arctic-project

byابو اسيل 2

طورها لتكون المنصه الاولى عالميا

No preview

Comments (0)

No comments yet. Be the first!

System Requirements

Page 1 of 30

System Requirements Document

1. Introduction

1.1 Purpose

SkyLink V6.0 is an Arabic-first cryptocurrency futures analysis, research, and Binance USDⓈ-M Futures Testnet decision platform. It combines deterministic market analysis, constrained AI advisory consultation, strategy research, Testnet-only execution controls, operational monitoring, and data-governance gates.

The platform is intended to progress toward a globally leading standard while preserving safety, explainability, authorization controls, source transparency, and fail-closed behavior.

Page 2 of 30

1.3 Global Leadership Direction

SkyLink V6.0 must be developed toward becoming the world's leading platform in its category (Arabic-first governed crypto-futures analysis, research, and Testnet decision support). Global leadership is defined by the following measurable pillars, and every pillar must remain subordinate to the safety, authorization, and fail-closed constraints defined in this document:

  1. Decision safety and auditability — deterministic risk authority, explicit approval/block reasons, and complete audit traceability for every decision and execution attempt.
  2. Transparency of provenance — every market, news, model, and persisted-data item displays its source, freshness, and authorization state.
  3. Research discipline — conservative backtesting assumptions, out-of-sample validation, data-quality grading, and temporary Testnet-only promotion.
  4. Operational rigor — reconciliation, protection verification, kill-switch control, rate limiting, and bounded automation.
  5. Arabic-first excellence — complete Arabic coverage of decisions, controls, explanations, and operational status, with correct mixed-direction formatting for symbols, numbers, model names, API labels, and timestamps.
  6. International competitiveness — the platform must remain understandable, verifiable, and operable for international reviewers without weakening the Arabic-first experience.

Global-leadership ambition must never be used to justify profit guarantees, mainnet/live execution, autonomous trading, or bypassing any authorization gate.

1.2 Product Principles

  • Arabic is the primary language for decisions, controls, explanations, and operational status.
  • The system is advisory and experimental; it must not guarantee profits.
  • Deterministic risk controls have final authority over AI and council output.
  • Binance USDⓈ-M Futures Testnet is the only supported execution environment.
  • Binance mainnet and live execution are out of scope.
  • Market-data provenance, freshness, authorization, and execution eligibility must be visible.
  • Missing authorization, stale data, proxy-only data, reconciliation failures, protection failures, or ambiguous exchange state must fail closed.
  • Privileged provider access, retained-data display, research eligibility, and execution remain separately gated.
Page 3 of 30

2. System Overview

SkyLink V6.0 provides the following platform capabilities:

  1. Arabic decision workspace for market analysis, risk checks, council decisions, positions, orders, strategies, automation, and system status.
  2. Real-time and historical market-data layer with Binance Futures Testnet preference and transparent fallback labeling.
  3. Market-universe scanner for USDT and USDC futures contracts, including liquidity, volatility, funding, basis, spread, and listing-related indicators.
  4. Deterministic trading-decision engine using technical indicators, market microstructure conditions, news, account state, and risk rules.
  5. Six-seat AI advisory council with optional Anthropic, OpenAI, Google, xAI, Alibaba, and Moonshot integrations.
  6. Backtesting and strategy promotion for four defined strategies using conservative trading assumptions.
  7. Safe simulated trading and controlled Testnet execution with strict authorization, reconciliation, protection, and token controls.
  8. Automation supervisor supporting monitoring, simulation, and gated Testnet operation.
  9. Data-foundation and compliance layer for provider status, authorization gates, data categories, onboarding, readiness, truth-center information, and lineage readiness.
  10. Source-bundle integrity awareness including original source-byte scope and source-file dependency visibility.

3. Functional Requirements as Story Points

3.1 Platform Direction and Arabic Experience

  • FR-001 — 5 points: As a crypto-trading analyst, I want to use an Arabic-first decision platform so that market analysis, risks, decisions, and operational feedback are understandable in Arabic.
  • FR-002 — 3 points: As a platform stakeholder, I want SkyLink to be developed toward globally leading platform quality so that its decision safety, transparency, research discipline, and operational rigor support international competitiveness.
  • FR-002a — 5 points: As a platform stakeholder, I want the platform to be developed toward becoming the world's leading platform in its category so that Arabic-first governed crypto-futures analysis, research, and Testnet decision support set the international benchmark.
  • FR-002b — 3 points: As a platform stakeholder, I want global-leadership progress to be measured against the six defined pillars (decision safety and auditability, provenance transparency, research discipline, operational rigor, Arabic-first excellence, and international competitiveness) so that "globally leading" is verifiable rather than promotional.
  • FR-002c — 5 points: As a risk controller, I want global-leadership ambition to never weaken safety constraints so that no leadership goal can justify profit guarantees, mainnet/live execution, autonomous trading, or bypassing any authorization gate.
  • FR-003 — 3 points: As a user, I want the platform to clearly identify Binance USDⓈ-M Futures Testnet as the operating environment so that I understand that production/mainnet execution is not supported.
  • FR-004 — 3 points: As a user, I want execution mode labeled as TESTNET or SAFE_SIMULATION so that I can distinguish exchange execution from local simulated trading.
  • FR-005 — 3 points: As a user, I want the platform to avoid profit guarantees so that all recommendations are presented as experimental and risk-controlled decisions rather than financial promises.
Page 4 of 30

3.2 Source Bundle and Delivery Integrity

  • FR-006 — 2 points: As a release reviewer, I want the platform source bundle to preserve its original source-byte scope so that the delivered source inventory can be verified against the declared original source bytes.
  • FR-007 — 2 points: As a release reviewer, I want source bundle files to be byte-counted and SHA256-verifiable so that source integrity can be audited.
  • FR-008 — 2 points: As a release reviewer, I want the source bundle scope to include Git-tracked source, UI, scripts, tests, SQL, workflows, environment templates, ignore files, and dependency definitions so that delivery coverage is explicit.
  • FR-009 — 2 points: As a release reviewer, I want Git internals, runtime data, research evidence, logs, reports, datasets, caches, and generated deliverables excluded from the source-bundle scope so that source delivery boundaries are controlled.
  • FR-010 — 2 points: As a build maintainer, I want the implementation to retain the from engine import atr dependency relationship so that backtesting uses the deterministic Average True Range calculation defined by the market-analysis engine.
  • FR-011 — 2 points: As a build maintainer, I want the implementation to retain import zipfile and from datetime import datetime dependencies so that official Binance historical ZIP archives can be downloaded, parsed, and dated safely.

