Trading Signals — AI Crypto Trading Signals for Agents

Install

Claude Code Plugin

/plugin install roman-rr/trading-skills click to copy

npx (Claude, Codex, Cursor, Windsurf)

npx skills add roman-rr/trading-skills click to copy

MCP Server (Claude Desktop / Code — native tools)

MCP endpoint: https://signals.x70.ai/mcp click to copy

Add to Claude Desktop config or claude mcp add trading-signals --transport http -- https://signals.x70.ai/mcp

Live Signal Preview

BTC

BULLISH 87%

Entry$68,450

SL$67,200

TP$71,800

Leverage3x

R/R2.7:1

ETH

BEARISH 82%

Entry$3,840

SL$3,920

TP$3,680

Leverage2x

R/R2.0:1

SOL

BULLISH 79%

Entry$142.50

SL$138.00

TP$152.00

Leverage2x

R/R2.1:1

How It Works

1. Register One-time

Your AI agent calls POST /api/skill/register with your name, email, and GitHub. Gets an API key instantly.

2. Get Signals Real-time

GET /api/skill/signals returns live trade setups with entry, SL, TP, leverage, and confidence scores.

3. Track Verification Automated

Every signal is automatically verified against real market prices. Check back with the signal ID to see if TP/SL was hit and actual ROI.

API Endpoints

Method	Path	Auth	Description
POST	/api/skill/register	No	Register & get API key
GET	/api/skill/signals	API key	List signals (active/verified/all)
GET	/api/skill/signals/:id	API key	Single signal + verification status
GET	/api/skill/stats	API key	Hit rate, ROI, performance stats

What You Get

Feature	Description
Live Signals	Bullish/bearish crypto signals with full trade setups — entry, stop-loss, take-profit, leverage, position size
50+ Coins	Top 30 by volume, top 20 by funding rate, top 20 by open interest — dynamically selected from crypto perpetual markets
Confidence Score	0-100 AI conviction level per signal. Higher confidence = tighter risk parameters, larger position sizing
Risk/Reward	Every signal includes R/R ratio (minimum 1.5:1 enforced), leverage caps tied to confidence level
Signal Types	Funding anomalies, volume spikes, price breakouts, whale activity, liquidation cascades, momentum shifts, trend reversals, and more
Auto-Verification	TP/SL monitored every minute against live mark prices. Rules engine resolves remaining signals with 1H candle data
ROI Tracking	Leverage-adjusted P&L per signal — wins capped at TP distance, losses capped at SL distance
Performance Stats	Hit rate, cumulative ROI, average leverage, breakdown by direction (long/short) and by coin
Market Data	Signals powered by funding rates, order book depth, whale trades, OI data, technical indicators (RSI, MACD, Bollinger Bands), and crypto news
Adaptive Risk	System automatically adjusts signal confidence, leverage caps, and position sizing based on recent track record
Multi-Expert	Each signal is validated by multiple independent AI analysis perspectives before publishing
24/7 Operation	Signals generated around the clock via market anomaly detection and trend monitoring — not just scheduled intervals

Scientific Foundation — 33 Research-Grounded Methods

Every algorithm is grounded in peer-reviewed research — 39 academic citations from quantitative finance, statistics, and machine learning.

Position Sizing

Kelly Criterion (half-Kelly), ATR-based normalization, Turtle Traders risk model. Adaptive sizing from paper balance metrics.

Anomaly Detection

CUSUM change-point detection, Modified Z-score (MAD), EWMA beta decorrelation, excess kurtosis fat-tail analysis, realized volatility ratio, Benjamini-Hochberg FDR control.

Technical Analysis

RSI, MACD, Bollinger Bands, ATR, ADX, VWAP, OLS regression slope + R², swing-level S/R, Pearson cross-asset correlation.

Multi-Expert AI

Mixture of Experts ensemble with Thompson sampling model selection, weighted composite scoring, regime-adaptive expert weighting. Each signal verified by independent Technical + Flow analysis.

Market Microstructure

CVD via tick rule (Lee & Ready 1991), funding rate contrarian signals, options expiry pin risk, order book imbalance detection.

Risk Management

Balance-based regime switching, profit factor over hit rate, walk-forward paper validation, ISQ multi-attribute quality scoring (Keeney-Raiffa MAUT).

Why Trading Signals?

Battle-tested in production — real signals from a live trading platform, not a demo

Every signal verified — automated tracking against actual market prices, wins and losses

Multi-expert consensus — signals fire only when multiple AI perspectives agree

17 real-time data sources — funding rates, order books, whale trades, smart money, news, TA indicators

33 research-grounded algorithms — from Kelly Criterion to CUSUM change-point detection, backed by 39 academic citations

Adaptive risk management — the system adjusts position sizing and confidence thresholds based on rolling performance

Trading Signalsfor AI Agents

Trading Signals
for AI Agents