Faraone-Dev/atomic-mesh

Added a new backtest report for the ATOMIC MESH engine, detailing performance metrics such as 98 round-trips, a 100% win rate, and a realized PnL of +$15.68 on $1.4M volume. The report includes specific engine parameters (e.g., half-spread, gamma) and notes that 11 bugs were fixed in the C++/Rust pipeline to achieve positive PnL. An accompanying CSV file provides the time-series of realized PnL and position quantities for the 13,334 synthetic BTCUSDT events.

results · high confidence

The system now supports live market data and order execution via Binance. A new \ExecutionGateway\ handles submitting and cancelling orders to the Binance REST API, while the \BinanceConnector\ manages WebSocket streams for depth and trade updates. The connector now supports a testnet mode for safe testing, and the normalizer has been updated to correctly parse symbols and convert float prices/quantities into integer-based internal types. Additionally, the QUIC mesh transport layer was updated to support heartbeat, event replication, and consensus vote messages, with corresponding unit tests added for the protocol.

crates/atomic-feed · high confidence

The \crates/atomic-strategy\ crate now implements the Avellaneda-Stoikov market-making strategy. This adds a new \MarketMaker\ struct that generates bid/ask quotes around a volume-weighted microprice, dynamically adjusting the spread based on trade-flow toxicity (VPIN) and short-term volatility. The strategy includes an \InventoryManager\ that tracks position and computes quote skew to naturally offload inventory, while a \ToxicityTracker\ detects informed order flow to widen spreads or pull quotes when conditions become toxic. The implementation also includes a \microprice\ module for calculating fair value from order book depth.

crates/atomic-strategy · high confidence

The \atomic-hotpath\ crate now provides a high-performance market-making engine backed by a C++ implementation, accessible via Rust FFI. This replaces previous Rust-based order book and strategy computation with cache-aligned flat arrays and SIMD-optimized microprice calculations, enabling nanosecond-scale hot-path processing. The change includes a new \build.rs\ script to compile the C++ source files with specific optimization flags (AVX2, LTO, etc.) and exposes a \HotPathEngine\ struct that wraps the C++ engine, offering methods like \on\_book\_update\, \on\_trade\, and \on\_fill\. Additionally, comprehensive benchmarks are added to measure the performance of the hot path, including steady-state updates, trade processing, and end-to-end tick cycles.

crates/atomic-hotpath · medium confidence

A new C++17 FFI header and implementation provide a nanosecond-scale market-making hot path. The engine processes book updates and trades to generate bid/ask commands, computing microprice, VPIN-based toxicity, and inventory skew. It uses a lock-free ring buffer for inter-thread communication and supports a runtime flag to enable or disable VPIN toxicity tracking.

cpp · high confidence

Users can now generate synthetic BTCUSDT backtest data via a new binary, and access a live dashboard (default port 3000) that streams real-time metrics, order book updates, and strategy status over WebSocket. The dashboard also exposes a kill-switch API allowing operators to stop all orders, cancel all open positions, or disconnect the node from the mesh, with the main loop checking for these actions. Configuration for the market-making strategy (spread, gamma, inventory limits) and exchange gateway (testnet/live mode, API keys) is now explicit in the node config.

crates/atomic-node/src · high confidence

A new dashboard page has been added that provides a dark-themed, responsive interface for real-time monitoring. The update introduces a complete UI structure including a top bar with cluster status indicators, a sidebar for navigation, and multiple content panels for displaying KPIs, latency metrics, order book data, and event streams. It also includes visual components for risk management and emergency kill-switches, along with toast notifications and price flash animations to support live data updates.

dashboard · high confidence

The SmartOrderRouter now supports multiple routing algorithms including BestVenue, VWAP, TWAP, and LiquiditySweep, allowing orders to be split across venues based on price, depth, or time. The implementation includes a test suite verifying that buy orders select the lowest ask, sell orders select the highest bid, and volume/time-weighted algorithms split orders proportionally to liquidity.

crates/atomic-replay, crates/atomic-router · medium confidence

The risk engine now enforces three new pre-trade controls: a maximum spread limit to prevent quoting into adverse selection, a circuit breaker that pauses trading after a configurable number of consecutive losses, and a drawdown limit that triggers a soft pause when PnL drops below a peak threshold. These are configurable via the \RiskLimits\ struct and tracked in the \RiskEngine\ state, with daily resets clearing the circuit breaker and loss counters.