3.3 Platform Status, Gates, and Readiness

  • FR-012 — 3 points: As a user, I want to view platform health and status so that I can see engine version, operating mode, safe-preview state, connectivity, execution state, and server time.
  • FR-013 — 3 points: As a user, I want to view separated public-connectivity and privileged-capability states so that public market access is not confused with privileged provider or execution access.
  • FR-014 — 5 points: As a compliance reviewer, I want the platform to expose the current data-authorization decision so that authorization-dependent capabilities remain traceable.
  • FR-015 — 5 points: As a compliance reviewer, I want the platform to fail closed when the data-authorization decision is not APPROVED_DATA_SCOPE_READY so that privileged provider access cannot activate prematurely.
  • FR-016 — 3 points: As a user, I want to view the gate registry so that I can see registered gates, the active privileged gate, and the next allowed gate.
  • FR-017 — 3 points: As a user, I want to view public-development mode status and its latest metadata-only diagnostic so that I can distinguish development connectivity from authorized data operations.
  • FR-018 — 5 points: As a compliance reviewer, I want the platform-readiness view to remain provider-free so that it does not contact exchanges, start models, open retained catalogs, or create authorization decisions.
  • FR-019 — 3 points: As a user, I want to view the authorization workflow stages—REGISTERED, AUTHORIZATION_APPROVED, SCOPE_APPROVED, RETRIEVAL_VERIFIED, STORAGE_PERMITTED, DATA_QUALITY_PASSED, LINEAGE_COMPLETE, and RESEARCH_ELIGIBLE—so that readiness progression is transparent.
  • FR-020 — 3 points: As a user, I want safe interface capabilities to be distinguished from blocked capabilities so that I know what can operate without market data or privileged access.
  • FR-021 — 3 points: As a user, I want blocked actions to return a DATA_AUTHORIZATION_REQUIRED response so that the reason for denial is explicit.
Page 5 of 30

3.4 Safe Preview and Access Gating

  • FR-022 — 5 points: As an operator, I want a safe-preview mode so that market-data, model, provider, trading, and action APIs fail closed when preview safety is enabled.
  • FR-023 — 3 points: As a user, I want safe-preview health responses to identify disabled streams and locked exchange positions so that the safe state is visible.
  • FR-024 — 3 points: As an operator, I want provider access to require an approved data scope, explicit provider-access enablement, and disabled safe-preview mode so that privileged provider paths are tightly controlled.
  • FR-025 — 3 points: As an operator, I want API actions to require a separate explicit enablement switch so that read access and mutating actions are independently controlled.
  • FR-026 — 3 points: As a compliance reviewer, I want persisted-data display to require a separate explicit gate so that rights-sensitive retained data is not displayed by default.
  • FR-027 — 2 points: As an operator, I want preview embedding to require explicit enablement so that cross-origin preview framing is not enabled by default.
Page 6 of 30

3.5 Market Data Acquisition and Source Transparency

  • FR-028 — 5 points: As a trading analyst, I want the system to retrieve Binance Futures Testnet candles when available so that analysis can use the preferred futures source.
  • FR-029 — 3 points: As a trading analyst, I want the system to retrieve recent spot klines from the official Binance public market-data endpoint so that recent market analysis can continue when Futures Testnet candles are unavailable.
  • FR-030 — 3 points: As a trading analyst, I want recent spot klines to support validated USDⓈ-M-style symbols and supported intervals so that fallback retrieval remains constrained to expected market formats.
  • FR-031 — 3 points: As a trading analyst, I want recent spot klines to be limited to a controlled range of 100–1,000 candles so that fallback data retrieval is bounded.
  • FR-032 — 3 points: As a trading analyst, I want recent spot klines to be cached for short periods so that repeated market requests do not unnecessarily contact the public provider.
  • FR-033 — 3 points: As a trading analyst, I want official Binance spot candles to be labeled as a display-only proxy so that they cannot be mistaken for Futures Testnet execution data.
  • FR-034 — 3 points: As a trading analyst, I want the system to use Kraken candles as a labeled display-only fallback so that the dashboard remains informative when Binance sources are unavailable.
  • FR-035 — 3 points: As a trading analyst, I want the system to use local synthetic candles only as a clearly labeled final fallback so that unavailable market connectivity does not appear to be real market data.
  • FR-036 — 3 points: As a user, I want every market-data fallback to display its source and warning so that I understand whether the information is direct Futures Testnet, proxy, or simulated.
  • FR-037 — 5 points: As an execution controller, I want Testnet execution to be rejected when the underlying decision is based on a fallback source rather than direct Binance Futures Testnet data.
  • FR-038 — 3 points: As a research analyst, I want historical data retrieval to prefer official Binance USD-M Futures archives so that backtesting uses longer completed historical series where available.
  • FR-039 — 3 points: As a research analyst, I want the system to retrieve and parse historical candle data from daily Binance ZIP archives so that completed immutable market periods can be used in research.
  • FR-040 — 3 points: As a research analyst, I want historical ZIP archive timestamps to support milliseconds and microseconds formats so that archive-format variation does not corrupt candle timing.
  • FR-041 — 3 points: As a research analyst, I want historical ZIP archive retrieval to skip missing daily archives but fail for other retrieval errors so that incomplete availability is handled safely.
  • FR-042 — 3 points: As a research analyst, I want historical fallback sources to be identified as Futures Testnet, Binance spot historical proxy, Kraken historical proxy, or local historical simulation so that test results remain interpretable.
  • FR-043 — 3 points: As a research analyst, I want supported historical intervals to include 1m, 5m, 15m, 30m, and 1h so that analysis and backtesting can operate across supported timeframes.
  • FR-044 — 3 points: As a researcher, I want the system to assess candle-data quality using candle count, completeness, gaps, time span, first timestamp, last timestamp, and grades A through D so that research suitability is measurable.
  • FR-045 — 3 points: As a research analyst, I want strategy promotion to require data quality grade A or B so that weak or incomplete candle history cannot be promoted.
Page 7 of 30

3.6 Real-Time Market Stream and Market Universe

  • FR-046 — 5 points: As a user, I want real-time Binance Futures Testnet mark-price streaming where connectivity is available so that displayed market prices can be current.
  • FR-047 — 3 points: As a user, I want the stream to provide stream connection state, source, last-event age, latency, errors, and reconnect count so that market-feed quality is visible.
  • FR-048 — 3 points: As a user, I want stale stream events older than ten seconds to be treated as disconnected so that old quotes are not represented as live.
  • FR-049 — 3 points: As a user, I want Kraken WebSocket prices to be labeled as display-only fallback quotes so that they cannot be mistaken for Binance execution prices.
  • FR-050 — 3 points: As a trading analyst, I want market details to include mark price, index price, funding rate, basis, next funding time, bid price, ask price, spread, and top-book notional when fresh so that futures conditions can be assessed.
  • FR-051 — 3 points: As a trading analyst, I want stale market-detail fields to be omitted when they exceed freshness thresholds so that outdated funding, mark, and order-book values do not influence decisions.
  • FR-052 — 5 points: As a market scanner, I want to view a universe of USDT and USDC futures symbols so that I can identify tradable market candidates.
  • FR-053 — 3 points: As a market scanner, I want to filter the market universe by symbol search, minimum volume, minimum change, maximum range, maximum funding, maximum basis, maximum spread, direction, quote asset, and result limit so that I can focus on suitable candidates.
  • FR-054 — 3 points: As a market scanner, I want the universe to identify newly listed contracts using exchange onboarding metadata and stream-baseline logic so that recent listings can be reviewed.
  • FR-055 — 3 points: As a market scanner, I want exchange metadata to provide contract status, onboarding date, contract type, and quote asset when available so that universe candidates can be evaluated accurately.
  • FR-056 — 3 points: As a user, I want a ranked opportunity scan for configured symbols so that I can compare direction, quality, price, RSI, trend, volume ratio, volatility, and source.

