a single execution approach, closing superseded PRs, and rebasing the chosen one.

---

Generated by Claude Code