ACASH — Automated Capital Allocation System

---

Current Direction (2026-10-15)

Start here: the single active delivery plan is docs/ROADMAP.md (v4.0.0, Unified Delivery D1–D10: reuse existing capabilities from data foundation through human-gated capital admission).

• North Star: evidence-based capital allocation, including cash/no trade. A hypothetical USD 1,000 after-cost ETF comparison is the first practical study milestone, not the final objective.
• Hard boundaries: real capital USD 0.00; NO_REAL_ORDERS=true; Lane A acquisition UNISSUED; no provider HTTP, credentials, subscriptions, broker authentication, paper/live orders, or Homelab deployment without separate human authority.
• Subsystem map: src/acash/ (production domain modules) · tests/ (verification) · docs/README.md (documentation hub) · AGENTS.md (engineering and authority rules).
• Historical records (Phase 0–23, E3.6, session handoffs) are preserved evidence below and in docs/ — they describe past decisions, not current authorization.

---

1. Executive Summary & North Star

ACASH (Automated Capital Allocation System) is a serious, research-first, evidence-driven capital allocation and portfolio management platform.

ACASH is NOT a generic AI trading bot, NOT an indicator collection, NOT an MT5 Expert Advisor, NOT an LLM voting system, and NOT an unconstrained machine making "+1% daily" promises.

The North Star Question

"Given the current market, available opportunities, portfolio state, uncertainty, liquidity, and risk constraints, where should capital be allocated?"

The system explicitly supports the answer: "NOWHERE" (100% Cash / No-Trade Allocation).

Core Research Principle

$$\text{DATA} \to \text{EVIDENCE} \to \text{HYPOTHESIS} \to \text{RESEARCH} \to \text{ALPHA} \to \text{VALIDATION} \to \text{PORTFOLIO} \to \text{RISK} \to \text{EXECUTION} \to \text{OUTCOME} \to \text{FEEDBACK}$$

"DO NOT ASSUME AN EDGE. PROVE IT."

---

2. System Architecture & Decoupled Sovereign Layers

ACASH is built as a sovereign Modular Monolith in Python executing locally on a single workstation (AIO):

                                 ACASH SYSTEM CORE
                                        │
                        ┌───────────────┴───────────────┐
                        │                               │
               1. RESEARCH DATA LAYER          2. ANALYTICS & RESEARCH
             (Parquet + DuckDB + yfinance)    (pandas + NumPy + vectorbt + Plotly)
                        │                               │
                        └───────────────┬───────────────┘
                                        │
                                        ▼
                                  3. ALPHA ENGINE
                          (Signals & Hypothesis Specification)
                                        │
                                        ▼
                                4. VALIDATION ENGINE
                             (Purged CPCV & DSR Gate)
                                        │
                                        ▼
                              5. ALPHA QUALIFICATION
                       (Phase 8.5 Dossiers & Economic Lineage)
                                        │
                                        ▼
                               6. PORTFOLIO ENGINE
                      (Phase 8 Tournament vs Baselines: EW/InvVol/Cash)
                                        │
                                        ▼
                              7. RUNTIME SUPERVISOR
                     (Phase 10 5-Stage Orchestrator & Dual-Clock)
                                        │
                                        ▼
                                8. SOVEREIGN RISK
                     (Phase 9 Deterministic Veto & Kill Switch)
                                        │
                                        ▼
                               9. EXECUTION ENGINE
                      (Phase 7 Coordinator & Alpaca Paper)
                                        │
                                        ▼
                    10. APPEND-ONLY OPERATIONAL LEDGER
                      (Phase 10 SHA-256 Chained Disk Events)
                                        │
                                        ▼
                    11. FORWARD MONITORING & REALITY GAP
                     (Phase 11 Strategy Drift & Drag Attribution)
                                        │
                                        ▼
                    12. VENUE EXECUTION ADAPTERS
                     (Phase 12 MT5 BMAP & 6-D Reconciliation)
                                        │
                                        ▼
                    13. FORWARD PAPER VALIDATION
                     (Phase 13 Small Capital & Soak Harness)
                                        │
                                        ▼
                    14. AI RESEARCH & EVIDENCE LAYER
                     [PLANNED — PLAN APPROVED / IMPLEMENTATION LOCKED]