3.7 News Collection and Sentiment Handling

  • FR-057 — 3 points: As a trading analyst, I want the system to retrieve news from CoinDesk so that current crypto headlines can be considered as weak advisory input.
  • FR-058 — 3 points: As a trading analyst, I want the system to retrieve news from Cointelegraph so that the news layer is not dependent on a single source.
  • FR-059 — 3 points: As a user, I want news headlines to be deduplicated so that repeated stories do not distort sentiment.
  • FR-060 — 3 points: As a user, I want news URLs to accept HTTPS destinations only so that unsafe links are excluded.
  • FR-061 — 3 points: As a user, I want future or invalid publication timestamps to be rejected so that stale or implausible articles are not treated as fresh.
  • FR-062 — 3 points: As a trading analyst, I want news freshness to be measured over a 48-hour window so that older headlines receive no decision weight.
  • FR-063 — 3 points: As a trading analyst, I want weighted headline sentiment to be capped as a weak advisory feature so that news cannot dominate deterministic market analysis.
  • FR-064 — 3 points: As a user, I want the platform to show a neutral unavailable-news message and reduce decision confidence when news sources cannot be reached so that missing information is not hidden.
Page 8 of 30

3.8 Deterministic Market Analysis and Council Decisions

  • FR-065 — 5 points: As a trading analyst, I want market analysis to calculate EMA 9, EMA 21, EMA 50, RSI, ATR, momentum, trend, volume ratio, volatility, technical score, news score, and composite score so that recommendations are explainable.
  • FR-066 — 3 points: As a trading analyst, I want the decision engine to classify directional outcomes as LONG, SHORT, or HOLD so that action intent is unambiguous.
  • FR-067 — 3 points: As a user, I want the council to include director, deputy, news, risk, quant, and intelligence roles so that decisions reflect distinct advisory perspectives.
  • FR-068 — 3 points: As a user, I want each council member to provide a vote, confidence, score, and Arabic rationale so that consensus is inspectable.
  • FR-069 — 3 points: As a trading analyst, I want council scoring to incorporate funding, basis, and spread when available so that crowded or costly futures conditions reduce confidence appropriately.
  • FR-070 — 5 points: As a risk controller, I want a deterministic risk gate to have final veto authority so that council or AI output cannot bypass safety constraints.
  • FR-071 — 3 points: As a risk controller, I want approved decisions to require at least four of six council votes in one direction so that weak consensus cannot authorize execution.
  • FR-072 — 3 points: As a risk controller, I want approved decisions to require confidence of at least 64% so that low-confidence decisions result in HOLD.
  • FR-073 — 3 points: As a risk controller, I want decisions to be blocked when daily loss reaches the configured limit of 3% so that loss exposure is bounded.
  • FR-074 — 3 points: As a risk controller, I want decisions to be blocked when volatility exceeds 2.5% so that excessive volatility prevents entry.
  • FR-075 — 3 points: As a risk controller, I want decisions to be blocked when funding exceeds the allowed threshold, basis exceeds the allowed threshold, or spread exceeds 0.25% so that adverse futures conditions prevent entry.
  • FR-076 — 3 points: As a risk controller, I want decisions to be blocked when open positions reach the maximum of three so that concentration is limited.
  • FR-077 — 3 points: As a risk controller, I want decisions to require at least 50 candles of current source data so that insufficient market context prevents execution.
  • FR-078 — 3 points: As a user, I want each decision to include entry, stop, target, quantity, risk amount, leverage, risk/reward ratio, metrics, votes, checks, source, and summary so that the decision is actionable and auditable.
  • FR-079 — 3 points: As a user, I want decisions to expire after 75 seconds so that execution cannot use outdated market analysis.
  • FR-080 — 3 points: As a user, I want the decision engine to calculate default risk sizing at 1% of equity with a maximum notional exposure of 2x equity so that position sizing remains conservative.
  • FR-081 — 3 points: As a user, I want the decision engine to use a default risk/reward target of 2.1 so that stop and target geometry are consistent.
Page 9 of 30

3.9 AI Advisory Council

  • FR-082 — 5 points: As an operator, I want optional advisory seats for Anthropic, OpenAI, Google, xAI, Alibaba, and Moonshot so that the council can use multiple configured model providers.
  • FR-083 — 3 points: As an operator, I want an advisory seat to activate only when both its API key and actual provider model ID are configured so that display names are not treated as executable provider identifiers.
  • FR-084 — 3 points: As a user, I want the council UI to display configured model names without exposing API keys or model-ID environment variable names.
  • FR-085 — 5 points: As a risk controller, I want AI providers to receive advisory-only prompts that explicitly prohibit order execution so that models cannot be positioned as execution engines.
  • FR-086 — 3 points: As a risk controller, I want news headlines included in model context to be treated as untrusted quoted data so that embedded prompt-injection instructions are ignored.
  • FR-087 — 3 points: As a system, I want provider output parsed as strict JSON containing vote, confidence, rationale, and key risk so that AI advice has a controlled structure.
  • FR-088 — 3 points: As a risk controller, I want invalid AI votes to default to HOLD and confidence to be bounded between 50 and 95 so that malformed responses cannot create unsafe confidence.
  • FR-089 — 3 points: As a risk controller, I want advisory rationale and risk text lengths to be bounded so that provider responses cannot overload the interface or logs.
  • FR-090 — 5 points: As a system, I want configured AI advisory calls to run concurrently so that multi-provider consultation does not unnecessarily delay decisions.
  • FR-091 — 3 points: As a user, I want failed advisory seats to fall back to deterministic local rules so that provider outages do not disable the decision engine.
  • FR-092 — 5 points: As a risk controller, I want local deterministic scoring to retain 55% weight and AI advice to contribute only 45% for each consulted seat so that AI cannot dominate the decision.
  • FR-093 — 5 points: As a risk controller, I want deterministic council and risk checks to be re-run after AI blending so that advisory changes cannot bypass consensus or safety rules.