crates/atomic-risk · medium confidence

The simulated exchange now enforces strict liquidity checks for Fill-or-Kill and Market orders, rejecting them immediately if insufficient liquidity is available. It also introduces support for Immediate-or-Cancel (IOC) orders, which are automatically cancelled if unfilled, and properly handles resting limit orders. Additionally, the execution engine now tracks stale orders that have been acknowledged but show no fill or cancel activity for a specified duration, aiding in the detection of stuck orders.

crates/atomic-execution · medium confidence

The \data/equity.csv\ file has been updated with a new dataset containing sequence numbers, realized PnL in USD, and position quantities. This change corrects the backtest data, which the commit message indicates fixes 11 bugs in the C++/Rust backtest pipeline, resulting in a +0.68 adjustment to the PnL.

data · medium confidence

Added new tests in crates/atomic-node/tests to validate the trading pipeline's behavior. The determinism test ensures that two independent pipeline instances processing the same event stream produce identical state hashes, confirming full determinism across the C++ engine, simulator, and execution layers. The end-to-end integration test verifies that synthetic events flow correctly through the HotPathEngine, SimulatedExchange, and RiskEngine, ensuring the pipeline produces fills and a non-zero state hash.

crates/atomic-node/tests · high confidence

Introduces the new \atomic-hotpath\ crate, which provides a C++ hot-path FFI for nanosecond-scale order book and strategy processing, including a \cc\ build dependency for C++ compilation. The crate is integrated into the workspace and added as a dependency to \atomic-node\ and \atomic-strategy\. Additionally, \criterion\ and its associated dependencies (e.g., \ciborium\, \half\, \crossbeam\) are added to enable benchmarking capabilities, with \criterion\ also added as a dev-dependency for the \atomic-hotpath\ crate.

(dependencies) · medium confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

• CAI 64 → 67 (+2.7)
• Rubric changed (rubric-2026.08.18 → rubric-2026.09.15) — scores are not directly comparable.

• Code Health 89 → 88 (-0.2)
• Architecture 100 → 96 (-3.6)
• Maturity 65 → 64 (-1.6)
• Readiness 64 → 69 (+4.9)
• Security 63 → 92 (+29.1)
• Domain Modelling 100 → 100 (+0.0)
• Event Sourcing 100 → 100 (+0.0)
• Accessibility 59 → 59 (+0.0)

• Coverage not included — suite not readable by the collector
• Dependency hygiene not measured — dependency manifest found but not parsed for hygiene
• High CVE: [GHSA redacted] (Cargo.lock)
• High CVE: [GHSA redacted] (Cargo.lock)
• High vulnerability: [GHSA redacted] (Cargo.lock)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• Medium advisory (unmaintained): RUSTSEC-2025-0141 (Cargo.lock)
• Medium advisory (unsound): RUSTSEC-2026-0097 (Cargo.lock)
• Medium advisory (unsound): RUSTSEC-2026-0190 (Cargo.lock)
• Medium vulnerability: RUSTSEC-2026-0204 (Cargo.lock)
• …and 6 more

• BinanceConnector::connect (cognitive 29) (crates/atomic-feed/src/binance.rs)
• Documentation: no installation or build instructions (README.md)
• Documentation: no usage examples (README.md)
• Duplicated block (12 lines × 3) (crates/atomic-node/src/main.rs)
• Duplicated block (15–16 lines × 2) (crates/atomic-feed/src/gateway.rs)
• Duplicated block (15–16 lines × 2) (crates/atomic-feed/src/gateway.rs)
• Duplicated block (17 lines × 2) (crates/atomic-node/src/main.rs)
• Duplicated block (4–15 lines × 3) (crates/atomic-node/src/main.rs)
• Duplicated block (4–5 lines × 3) (crates/atomic-node/src/main.rs)
• Duplicated block (5 lines × 2) (crates/atomic-node/src/dashboard.rs)
• Duplicated block (6 lines × 2) (crates/atomic-core/src/types.rs)
• Duplicated block (6 lines × 2) (crates/atomic-core/src/types.rs)
• Duplicated block (8 lines × 2) (crates/atomic-node/src/main.rs)
• Duplicated block (9 lines × 2) (crates/atomic-node/src/main.rs)
• Duplicated block (9 lines × 3) (crates/atomic-feed/src/gateway.rs)
• FunctionTooLong: atomic_node::main (crates/atomic-node/src/main.rs)
• FunctionTooLong: atomic_node::run_backtest (crates/atomic-node/src/main.rs)
• High CVE: [GHSA redacted] (Cargo.lock)
• High CVE: [GHSA redacted] (Cargo.lock)
• High CVE: [GHSA redacted] (Cargo.lock)
• …and 35 more

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

Faraone-Dev/atomic-mesh was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.