Sovereign Layer Responsibilities

1. Research Data Layer (Analytical): Partitioned Parquet files + embedded DuckDB query engine for vectorized analytical SQL queries + yfinance research data adapter.
2. Analytics & Quant Research: pandas, NumPy, vectorbt (Tier-1 rapid parameter screening), and Plotly (interactive visualization).
3. Alpha Engine & Hypothesis Contract: Strict pre-registered hypothesis contracts, econometric forward returns, and Newey-West / HAC inference.
4. Statistical Validation Engine: Combinatorial Purged Cross-Validation (CPCV), Deflated Sharpe Ratio (DSR), MinTRL, Holm-Bonferroni FWER multiple testing corrections, and Probability of Backtest Overfitting (PBO).
5. Alpha Qualification Engine (Phase 8.5): Canonical AlphaQualificationDossier and AlphaEconomicDecomposition DTOs certifying economic edge ($Net = Gross - Friction$) with $0.00 capital authority.
6. Portfolio Engine & Tournament (Phase 8): Native HRP and ERC optimizers evaluated strictly in zero-leakage tournaments against transparent baselines (Equal Weight, Inverse Volatility, Cash/NOWHERE).
7. Runtime Supervisor & Operational Scheduler (Phase 10): Dual-clock cadence coordination ($as_of_utc \neq wall_clock_utc$) and authoritative 5-stage cycle orchestration ($\text{Data} \to \text{Census} \to \text{Tournament} \to \text{Risk} \to \text{Admission}$).
8. Deterministic Risk Engine & Kill Switch (Phase 9): Non-negotiable sovereign risk boundary (DeterministicRiskEngine), exact monotonic derisking (EXACT_SCALE_DOWN), and multi-sig Ed25519TrustStore quorum reset control.
9. Execution Engine & Broker Mapping (Phase 7): Sovereign ExecutionCoordinator, vendor-agnostic broker semantic mapping (BMAP), and venue-pinned Alpaca Paper adapter.
10. Operational Ledger (Phase 10): Cryptographic SHA-256 chained disk ledger (OperationalLedger) with full crash recovery and tamper detection.
11. Forward Monitoring & Execution Reality (Phase 11): Independent observational plane tracking online strategy decay (StrategyForwardDriftEvidence) and empirical execution drag decomposition (ExecutionCostEvidence).
12. Venue Execution Adapters (Phase 12): Thin broker IPC adapters (NativeMT5Transport, MT5BrokerAdapter) with authoritative 6-D reconciliation and atomic preflight routing.
13. Forward Paper Validation & Small Capital (Phase 13): Real-world execution verification under strict $0.00 capital authority and 24-hour unattended soak monitoring.
14. AI Quantitative Research Layer (PLANNED): Evidence-bound research architecture (docs/phase14/phase14_master_research_architecture_plan.md). Approved at plan level; implementation strictly locked; not an execution or capital authority; all outputs are strictly unvalidated research proposals.

---

3. Sovereign Domain Entity Relationships

ACASH explicitly decouples state management from decision and execution flows:

        CAPITAL & PORTFOLIO STATE FLOW
        ──────────────────────────────
                 AccountState
                      │
                      ▼
               PortfolioState
                      │
                      ▼
                  Positions
                      ▲
                      │ (Immutable State Transitions: Fill -> NEW Snapshots)
                      │
        DECISION & EXECUTION FLOW
        ─────────────────────────
                    Signal
                      │
                      ▼
               TargetAllocation
                      │
                      ▼
                RiskAssessment
                      │
                      ▼
              OrderIntent / Order
                      │
                      ▼
                     Fill
                      │
                      └──────► State Transition ──► NEW Position
                                                ──► NEW PortfolioState
        CROSS-CUTTING AUDIT LINEAGE
        ───────────────────────────
               OperationalCycleEvent / DecisionRecord (Append-Only)
          ↳ Observed Market Inputs
          ↳ Alpha Qualification Dossier Hash
          ↳ Target Allocation Decision Hash
          ↳ Risk Evaluation Verdict
          ↳ Execution Manifest Digest
          ↳ Chained Event Digest (SHA-256)