3.10 Dashboard, Account State, Positions, Orders, and Audit

  • FR-094 — 5 points: As a user, I want a dashboard showing market source, warnings, current price, quote source, recent candles, market metrics, stream status, and current symbol context so that I can assess the market in one place.
  • FR-095 — 3 points: As a user, I want the dashboard to show equity, available balance, realized daily PnL, daily PnL percentage, unrealized PnL, and open-position count so that account exposure is visible.
  • FR-096 — 3 points: As a user, I want positions to be marked against current or fallback prices so that unrealized PnL and portfolio equity can be calculated.
  • FR-097 — 3 points: As a user, I want the dashboard to show positions, recent orders, latest council decision, news, risk profile, kill-switch state, automation state, promotion state, exchange lock state, audit records, opportunities, strategies, and worker statuses so that operational context is consolidated.
  • FR-098 — 3 points: As an operator, I want local state to persist equity, positions, orders, audit entries, risk profile, automation state, and decision state so that the system can recover its operating context.
  • FR-099 — 3 points: As an operator, I want state saves to use a temporary file and backup behavior so that local state writes reduce corruption risk.
  • FR-100 — 3 points: As a user, I want daily trading state to roll over according to the Asia/Baghdad day so that daily PnL and daily-risk boundaries are consistently applied.
  • FR-101 — 3 points: As a user, I want daily history retained for up to 90 days so that recent daily performance is reviewable.
  • FR-102 — 3 points: As an operator, I want audit records for significant actions, warnings, errors, execution, position changes, strategy promotion, synchronization, and automation events so that activity is traceable.
  • FR-103 — 3 points: As an operator, I want audit history bounded to 60 entries so that local state remains controlled.
Page 10 of 30

3.11 Simulation and Manual Trade Lifecycle

  • FR-104 — 5 points: As a user, I want approved LONG or SHORT decisions to be executable in simulation so that strategies can be evaluated without exchange orders.
  • FR-105 — 3 points: As a user, I want simulated entries to apply adverse slippage and entry fees so that simulation does not present unrealistic results.
  • FR-106 — 3 points: As a user, I want simulated exits to apply adverse slippage and exit fees so that simulated PnL includes trading costs.
  • FR-107 — 3 points: As a user, I want to close an open simulated or Testnet position so that I can exit exposure manually.
  • FR-108 — 3 points: As a user, I want closed orders to record close price, fees, PnL, and status so that performance is visible.
  • FR-109 — 5 points: As a risk controller, I want execution blocked when the kill switch is active so that emergency suspension has immediate effect.
  • FR-110 — 3 points: As a risk controller, I want execution blocked when a decision is not approved, is not directional, or has expired so that only current approved decisions can create positions.
  • FR-111 — 3 points: As a risk controller, I want execution blocked when the maximum position count has been reached so that portfolio limits are enforced.
Page 11 of 30

3.12 Binance Futures Testnet Execution and Reconciliation

  • FR-112 — 5 points: As an operator, I want Binance execution to support only allow-listed Binance USDⓈ-M Futures Testnet hosts so that mainnet endpoints cannot be used.
  • FR-113 — 3 points: As an operator, I want Binance API credentials to be read from environment variables rather than stored in application state so that secrets are not persisted by the platform.
  • FR-114 — 3 points: As an operator, I want Testnet orders to require configured credentials, ALLOW_TESTNET_ORDERS=true, and a configured council control token so that execution requires deliberate activation.
  • FR-115 — 5 points: As a system, I want signed Binance requests to require valid Binance server-time synchronization so that stale local clocks cannot authorize signed requests.
  • FR-116 — 3 points: As an operator, I want time synchronization to retry with bounded delays and respect provider Retry-After signals so that timing recovery is controlled.
  • FR-117 — 3 points: As a system, I want failed time synchronization to block signed requests so that requests cannot proceed with unreliable timestamps.
  • FR-118 — 3 points: As a user, I want Binance Testnet execution to require a fresh direct Binance mark price so that exchange orders are not placed against stale or proxy data.
  • FR-119 — 3 points: As a user, I want Binance Testnet execution to require a fresh executable bid or ask price and spread at or below 0.25% so that order-book conditions are acceptable.
  • FR-120 — 3 points: As a user, I want Binance Testnet execution to be rejected when the executable price materially departs from planned entry so that stale plans are not executed.
  • FR-121 — 3 points: As a user, I want Binance Testnet execution to be rejected when price has crossed the planned stop or target geometry so that invalid trade setup geometry cannot execute.
  • FR-122 — 5 points: As a user, I want Testnet market entries to attach stop-market and take-profit-market close-all protection so that every exchange position has protective orders.
  • FR-123 — 5 points: As a risk controller, I want an emergency close attempt if protective orders cannot be established after a Testnet entry so that an unprotected Testnet position is not silently left open.
  • FR-124 — 3 points: As an operator, I want Testnet order quantities and prices normalized using exchange lot-size and price-filter metadata so that requests conform to exchange constraints.
  • FR-125 — 3 points: As an operator, I want leverage limited to a maximum of 2 so that Testnet exposure remains conservative.
  • FR-126 — 3 points: As an operator, I want Hedge Mode to be rejected for Testnet execution so that the system operates only in One-way Mode.
  • FR-127 — 5 points: As an operator, I want Binance account synchronization to reconcile wallet balance, available balance, unrealized PnL, positions, position mode, and protective algorithmic orders so that exchange state is understood before execution.
  • FR-128 — 5 points: As a risk controller, I want account synchronization to lock Testnet execution when Hedge Mode, external positions, missing exchange positions, quantity mismatches, or unprotected positions are detected so that discrepancies fail closed.
  • FR-129 — 3 points: As a risk controller, I want every active exchange position to require a valid adverse stop and favorable target based on mark-price close-all protection with price protection enabled so that exchange protection is complete.
  • FR-130 — 3 points: As an operator, I want Binance API synchronization failures to create an exchange-position lock rather than being interpreted as an empty account so that outages fail closed.
  • FR-131 — 3 points: As a user, I want Testnet position closing to close active exchange positions and attempt to cancel remaining standard and algorithmic protection orders so that closure cleanup is performed.
Page 12 of 30

3.13 Protective Stop Management

  • FR-132 — 5 points: As a user, I want positions to use staged trailing-stop logic so that protective stops can tighten as favorable movement develops.
  • FR-133 — 3 points: As a user, I want trailing-stop stage one to activate at 0.75R favorable movement so that the stop can advance near breakeven.
  • FR-134 — 3 points: As a user, I want trailing-stop stage two to activate at 1.25R favorable movement so that more gains are protected.
  • FR-135 — 3 points: As a user, I want trailing-stop stage three to activate at 1.75R favorable movement so that strong favorable moves receive tighter protection.
  • FR-136 — 3 points: As a user, I want Binance Testnet stop replacement to place the tighter stop before attempting cancellation of the prior stop so that protection is not removed before replacement is active.
  • FR-137 — 3 points: As a user, I want simulated positions to close automatically when stop or target conditions are met so that simulation includes protective lifecycle handling.
  • FR-138 — 3 points: As an operator, I want stop-adjustment failures recorded in the audit trail so that protective-operation issues are visible.
Page 13 of 30

