d8b44e0829
Task-by-task plan for the SP11 spec at HEAD 9395b983c. Three layers:
Layer A (3 commits, additive infrastructure, no behavior change):
A0 — 20 ISV slots [340..360) + 20 reset entries + novelty-hash arm
A1 — 3 canary producers (val_sharpe_delta, saboteur_engagement,
reward_component_grad_ratio) with Pearls A+D chained
A2 — controller kernel + SimHash novelty buffer; HEALTH_DIAG sp11_reward
Layer B (1 atomic commit, ~750 LOC):
B1 — every consumer migrates: cf_weight (mse + c51), audit-discovered
shaping sites, saboteur multiplier, replay-time curiosity bonus,
novelty hash lookup+update scheduled at replay
Layer C (validation + close-out):
C1-C2 — local + L40S smoke; T10 3×3×50 full validation
C3-C5 — Fix 39 audit doc, pearl_reward_as_controlled_subsystem, MEMORY.md
C6 — merge to main
Plan provides exact file paths, kernel signatures, GPU oracle test
skeletons, commit messages, and validation pass criteria from spec
§9.1-9.3. ~1550 LOC total estimate; ~3 hours subagent work plus
validation wall-clock (smoke ~25 min, T10 ~4 hr).
Spec: docs/superpowers/specs/2026-05-04-sp11-reward-as-controlled-subsystem.md
Foxhunt
Production HFT trading system in Rust.
Architecture
The workspace contains 32 crates organized as follows:
Core Libraries (16)
| Crate | Purpose |
|---|---|
trading_engine |
Order processing, FIX 4.4, IB TWS, SIMD, RDTSC timing |
risk |
VaR, Kelly, circuit breakers, kill switches, compliance |
risk-data |
Risk data types and shared structures |
trading-data |
Trading data types |
ml |
DQN Rainbow, PPO, TFT, Mamba2, ensemble inference |
ml-data |
ML data types and feature definitions |
data |
Market data ingestion and storage |
backtesting |
Replay engine, strategy tester |
adaptive-strategy |
Ensemble execution, microstructure analysis |
common |
Shared types, resilience, error handling |
storage |
S3 and local model storage |
model_loader |
Model serialization and loading |
market-data |
Market data feed handlers |
database |
PostgreSQL access layer (SQLx) |
config |
Configuration management |
tli |
CLI commands and tooling |
Services (8)
| Service | Purpose |
|---|---|
backtesting_service |
gRPC backtesting service |
broker_gateway_service |
FIX routing, broker connectivity |
trading_service |
Core trading operations |
ml_training_service |
Model training orchestration |
data_acquisition_service |
Market data acquisition |
trading_agent_service |
Autonomous trading agents |
api_gateway |
gRPC API gateway with auth |
web-gateway |
Axum REST + WebSocket gateway |
Frontend
web-dashboard/ -- React 19 + TypeScript + Vite + TradingView charts.
Building
# Check compilation (no PostgreSQL required)
SQLX_OFFLINE=true cargo check --workspace
# Run tests for a specific crate
SQLX_OFFLINE=true cargo test -p <crate> --lib
# Clippy
SQLX_OFFLINE=true cargo clippy --workspace
ML Models
Four production model architectures on Candle v0.9.1 with CUDA:
- DQN Rainbow -- Deep Q-Network with prioritized replay, dueling heads, noisy nets
- PPO -- Proximal Policy Optimization with GAE and LSTM policies
- TFT -- Temporal Fusion Transformer for multi-horizon forecasting
- Mamba2 -- State space model for sequence prediction
Each model has a standalone trainer and a UnifiedTrainable adapter for the hyperopt pipeline.
Infrastructure
- Git: Gitea at
git.fxhnt.ai(Tailscale-only), Scaleway DEV1-S - Observability: OpenTelemetry OTLP (env
OTEL_EXPORTER_OTLP_ENDPOINT) - Database: PostgreSQL with SQLx offline mode for CI
License
Proprietary. All rights reserved.