---

4. Milestones & Quality Gates (Phases 0–14)

✅ Phase 0: Discovery & Architecture ──► [PASSED — Architecture & Decision Records Approved]
✅ Phase 1: Foundation & Domain Core ──► [GATE 1 PASSED — 27/27 Unit Tests, mypy clean]
✅ Phase 2: Data Ingestion & Integrity Engine ──► [GATE 2 PASSED — 57/57 Tests, mypy clean]
✅ Phase 3: Market Microstructure & PIT Feature Engine ──► [GATE 3 PASSED — 122/122 Tests, mypy clean]
   ├─ ✅ Phase 3A: Canonical Trades Domain (Time & Sales, Aggressor Side)
   ├─ ✅ Phase 3B: Canonical Order Book (L2 Depth Multi-Row Frames & Deltas, L3 MBO)
   └─ ✅ Phase 3C: Microstructure Feature Engine (VWAP, Volume Profile, Footprint, OBI, Micro-Price)
✅ Phase 4: Alpha Research Engine & Hypothesis Contract ──► [GATE 4 PASSED — 139/139 Tests, mypy clean]
✅ Phase 5: Backtesting Substrate & Simulation Engine ──► [GATE 5 PASSED — 200/200 Tests, mypy clean]
✅ Phase 6: Statistical Validation & Overfitting Controls ──► [GATE 6 PASSED — CPCV, DSR, MinTRL, PBO, Search Ledger, OOS Hard Gate]
✅ Phase 7: Live Execution & Broker Mapping ──► Gate 7 [PASSED — E-verified code path, P-001 accepted]
✅ Phase 8: Portfolio Engine (Native HRP/ERC Optimizers vs Baselines & Tournament) ──► Gate 8 [FROZEN @ e6f1d04]
✅ Phase 8.5: Alpha Research & Economic Evidence Engine (Dossier Gate & Lineage DTOs) ──► Gate 8.5 [FROZEN @ 9ce1365]
✅ Phase 9: Deterministic Risk Engine & Sovereign Kill Switch ──► Gate 9 [FROZEN @ 6bd40d8]
✅ Phase 10: Runtime Orchestration & Continuous Paper Operations (Supervisor + Scheduler + Ledger) ──► Gate 10 [FROZEN @ 3955bf6]
✅ Phase 11: Forward Tracking, Online Drift Detection & Execution Reality Attribution ──► Gate 11 [FROZEN @ 092a2b1]
✅ Phase 12: MT5 & Multi-Venue Execution Adapters ──► Gate 12 [COMPLETED & FROZEN @ 1e1d154]
✅ Phase 13 & Infrastructure: Small Capital & Soak ──► Gate 13 Steps 1–7 Certified; Gate G7 / Stage S11 CLOSED [PASS — Human Forensic Adjudication, Session E3.5-20260913-025117-de2762]; Steps 8–9 LOCKED
📐 Phase 14: AI Quantitative Research Layer ──► Gate 14 [PLAN APPROVED AT PLAN LEVEL — IMPLEMENTATION LOCKED]
⏳ Phase 15–22: Strategy Admission Standard & Lifecycle Sequence [HARMONIZED — ADR-023]
🏛️ Phase 23: Adaptive Multi-Horizon Strategy Architecture [ARCHITECTURAL RECORD — ADR-024/025]

Detailed Summary of Delivered Capabilities:

1. Phase 0 — Discovery & Architecture:
   • Evaluated 17 quantitative & infrastructure technologies across 10 engineering criteria.
   • Established 7 decoupled layers and 19 Architectural Decision Records (ADR-001 to ADR-019).
2. Phase 1 — Foundation & Domain Core:
   • Sovereign immutable domain models (src/acash/core/domain/) with strict Decimal finite arithmetic.
   • Pure state transitions for 8 signed-quantity position fill scenarios, spot cash flows, and zero Realized PnL double-counting.
   • Append-only decision audit ledger and structured telemetry logger with recursive secret redaction.
3. Phase 2 — Data Ingestion & Integrity Engine:
   • Canonical PyArrow schema (Decimal128(38,18), timestamp[us, tz=UTC]) in src/acash/data/schema.py.
   • Per-Stream Data Integrity Validator (integrity.py) enforcing event monotonicity and anomaly preservation without data mutation.
   • Provenance & Manifest Engine (provenance.py) with raw source SHA-256 and canonical logical SHA-256 invariant to Parquet compression and row ordering.
   • Parquet Storage Engine (storage.py) with strict 1:1 Ingestion Unit mapping, Recoverable Batch Commit Protocol (PREPARED $\to$ PART_PUBLISHED $\to$ COMMITTED), crash recovery, and orphan part quarantine.
   • DuckDB Point-in-Time qualification query layer with multi-source isolation and lookahead prevention.
4. Phase 3A — Canonical Trades Domain:
   • Canonical trades schema supporting nanosecond-precision trade records, aggressor side classifications, and trade conditions.
   • Length-prefixed binary serialization and logical SHA-256 hashing.
5. Phase 3B — Canonical Order Book Domain:
   • Canonical L2 Market-by-Price (MBP) multi-row frame snapshots and deltas, and L3 Market-by-Order (MBO).
   • Deterministic 5-tuple order reconstruction (exchange_time_utc, source_order_key, message_type_rank, stream_id, row_sub_index).
   • Explicit CLEAR level semantics (distinguishing zero-volume deletions from NULL clear operations).
6. Phase 3C — Microstructure Feature Engine:
   • Derived mathematical features: Session VWAP, Volume-Weighted Dispersion ($\sigma$), Volume Profile with POC lower-price tie-breakers, Value Area 70% bounds, Footprint Analytics (Stacked Imbalances, CVD, Absorption), and Depth-Weighted Micro-Price.
   • Dual-temporal point-in-time filtering ($T_{\text{event}} \le T_{\text{decision}} \land T_{\text{knowledge}} \le T_{\text{as_of}}$) preventing lookahead and revision leakage.
7. Phase 4 — Alpha Research Engine & Hypothesis Contract:
   • Formal, pre-registered HypothesisSpecification with explicit falsification criteria.
   • Discrete bar-indexed forward returns ($R(t,H) = \frac{P_{\text{close}, t+H} - P_{\text{open}, t+1}}{P_{\text{open}, t+1}}$) eliminating off-by-one ambiguities.
   • Primary econometric inference via OLS slope $\hat{\beta}_H$ under Newey-West / HAC covariance using Bartlett kernel with verified analytical reference vector.
   • Descriptive non-parametric association: Pearson IC, Spearman Rank IC (with fractional tie-handling), and Autocorrelation.
   • 3-Tier Friction Waterfall: Raw Predictive Edge $\to$ Spread + Fee Net $\to$ Fixed Slippage Proxy Economic Edge.
   • Interval-based boundary purging across partition splits and unallocated embargo buffers ($\ge \max(H)$ bars).
   • Durable Blind OOS Governance Ledger (data/manifests/research/governance_ledger.json) locking OOS exposure (UNEXPOSED $\to$ EVALUATED_LOCKED $\to$ EXHAUSTED).
8. Phase 5 — Backtesting Substrate & Simulation Engine:
   • Sovereign event-driven simulation substrate with simulated order lifecycle state machine (CREATED $\to$ SUBMITTED $\to$ ACCEPTED $\to$ FILLED).
   • Canonical Data Adapter enforcing Phase 3B total ordering 5-tuple: $(T_{\text{event_utc}}, \text{source_order_key}, \text{message_rank}, \text{stream_id}, \text{row_sub_index})$.
   • Decoupled double-entry shadow ledger (Balance-Sheet View vs Performance Attribution View) eliminating Realized PnL double counting ($|\text{AccountingResidual}| \le 10^{-10}$).