3.14 Backtesting and Strategy Evaluation

  • FR-139 — 5 points: As a research analyst, I want to backtest the council_ensemble strategy so that I can evaluate combined trend, momentum, and breakout behavior.
  • FR-140 — 3 points: As a research analyst, I want to backtest the trend_follow strategy so that I can evaluate EMA 9/21 with EMA 50 trend filtering.
  • FR-141 — 3 points: As a research analyst, I want to backtest the breakout strategy so that I can evaluate 20-candle range breaks with volume confirmation.
  • FR-142 — 3 points: As a research analyst, I want to backtest the momentum strategy so that I can evaluate RSI and momentum behavior with a trend filter.
  • FR-143 — 3 points: As a research analyst, I want backtests to require at least 100 candles so that insufficient history is rejected.
  • FR-144 — 5 points: As a research analyst, I want backtests to include fees, adverse slippage, risk-per-trade sizing, maximum 2x leverage, stop logic, target logic, signal-flip exits, and cooldown behavior so that results are conservative.
  • FR-145 — 3 points: As a research analyst, I want backtests to calculate return, final equity, maximum drawdown, trade count, wins, losses, win rate, profit factor, Sharpe proxy, and fees paid so that strategy performance is assessable.
  • FR-146 — 3 points: As a research analyst, I want backtests to reserve the final 30% of the sample as an out-of-sample holdout so that strategy promotion considers unseen-period behavior.
  • FR-147 — 3 points: As a research analyst, I want strategy comparison results ranked by out-of-sample return and drawdown so that candidates can be compared consistently.
  • FR-148 — 3 points: As a user, I want the backtest equity curve to be compacted while preserving endpoints so that results remain usable in the interface.
  • FR-149 — 3 points: As a research analyst, I want a promotion result of INSUFFICIENT when fewer than 20 full-sample trades or five out-of-sample trades exist so that inadequate samples cannot be promoted.
  • FR-150 — 3 points: As a research analyst, I want a promotion result of CANDIDATE only when full and out-of-sample returns are positive, full drawdown is below 8%, out-of-sample profit factor is at least 1.2, and out-of-sample Sharpe proxy is positive so that promotion requires conservative evidence.
  • FR-151 — 3 points: As a research analyst, I want strategies failing promotion criteria to be marked REJECTED so that inadequate performance is explicit.
  • FR-152 — 5 points: As an operator, I want strategy promotion to require direct Binance Futures Testnet or official Binance USD-M Futures Archive data so that promoted strategies are not based on proxy or synthetic data.
  • FR-153 — 3 points: As an operator, I want promoted strategies to be scoped to Testnet only and expire after 24 hours so that promotion is temporary and environment-limited.
  • FR-154 — 3 points: As an operator, I want promotion validity to require the same symbol, strategy, current engine version, and unexpired promotion record so that stale evaluations cannot authorize automation.
  • FR-155 — 5 points: As a user, I want a full candidate evaluation to combine four-strategy comparison, council analysis, direct-data validation, quality grade, and promotion outcome so that readiness for experimental entry is comprehensively assessed.
Page 14 of 30

3.15 Automation Supervisor

  • FR-156 — 3 points: As an operator, I want automation mode OFF so that an armed session can be explicitly stopped.
  • FR-157 — 3 points: As an operator, I want automation mode MONITOR so that the platform can analyze without creating positions.
  • FR-158 — 3 points: As an operator, I want automation mode SIMULATION so that approved matched signals can create simulated positions.
  • FR-159 — 5 points: As an operator, I want automation mode TESTNET so that approved matched signals can execute only in Binance Futures Testnet after all Testnet safeguards pass.
  • FR-160 — 3 points: As an operator, I want SIMULATION mode to require typed confirmation SIMULATION so that simulated automation is deliberate.
  • FR-161 — 3 points: As an operator, I want TESTNET mode to require typed confirmation TESTNET so that Testnet automation is deliberate.
  • FR-162 — 5 points: As an operator, I want TESTNET automation to require enabled Testnet orders, a current promoted strategy, successful account synchronization, and no exchange-position lock so that automation cannot bypass execution readiness.
  • FR-163 — 3 points: As an operator, I want automation sessions limited to 5–30 minutes so that unattended operation is time-bounded.
  • FR-164 — 3 points: As an operator, I want automation scans to run on a configured interval with a default of 20 seconds so that monitoring is controlled.
  • FR-165 — 3 points: As an operator, I want automation to require agreement between the selected strategy signal and the council decision so that disagreement results in HOLD.
  • FR-166 — 3 points: As an operator, I want MONITOR mode to avoid execution even when a matched signal exists so that observation remains non-trading.
  • FR-167 — 3 points: As an operator, I want automation to avoid new execution while positions exist or a cooldown is active so that it does not stack exposure.
  • FR-168 — 3 points: As an operator, I want automation limited to two trades per hour so that excessive trade frequency is prevented.
  • FR-169 — 3 points: As an operator, I want automation to apply a 10-minute cooldown after execution so that repetitive re-entry is controlled.
  • FR-170 — 5 points: As a risk controller, I want automation to activate the kill switch and disarm when realized daily loss reaches the daily-loss limit so that automatic trading stops after unacceptable loss.
  • FR-171 — 5 points: As a risk controller, I want automation to activate the kill switch and disarm after three consecutive losing closed trades so that a loss circuit breaker is enforced.
  • FR-172 — 3 points: As an operator, I want Testnet automation blocked when direct Binance Futures Testnet candles are unavailable so that proxy data cannot drive Testnet automation.
  • FR-173 — 3 points: As an operator, I want a single unexpected Testnet automation error to disarm the session so that sensitive execution does not repeatedly retry after faults.
  • FR-174 — 3 points: As an operator, I want three consecutive MONITOR or SIMULATION errors to disarm the session so that non-Testnet automation also fails safely.
  • FR-175 — 3 points: As a user, I want automation status to display armed state, mode, symbol, strategy, interval, session expiry, scans, next scan, last signal, last result, cooldown, and errors so that automated behavior is observable.
  • FR-176 — 3 points: As a user, I want full autonomous trading mode to remain disabled in the MVP so that execution continues to require the defined approval and automation safeguards.
Page 15 of 30

3.16 Kill Switch, Controls, and API Safety

  • FR-177 — 5 points: As an operator, I want an emergency kill switch that disarms automation and prevents execution so that trading activity can be stopped immediately.
  • FR-178 — 3 points: As an operator, I want cancellation of the kill switch to require control-token validation when Binance credentials are configured so that sensitive recovery is controlled.
  • FR-179 — 3 points: As an operator, I want sensitive account synchronization, Testnet execution, Testnet position closing, Testnet automation arming, and protected demo reset to require a council control token so that privileged actions require explicit authorization.
  • FR-180 — 3 points: As a system, I want control-token comparison to use constant-time comparison so that token checks reduce timing-attack exposure.
  • FR-181 — 3 points: As an API consumer, I want request bodies limited to 100 KB so that oversized payloads are rejected.
  • FR-182 — 3 points: As an API consumer, I want action requests with bodies to require application/json content type so that unsupported request formats are rejected.
  • FR-183 — 3 points: As an API consumer, I want expensive endpoints to be rate limited separately from ordinary write endpoints so that analysis and backtesting cannot be abused.
  • FR-184 — 3 points: As an API consumer, I want rate-limit responses to return HTTP 429 so that clients receive a standard throttling signal.
  • FR-185 — 3 points: As an operator, I want demo reset capability to restore the defined default state so that safe local testing can restart from a known baseline.

3.17 Data Foundation, Truth Center, and Provider Onboarding

  • FR-186 — 3 points: As a compliance reviewer, I want the platform to expose provider registry entries including provider identity, classification, status, technical role, code license, data authorization, and retrieval state so that provider readiness is reviewable.
  • FR-187 — 3 points: As a compliance reviewer, I want the platform to expose data-category registry entries including authorization status, candidate provider, endpoint identifier, retention requirement, and intended research use so that data scope is reviewable.
  • FR-188 — 3 points: As a user, I want the platform to expose Tardis onboarding status as a documented local trial path so that provider onboarding can be assessed without creating a provider account, session, or download.
  • FR-189 — 3 points: As a user, I want a truth-center view for persisted data when display is explicitly enabled so that instruments, coverage, providers, and status can be reviewed.
  • FR-190 — 3 points: As a user, I want separate persisted-data endpoints for instruments, data coverage, and providers so that governed data views can be accessed independently.
  • FR-191 — 5 points: As a compliance reviewer, I want provider retrieval, persisted market display, training, backtesting, paper execution, live execution, and orders blocked in the safe data-foundation state so that unapproved capabilities cannot operate.
  • FR-192 — 3 points: As a compliance reviewer, I want the product to state that the data-foundation catalog schema is design-only and does not itself create market rows so that schema availability is not confused with data acquisition.
  • FR-193 — 3 points: As a compliance reviewer, I want the system to report zero or governed counts for approved providers, approved categories, approved scopes, retrieval events, market rows, and research-eligible rows so that readiness claims remain measurable.

4. User Personas

4.1 Trading Analyst

Uses the Arabic dashboard to inspect market data, news, indicators, council votes, trade proposals, opportunities, positions, and research results.

Page 16 of 30

4.2 Research Analyst

Runs historical backtests, compares strategies, evaluates candle quality, reviews out-of-sample results, and promotes qualified strategies for Testnet-only experimentation.

4.3 Automation Operator

Arms, monitors, and disarms automation sessions; selects symbol, strategy, interval, and mode; supplies required confirmations; and reviews automation outcomes.

4.4 Risk and Compliance Reviewer

Reviews authorization gates, provider status, data-category status, platform readiness, truth-center information, audit activity, exchange reconciliation, protection status, and blocked capability states.

4.5 Testnet Execution Controller

Performs controlled Binance Futures Testnet synchronization, validates exchange reconciliation, executes approved decisions, closes positions, and uses the emergency kill switch.

4.6 Release and Build Maintainer

Verifies source-bundle scope, original source bytes, hashes, Python dependencies, required imports, workflow checks, and release reproducibility.

Page 17 of 30

4.7 External System Actors

  • Binance USDⓈ-M Futures Testnet.
  • Binance official public-data services.
  • Kraken.
  • CoinDesk.
  • Cointelegraph.
  • Anthropic.
  • OpenAI.
  • Google.
  • xAI.
  • Alibaba.
  • Moonshot.

5. Core User Flows

Page 18 of 30

5.1 Trading Analysis Flow

  1. The Trading Analyst opens the Arabic dashboard.
  2. The system displays environment, platform mode, source status, safe-preview state, and execution mode.
  3. The analyst selects a valid USDⓈ-M trading symbol.
  4. The system attempts direct Binance Futures Testnet candle retrieval.
  5. If necessary, the system uses recent official Binance spot klines.
  6. If Binance spot data is unavailable, the system uses Kraken candles.
  7. If external data is unavailable, the system uses clearly labeled local synthetic candles.
  8. The platform shows market-data source and warnings.
  9. The analyst runs analysis.
  10. The system refreshes candles, news, market details, account context, and position state.
  11. The deterministic engine calculates technical and market-risk metrics.
  12. Configured advisory providers are consulted in parallel.
  13. The system blends validated AI advice with deterministic scores.
  14. The deterministic risk gate produces an APPROVED or BLOCKED result.
  15. The decision is stored, audited, and shown with checks, votes, entry, stop, target, and expiry.

5.2 Simulation Execution Flow

  1. The user reviews an APPROVED and unexpired LONG or SHORT decision.
  2. The user requests execution in simulation.
  3. The system verifies kill-switch status, decision validity, expiry, and position limits.
  4. The system creates a simulated position with slippage and entry fees.
  5. The platform persists the order, position, audit event, and account effects.
  6. The dashboard updates PnL, equity, available balance, and exposure.
  7. The user closes the position manually, or automated stop/target monitoring closes it.
Page 19 of 30

5.3 Binance Futures Testnet Execution Flow

  1. The Testnet Execution Controller configures Testnet credentials, enables orders, and configures the control token.
  2. The controller requests Binance synchronization.
  3. The system validates Binance server time, account status, One-way Mode, local/exchange position matching, quantity matching, and protection orders.
  4. The system sets an exchange-position lock if reconciliation fails.
  5. The controller runs fresh analysis using direct Binance Futures Testnet data.
  6. The system verifies decision approval, direction, expiry, source, mark-price freshness, executable bid/ask freshness, spread, and planned-entry deviation.
  7. The controller supplies the required control token.
  8. The system places a Testnet market order.
  9. The system creates a stop-market and take-profit-market close-all protection pair.
  10. If protection creation fails, the system attempts an emergency close.
  11. The platform records the resulting exchange, position, order, and audit state.

5.4 Strategy Research and Promotion Flow

  1. The Research Analyst selects a symbol, strategy, and interval.
  2. The platform retrieves historical data, preferring official Binance Futures archives.
  3. The archive layer uses ZIP parsing and date handling to locate and interpret completed daily archives.
  4. The system assesses candle quality.
  5. The platform runs the strategy using fees, slippage, leverage limits, stops, targets, cooldowns, and out-of-sample testing.
  6. The analyst reviews metrics and promotion status.
  7. The analyst may compare all four strategies.
  8. The analyst promotes only a candidate strategy based on direct Binance Futures data with quality grade A or B.
  9. The platform creates a Testnet-only promotion record that expires in 24 hours.
Page 20 of 30

5.5 Automation Supervisor Flow

  1. The Automation Operator selects MONITOR, SIMULATION, or TESTNET.
  2. The operator selects symbol, strategy, interval, and duration.
  3. For simulation, the operator enters SIMULATION.
  4. For Testnet, the operator enters TESTNET and supplies the control token.
  5. The platform verifies current strategy promotion, exchange synchronization, no exchange lock, and Testnet order readiness.
  6. The platform arms the session.
  7. Each scan calculates the current strategy signal and council decision.
  8. The platform requires both to agree on a non-HOLD direction.
  9. MONITOR records findings only.
  10. SIMULATION or TESTNET execution proceeds only if all risk, timing, position, cooldown, and trade-rate conditions pass.
  11. The session disarms on expiry, kill switch, loss circuit breaker, daily-loss violation, or defined error conditions.