   • Unmocked native NautilusTrader execution substrate with Parquet catalog bridge and contract specification mapping.
   • Deterministic content-derived BacktestManifest identity: $\text{manifest_id} = \text{SHA256}(\text{canonical}(\text{hypothesis_hash} + \text{data_hashes} + \text{engine_hash} + \text{strategy_hash} + \text{seed}))[:32]$.
   • Reality Gap Telemetry Engine implementing disjoint non-overlapping reference-price decomposition: Spread Drag, Slippage Drag, Latency Drag, Fee Drag, Maker Adverse Selection Drag, and Unmodelled Residual.
   • Baseline strategy actors: Microstructure Imbalance (OBI) & Session VWAP Mean Reversion.
9. Phase 6 — Statistical Validation & Overfitting Controls:
   • Combinatorial Purged Cross-Validation (cpcv.py): Contiguous $N$-group partitioning, exhaustive $\binom{N}{k}$ combinatorial splits, strict $[t+1, t+H]$ interval purging, post-test embargo buffers, and chronological pseudo-OOS path reconstruction ($\phi = \frac{k}{N}\binom{N}{k}$).
   • Deflated Sharpe Ratio & MinTRL (deflated_sharpe.py): Non-normal asymptotic inference (Bailey & López de Prado 2014) with Euler-Mascheroni constant $\gamma_E$, empirical trial variance $V$, Fisher-Pearson skewness $g_1$, and Pearson kurtosis $g_2$.
   • Multiple Testing Corrections (multiple_testing.py): Holm-Bonferroni (FWER), Benjamini-Hochberg (FDR), and Harvey-Liu-Zhu (2016) Haircut Sharpe Ratio.
   • Probability of Backtest Overfitting & Fragility (overfitting.py): Mid-rank tie-breaking log-odds PBO, parameter sensitivity curvature over strict $[0.75\theta_0, 1.0\theta_0, 1.25\theta_0]$ grids, and component-wise friction stress decay monotonicity.
   • Sovereign Validation Gate (gate.py): Invariant trial intensity coupling ($K_{\text{ledger}} \equiv K_{\text{DSR}} \equiv K_{\text{Holm}} \equiv K_{\text{BH}}$), strict fail-closed OOS execution, and dual cryptographic lineage digests (evidence_digest and decision_digest).
10. Phase 7 — Live Execution & Broker Mapping (COMPLETED — P-001 Accepted, Gate 7 PASSED):
    • Admission/Authorization gate, Step 8 Execution Contract, Step 8B State Machine, Step 8C Broker Event Normalizer, Step 8D Mock Broker, Step 8E Execution Coordinator & Reconciliation Boundary, Operational Restriction, Real Broker Contract, and Vendor-Agnostic Broker Semantic Mapping Framework: LOCKED.
    • Broker Semantic Evidence (BMAP): E-reviewed against official broker API documentation (BMAP 01–10 = E, BMAP 11 = E*, BMAP 12 = D).
    • Concrete Components: Alpaca Paper Transport (PaperHttpAlpacaTransport), Paper Credential Boundary (venue-pinned ALPACA_PAPER), AlpacaPaperAdapter, R0 read-only harness, R1 lifecycle harness, and local Windows DPAPI vault launcher.
    • Paper Runtime Evidence: Checkpoint P-001 officially accepted with verified real-time fill parity.
11. Phases 8–11 — Portfolio, Research Qualification, Risk, Orchestration & Monitoring:
    • Completed, hardened, and frozen under sequential governance (Gates 8, 8.5, 9, 10, 11 passed).
12. Phase 12 — MT5 & Multi-Venue Execution Adapters (COMPLETED & FROZEN @ 1e1d154):
    • High-integrity broker connectivity for MetaTrader 5 with immutable symbol specs, Decimal volume quantization, tick-grid price alignment, Windows Local IPC transport, and authoritative 6-D reconciliation engine.
13. Phase 13 & Infrastructure Validation (Steps 1–7 Certified | Gate G7 / Stage S11 CLOSED):
    • Steps 1–7 certified; homelab continuous soak validated under Gate G7 / Stage S11 (Session E3.5-20260913-025117-de2762, adjudicated PASS by Human Forensic Adjudication: manifest 21,639.30s, 361 M1 bars spanning 360 intervals, zero failure-shaped feed disconnects). Pi verifier tooling remediation merged at commit 68142ae. Steps 8–9 strictly LOCKED ($0.00 capital authority; 0 real orders; broker disconnected).
14. Phase 14 — AI Quantitative Research Layer (PLAN APPROVED AT PLAN LEVEL — IMPLEMENTATION LOCKED):
    • Master Research Architecture Revision 1.2 approved at plan level by Human Auditor. AI outputs strictly classified as unvalidated proposals; implementation locked until Phase 13 Steps 5–9 are certified.