5.6 Governance and Readiness Review Flow

  1. The Risk and Compliance Reviewer opens platform readiness.
  2. The platform reads local fixed readiness artifacts only.
  3. The reviewer inspects authorization decision, gates, provider registry, data-category registry, module status, counts, public-development status, and onboarding status.
  4. The reviewer confirms that blocked capabilities remain blocked.
  5. When persisted-data display is enabled, the reviewer accesses truth-center, instrument, coverage, and provider records.
  6. The reviewer uses audit history, status responses, and reconciliation results for operational review.
Page 21 of 30

5.7 Release Integrity Review Flow

  1. The Release and Build Maintainer reviews source-bundle metadata.
  2. The maintainer verifies declared original source bytes and file counts.
  3. The maintainer verifies file hashes.
  4. The maintainer validates the dependency lock and Python version.
  5. The maintainer confirms required modules—including ATR import usage, ZIP support, and datetime support—are available for research processing.
  6. The maintainer runs compilation, JavaScript syntax checks, deterministic tests, and frozen-evidence validation.

6. Visuals Colors and Theme

6.1 Visual Direction

A professional Arabic financial-intelligence workspace emphasizing safety, traceability, source provenance, and operational clarity. The product should avoid speculative or promotional trading visuals.

6.2 Theme

  • Primary experience: dark analytical workspace for prolonged market monitoring.
  • Secondary experience: restrained light mode suitable for governance and research review.
  • Interface direction: RTL-first.
  • Market symbols, numbers, model names, API labels, and timestamps must retain readable mixed-direction formatting.
Page 22 of 30

6.3 Color System

  • Background: deep navy or charcoal.
  • Surface panels: dark slate.
  • Primary action: restrained electric blue.
  • Approved, protected, or healthy: muted green.
  • Warning, proxy source, or review-required: amber.
  • Blocked, failed, stale, or kill-switch state: red.
  • Simulation and informational states: neutral cyan or slate blue.
  • Council members: retain their defined distinct colors while satisfying contrast requirements.

6.4 Data Visualization

  • LONG must use green with textual labels and non-color indicators.
  • SHORT must use red with textual labels and non-color indicators.
  • HOLD and BLOCKED must use visible explanatory labels.
  • Proxy, fallback, synthetic, stale, and unauthorized data must remain visually distinct.
  • Charts must identify source, interval, freshness, and Testnet execution eligibility.

7. Signature Design Concept

Page 23 of 30

The Governed Council Desk

The signature design concept is a three-layer decision workspace:

  1. Market Reality

    • Price, source, freshness, chart, liquidity, funding, basis, spread, and news.
  2. Council Deliberation

    • Six visible advisory seats with role, model status, vote, confidence, rationale, and key risk.
  3. Deterministic Authority

    • Prominent risk-gate panel with every pass/fail condition, decision expiry, stop/target geometry, execution readiness, exchange lock, protection status, and kill-switch state.

The central UX message is explicit: AI advisors can inform a decision, but deterministic controls authorize or block it.

8. Interaction Model & Motion Direction

8.1 Interaction Model

  • RTL-first navigation and content flow.
  • Structured operational density rather than a consumer-trading interface.
  • Progressive disclosure for advanced research, governance, reconciliation, and exchange details.
  • Sensitive actions require explicit confirmation and, where required, typed confirmation and control-token validation.
  • Disabled actions must explain the missing authorization, safety condition, or required gate.
  • Data source and freshness must remain close to the information they qualify.
Page 24 of 30

8.2 Motion Direction

  • Use restrained operational transitions only.
  • Refreshes should use subtle loading and status-state transitions.
  • Decision state may transition visibly between analysis, approved, and blocked.
  • Kill-switch activation, exchange locks, stale data, and protection failures must use immediate high-priority feedback.
  • Avoid celebratory profit animations, gamified execution effects, or motion that implies investment certainty.

9. Non-Functional Requirements

9.1 Safety and Risk

  • Production/mainnet trading must not be implemented.
  • Binance USDⓈ-M Futures Testnet is the only exchange execution target.
  • AI outputs must remain advisory and must not override deterministic risk controls.
  • Unavailable, stale, unauthorized, mismatched, unprotected, or ambiguous conditions must fail closed.
  • Proxy and synthetic data must never independently authorize Testnet execution.
  • Testnet execution requires fresh direct Binance market data, valid market conditions, reconciliation, protection, and token-controlled authorization.
  • Full autonomous trading remains disabled in the MVP.
Page 25 of 30

9.2 Security

  • Credentials and API keys must be sourced from environment variables.
  • Secrets must not be exposed in UI responses, logs, source metadata views, or council information.
  • Sensitive actions must use a server-side control token.
  • Token comparison must use constant-time comparison.
  • Request bodies must be limited to 100 KB.
  • Requests with payloads must require JSON content type.
  • Sensitive and expensive endpoints must be rate limited.
  • Static-path resolution must prevent directory traversal.
  • API responses must use no-store caching behavior.
  • Security response headers must include content-type protection, referrer protection, restricted permissions, same-origin resource policies, frame restrictions unless explicitly overridden, and restrictive content-security policy.

9.3 Performance and Resilience

  • Market, news, historical-data, opportunity, backtest, and exchange metadata caching must reduce unnecessary provider requests.
  • Market streaming must reconnect after failures.
  • AI provider requests must run concurrently where multiple providers are configured.
  • AI provider failures must degrade to deterministic local rules.
  • Binance time synchronization must use bounded retry behavior and failure cooldown.
  • Local state persistence must use temporary-write replacement and backup behavior.
  • Historical archive parsing must enforce safe archive layout expectations and bounded retrieval sizes.
Page 26 of 30

9.4 Data Quality and Provenance

  • Every displayed market source must be identifiable.
  • Proxy and synthetic data must show warnings.
  • Historical backtests must identify source and quality grade.
  • Stream freshness thresholds must govern quote and order-book usability.
  • News must be deduplicated, timestamp-validated, freshness-weighted, and capped in influence.
  • Data retention, display, retrieval, research, and execution must remain separately authorization-gated.
  • Source-bundle metadata must preserve verifiable original source-byte scope and file integrity.

9.5 Global Leadership Quality Bar

  • Global-leadership claims must be backed by the measurable pillars defined in Section 1.3 and must never be expressed as profit guarantees or investment certainty.
  • Every capability that supports global competitiveness must remain subject to the same authorization gates, fail-closed behavior, and Testnet-only execution scope defined elsewhere in this document.
  • International readability must be preserved without weakening the Arabic-first interface: mixed-direction formatting for symbols, numbers, model names, API labels, and timestamps must remain correct at every viewport.
  • Global-leadership progress must remain auditable through the existing audit trail, readiness counts, and release-integrity evidence.

9.6 Testing and Delivery Quality

  • The platform must support deterministic tests for API security, backtesting, Binance client behavior, platform shell behavior, data foundation, data quality, engine logic, provider gating, market streams, news quality, HTTP controls, ingestion, storage, truth-center behavior, reconciliation, safe preview, time synchronization, UI professionalization, public-development mode, and universe filtering.
  • Continuous integration must:
    • use Python 3.13;
    • use Node.js 20 for JavaScript syntax validation;
    • install hash-pinned runtime dependencies;
    • compile Python source;
    • validate dashboard JavaScript;
    • run deterministic tests; and
    • validate frozen evidence.

10. Tech Stack

Page 27 of 30

10.1 Application Runtime

  • Python 3.13.
  • Python standard library HTTP serving using ThreadingHTTPServer and SimpleHTTPRequestHandler.
  • Standard-library modules for JSON, HTTP requests, HMAC signing, threading, XML RSS parsing, local persistence, ZIP parsing, and date/time handling.

10.2 Frontend

  • Static HTML.
  • Static CSS.
  • Vanilla JavaScript.
  • Arabic RTL interface.

10.3 Market Connectivity

  • Binance USDⓈ-M Futures Testnet REST API.
  • Binance Futures Testnet WebSocket.
  • Official Binance public spot data endpoint for recent spot klines.
  • Official Binance Vision USD-M Futures daily ZIP archives.
  • Kraken REST OHLC endpoint.
  • Kraken WebSocket ticker endpoint.

10.4 AI Advisory Providers

  • Anthropic API.
  • OpenAI-compatible API.
  • Google Generative Language API.
  • xAI OpenAI-compatible API.
  • Alibaba DashScope-compatible API.
  • Moonshot OpenAI-compatible API.
Page 28 of 30

10.5 Dependencies

  • websocket-client>=1.8,<2.
  • Hash-pinned dependency lock for websocket-client==1.9.0.
  • Python ZIP archive handling through zipfile.
  • Python date/time handling through datetime.

10.6 Data Foundation

  • DuckDB catalog schema is included as a design-only specification.
  • DuckDB is not installed or selected as a runtime dependency in the provided scope.
Page 29 of 30

11. Assumptions and Constraints

  1. Binance Futures Testnet—not Binance mainnet—is the only supported exchange execution environment.
  2. Mainnet URLs are intentionally unsupported.
  3. Testnet orders are disabled by default.
  4. Testnet execution requires environment configuration, explicit order enablement, a control token, direct Binance Testnet data, fresh prices, valid spread, valid geometry, successful reconciliation, and active protection.
  5. The platform may operate in DEV_PUBLIC_READONLY, where privileged provider paths remain disabled.
  6. Safe-preview mode blocks data, provider, trading, and action capabilities.
  7. Binance spot, Kraken, and synthetic data are analysis and display fallbacks only.
  8. Proxy and synthetic data cannot independently authorize Futures Testnet execution.
  9. AI provider configuration is optional; deterministic analysis must function without AI providers.
  10. Provider display names may differ from actual API model identifiers.
  11. AI responses are untrusted advisory data and cannot execute trades.
  12. Persisted data display, provider retrieval, training, research, paper execution, and live execution are independently gated.
  13. Live execution is outside the current phase.
  14. Full autonomous trading remains disabled in the MVP.
  15. Default simulated starting equity is 10,000.
  16. Default risk profile is 1% risk per trade, 3% daily loss limit, three maximum positions, 2x maximum leverage, and 64% minimum confidence.
  17. Trading-day rollover follows the Asia/Baghdad timezone.
  18. Strategy promotion is symbol-specific, strategy-specific, engine-version-specific, Testnet-only, and valid for 24 hours.
  19. Historical ZIP archives may be unavailable for some dates; missing archive days may be skipped, but insufficient returned data must prevent reliable backtesting.
  20. Original source-byte metadata is a release-integrity record and does not itself represent runtime trading data.
Page 30 of 30

12. Glossary

TermDefinition
AI AdvisoryOptional model-provider output used as constrained input to council scoring; it cannot execute trades or override deterministic risk controls.
Approved DecisionA directional LONG or SHORT decision that has passed all deterministic checks.
ATRAverage True Range; a volatility measure used for risk distance, stops, and strategy calculations.
Binance Futures TestnetBinance’s non-production USDⓈ-M Futures environment used for experimental order execution.
BasisPercentage difference between mark price and index price.
CandleOHLCV market-data record containing open, high, low, close, volume, and timestamp.
CouncilSix-seat advisory framework consisting of director, deputy, news, risk, quant, and intelligence roles.
Control TokenServer-side secret supplied in the X-Council-Token header for sensitive operations.
Data AuthorizationHuman-governed approval state required before protected data capabilities can activate.
Direct DataBinance Futures Testnet or official Binance USD-M Futures archive data.
Fail ClosedSafety behavior that blocks an operation when authorization, freshness, integrity, reconciliation, or required data is unavailable.
Funding RateFutures funding value used as a market-crowding and carry-cost signal.
Hedge ModeBinance position mode that supports simultaneous long and short exposure; unsupported for SkyLink Testnet execution.
Holdout / Out-of-SampleFinal 30% of a backtest period used to evaluate performance outside the primary sample.
Kill SwitchEmergency control that disables automation and blocks execution.
LONGDirectional expectation that price will rise.
Mark PriceExchange-calculated reference price used for futures valuation and conditional protection.
One-way ModeBinance position mode allowing a single net position direction per symbol; required for Testnet execution.
Original Source BytesDeclared total byte count of the original Git-tracked source bundle used as an integrity and delivery-scope reference.
Proxy DataNon-Futures-Testnet market data such as official Binance spot or Kraken data used for display or analysis fallback.
Recent Spot KlinesRecent official Binance spot OHLCV candles used as a labeled fallback for analysis and display only.
Risk GateDeterministic set of checks that must pass before a council decision can be approved for execution.
SAFE_SIMULATIONLocal simulated execution mode that does not place exchange orders.
Safe PreviewOperating mode that blocks data, provider, trading, and action capabilities.
SHORTDirectional expectation that price will fall.
SlippageDifference between intended and modeled or actual execution price.
SpreadPercentage difference between the current bid and ask prices.
Strategy PromotionTemporary Testnet-only approval of a strategy after satisfying data-quality and backtesting criteria.
Testnet ExecutionControlled placement of orders on Binance USDⓈ-M Futures Testnet.
Truth CenterGoverned view of persisted-data status, instruments, coverage, and providers.
USDⓈ-MBinance futures contracts margined and settled using stablecoin assets such as USDT or USDC.
ZIP ArchiveCompressed Binance historical-data package containing daily market candle records.

No completed page designs yet.

Completed design pages will appear here when they are ready to preview.

Landing: Open platform home
Login: Sign in
Automation: Select mode and symbol
Automation: Arm automation session
Automation: Monitor scans and signal
Automation: Disarm on expiry or error
Kill Switch: Activate emergency kill switch
Audit: Review automation events

No completed page designs yet.

Completed design pages will appear here when they are ready to preview.

Landing: Open platform home
Login: Sign in
Automation: Select mode and symbol
Automation: Arm automation session
Automation: Monitor scans and signal
Automation: Disarm on expiry or error
Kill Switch: Activate emergency kill switch
Audit: Review automation events