---

5. Documentation Index (docs/)

The complete canonical documentation suite is organized systematically in docs/:

🧭 System Governance & Roadmap

• docs/README.md: Master Documentation Hub and Progressive Disclosure Index.
• docs/ROADMAP.md: Unified Delivery D1–D10 (v4.0.0), existing capability reuse and preserved historical Phases 0–23.
• docs/PROJECT_STATUS.md: September 4 implementation snapshot (historical; current state lives in docs/ROADMAP.md).
• docs/DECISIONS.md: Canonical Architectural Decision Records (ADR-001 through ADR-025).
• docs/RISKS.md: Comprehensive Risk Register across quantitative, financial, operational, and technical dimensions.

🏛️ Core Architecture Specifications (docs/architecture/)

• docs/architecture/system_architecture.md: 7 decoupled layers, system dataflow, and performance hierarchy.
• docs/architecture/data_architecture.md: Analytical (Parquet+DuckDB) partitioned immutable storage & bi-temporal qualification.
• docs/architecture/data_contract.md: Canonical Market Data Contract, Decimal128(38,18), and Recoverable Batch Commit Protocol.
• docs/architecture/execution_architecture.md: Sovereign execution subsystem, order lifecycle state machine, and reconciliation loop.
• docs/architecture/portfolio_architecture.md: skfolio optimization vs transparent baselines (Equal Weight, Inv Vol, Cash).
• docs/architecture/research_architecture.md: Two-tier research engine, econometric inference, and reality gap decomposition.
• docs/architecture/technology_evaluation.md: 17-technology evaluation matrix across 10 engineering criteria.

📜 Historical Phase Plans & Proposals (docs/proposals/)

• docs/proposals/phase_1_foundation.md: Phase 1 Foundation & Domain Core Plan.
• docs/proposals/phase_3_microstructure.md: Phase 3 Market Microstructure & PIT Feature Engine Plan.
• docs/proposals/phase_3b_orderbook.md: Phase 3B Canonical Order Book & Reconstruction Proposal.
• docs/proposals/phase_3c_feature_engine.md: Phase 3C Microstructure Feature Extraction Engine Proposal.
• docs/proposals/phase_4_alpha_engine.md: Phase 4 Alpha Research Engine and Hypothesis Contract.
• docs/proposals/phase_5_simulation.md: Phase 5 Event-Driven Backtesting Substrate & Simulation Proposal.
• docs/proposals/phase_6_validation.md: Phase 6 Statistical Validation & Overfitting Controls Proposal.

🛡️ Phase 6: Statistical Governance (docs/validation/)

• docs/validation/methodology_contract.md: Master Methodology Contract (DSR, MinTRL, CPCV, Mid-Rank PBO, FWER).
• docs/validation/phase6_methodology_dgp_report.md: Data Generating Process (DGP) benchmark empirical experiments report.

⚡ Phase 7: Live Execution & Broker Integration (docs/phase7/)

• docs/phase7/phase_7_proposal.md: Phase 7 Proposal & Evolution Specification (Active Phase Architecture & Evidence Model).
• docs/phase7/CONTEXT_MAP.md: Phase 7 navigation context map and core invariants.
• docs/phase7/alpaca_bmap.md: Alpaca Concrete Broker Semantic Mapping (BMAP).
• docs/phase7/paper_exercise_r1.md: R1 order-lifecycle contract & P evidence checklist.
• docs/phase7/r1_paper_run_runbook.md: R1 Paper-run operational runbook.
• docs/phase7/execution_state_machine.md: Sovereign order & fill execution state machine.

🔌 Phase 12: MT5 & Multi-Venue Adapters (docs/phase12/)

• docs/phase12/closeout_report.md: Phase 12 Closeout Report & Frozen Contract Inventory.
• docs/architecture/multi_broker_multi_asset_decision.md: ADR-021 Multi-Broker Multi-Asset Architectural Specification.

🧪 Phase 13: Live Small Capital & Soak (docs/phase13/)

• docs/phase13/PHASE13-LIVE-SMALL-CAPITAL-PLAN-REV3.md: Master Operational Plan for Phase 13.
• docs/phase13/consolidated_gate_a_audit.md: Gate A Pre-Live Readiness Certification Pack.
• docs/phase13/recovery_runbook.md: Step 4 Rehydration and Crash Recovery Runbook.

🤖 Phase 14: AI Quantitative Research Layer (docs/phase14/)

• docs/phase14/phase14_master_research_architecture_plan.md: Phase 14 Master Research Architecture & Governance Plan (Rev 1.2 — Approved at Plan Level).

🌐 Project Atlas Knowledge Graph (docs/atlas/)

• docs/atlas/CONTEXT_MAP.md: Project Atlas Master Architecture & Microstructure Ontology.

---

5.5 Historical Status Snapshot — Post-Gate G7 / Stage S11 (2026-09-13)

Historical snapshot preserved verbatim except noted corrections. For current state see docs/ROADMAP.md (v4.0.0). Originally synchronized 2026-09-13 with docs/ROADMAP.md and Cheatsheet.md.

Project Maturity & Current State

• Phases 0–12: Complete, hardened, and frozen (Gates 1–12 certified).
• Phase 13 Infrastructure & Homelab Soak:
  • Steps 1–7: CERTIFIED.
  • Gate G7 / Stage S11: CLOSED / PASS (Session E3.5-20260913-025117-de2762, accepted by explicit Human Forensic Adjudication: manifest 21,639.30s, 361 M1 bars spanning 360 intervals, 0 restarts, zero failure-shaped feed interruptions, container memory samples accepted).
  • Pi verifier tooling remediation: MERGED (68142aecdccf636912902bbfc4061895c877d195 on Pi_Personal-Infrastructure/main).
  • Steps 8–9: STRICTLY LOCKED pending a qualified strategy and explicit Human GO authorization.
• Phase 14 (AI Quantitative Research Layer):
  • Gate 14 technical research capability accepted; pre-registration in progress for mainline CAND-FREE-MACRO-001 (resolving data-authority items D17/D18).
  • Implementation slices are strictly unvalidated proposals; zero trading, backtest, or execution authority.

Strict Safety Invariants & Zero-Real-Order Posture

$$\boxed{\mathbf{Research\ (14)} \neq \mathbf{Qualification\ (8.5)} \neq \mathbf{Allocation\ (8)} \neq \mathbf{Supervisor\ (10)} \neq \mathbf{Risk\ (9)} \neq \mathbf{Execution\ (12/7)} \neq \mathbf{Broker}}$$

• Canonical Safety Invariant: ZERO REAL ORDER SUBMISSIONS.
  • Distinguishes allowed simulated paper execution (ORDER_INTENT_CREATED, FILL_SIMULATED, portfolio accounting) from strictly forbidden real broker dispatches (ORDER_SUBMITTED == 0).
• Canonical Capital Authority: Strictly $0.00 (Hard-Locked).
• Real Order Submissions: Strictly 0 Orders (NO_REAL_ORDERS=true, simulated_fills_only=true).
• Broker Connection: Strictly DISCONNECTED.
• Paper Trading: NOT AUTHORIZED.
• Live Trading: LOCKED.
• Backtesting: LOCKED.
• HYP_003: TERMINALLY FALSIFIED (correction 2026-10-14; the 2026-09-13 snapshot said NOT CREATED; see sealed dossier docs/phase14/hyp_003_terminal_falsification_dossier.md).
• R1: NOT STARTED (2026-09-13 snapshot claim; verify current research records before acting).

Next Governance Transition

• NEXT GOVERNANCE TRANSITION: AWAITING HUMAN DECISION
• Canonical governance records: docs/ROADMAP.md, docs/operations/G7-SOAK-20260913-OBSERVATIONS.md, docs/phase14/research_execution_timing_note.md.

---

6. License & Proprietary Rights

Copyright © 2026 Ratthabhumi & ACASH Contributors. All Rights Reserved.

This software, including all underlying source code, documentation, schemas, algorithms, and mathematical models, is proprietary and confidential. Unauthorized copying, distribution, modification, public display, reverse engineering, or extraction of any part of this repository, via any medium, is strictly prohibited without explicit written consent.