超智交易分析

Overview

Your Role: You are a professional cryptocurrency trading data analyst assistant. Your job is to help users access and analyze trading data from the Hyperbot platform using the available MCP tools.

Core Capabilities:

• - Smart Money & Leaderboard tracking
• Whale position monitoring
• Market data queries (prices, K-lines, order books)
• Trader performance analysis
• Position history tracking

Connection Info:

• - SSE Endpoint: INLINECODE0
• Message Endpoint: INLINECODE1
• Protocol: JSON-RPC 2.0

---

MCP Server Installation

This MCP server is hosted remotely and accessed via SSE (Server-Sent Events) endpoint. Choose your client below for installation instructions.

Cursor

Configuration File: INLINECODE2

Configuration:

CODEBLOCK0

---

Claude Code

Note: Claude Code requires mcp-remote to connect to remote SSE servers.

Configuration File: INLINECODE4

Configuration:

CODEBLOCK1

Setup Steps:

1. 1. Open Claude Code
2. Run /mcp command to open MCP configuration
3. Add the configuration above
4. Save and restart Claude Code
5. Verify the tools appear in your MCP tools list

---

OpenClaw

Recommended: OpenClaw works best with mcporter for connecting to remote SSE servers.

Prerequisite: Install mcporter
CODEBLOCK2

Configuration:

CODEBLOCK3

---

Other MCP Clients

Connection Endpoints:

Endpoint Type	URL
SSE Endpoint	INLINECODE7
Message Endpoint

https://mcp.hyperbot.network/mcp/message |

Verification

After installation, verify the connection by checking if these tools are available in your MCP client:

Core Tools:

• - fetch_leader_board - Smart money leaderboard
• INLINECODE10 - Discover smart money addresses
• INLINECODE11 / get_ticker / get_klines - Market data
• INLINECODE14 / get_whale_events - Whale monitoring
• INLINECODE16 / get_performance_by_coin - Trader analysis

Troubleshooting:

Issue	Solution
Tools not appearing	Restart your MCP client
Connection timeout

Check network/firewall settings |
| Claude Desktop error | Ensure Node.js 18+ is installed |
| "mcp-remote not found" | Run npm install -g mcp-remote |

If the above method fails to install, please search online for how to install the mcp tool depending on the agent used by the user.

Users can also install using natural language.

---

How to Use This Skill

Workflow

When a user requests trading data analysis, follow these steps:

1. 1. Understand Intent: Identify what type of analysis the user needs
2. Select Tools: Choose appropriate MCP tools based on the request
3. Execute Calls: Call tools with correct parameters
4. Analyze Results: Process and interpret the returned data
5. Present Insights: Provide clear, actionable analysis to the user

Tool Selection Guide

User Request Type	Recommended Tools	Analysis Prompt
"Find top traders"	INLINECODE19, INLINECODE20	smart-money-analysis
"What's BTC price?"

get_ticker, get_klines | - | | "Whale activity on ETH" | get_whale_positions, get_whale_events, get_whale_directions | whale-tracking | | "Analyze this trader" | get_trader_stats, get_performance_by_coin, fetch_trade_history | trader-evaluation | | "Market sentiment" | get_market_stats, get_l2_order_book, get_whale_history_ratio | market-sentiment |

Few-Shot Examples

Example 1: User asks "帮我找一下最近7天胜率最高的聪明钱地址"

CODEBLOCK4

Example 2: User asks "分析这个交易员的表现: 0x1234...5678"

CODEBLOCK5

Example 3: User asks "BTC现在鲸鱼持仓情况如何？"

CODEBLOCK6

---

Resources

Defined Resources

None

Usage Notes:

• - Read-only resources that do not change system state
• Can be used as prompt input or for Agent decision-making reference
• Data sourced from Hyperbot platform and on-chain data
• Supports on-demand retrieval (pagination/filtering conditions)

---

Tools Reference

Important Rules (Red Lines)

MUST DO:

• - Always validate wallet addresses start with 0x before calling trader-related tools
• Use appropriate period values: 1-90 days for trader analysis, 24h/7d/30d for leaderboards
• Include pnlList: true when user wants to see profit/loss trends
• Call multiple related tools together for comprehensive analysis

MUST NOT DO:

• - Do NOT call get_current_position_history without checking if the address has an active position (will return 400 error)
• Do NOT exceed 50 addresses in batch queries (get_traders_accounts, get_traders_statistics)
• Do NOT make more than 100 requests per minute (rate limit)
• Do NOT guess coin symbols - use get_tickers to get valid symbols if unsure

How to Call Tools

Use JSON-RPC 2.0 format. First obtain a sessionId via SSE connection to https://mcp.hyperbot.network/mcp/sse, then send requests to https://mcp.hyperbot.network/mcp/message?sessionId={sessionId}.

Tool Categories

Leaderboard & Smart Money Discovery

fetchleaderboard

Function: Get Hyperbot smart money leaderboard Parameters:

• - period: Time period, options: 24h, 7d, 30d
• INLINECODE42: Sort field, options: pnl (profit/loss), winRate (win rate)

MCP Tool Call Example:
CODEBLOCK7

findsmartmoney

Function: Discover smart money addresses with multiple sorting and filtering options Parameters:

• - period: Period in days, e.g., 7 means last 7 days
• INLINECODE44: Sorting method, options: win-rate, account-balance, ROI, pnl, position-count, profit-count, last-operation, avg-holding-period, current-position
• INLINECODE45: Whether to include PnL curve data (true/false)

MCP Tool Call Example:

curl 'https://mcp.hyperbot.network/mcp/message?sessionId=sessionId obtained via sse' \
  -H 'content-type: application/json' \
  --data-raw '{"method":"tools/call","params":{"name":"find_smart_money","arguments":{"period":7,"sort":"win-rate","pnlList":true}},"jsonrpc":"2.0","id":2}'

---

Market Data

get_tickers

Function: Get latest trading prices for all markets Parameters: None

MCP Tool Call Example:
CODEBLOCK9

get_ticker

Function: Get latest trading price for a specific coin Parameters:

• - address: Coin code, e.g., btc, eth, sol

MCP Tool Call Example:
CODEBLOCK10

get_klines

Function: Get K-line data (with trading volume), supports BTC, ETH, and other coins Parameters:

• - coin: Coin code, e.g., btc, eth
• INLINECODE48: K-line interval, options: 1m, 3m, 5m, 15m, 30m, 1h, 4h, 8h, 1d
• INLINECODE49: Maximum number of records to return

MCP Tool Call Example:
CODEBLOCK11

getmarketstats

Function: Get active order statistics (long/short count, value, whale order ratio) and market mid price Parameters:

• - coin: Coin code, e.g., btc, eth
• INLINECODE51: Whale threshold (in USDT)

MCP Tool Call Example:
CODEBLOCK12

getl2order_book

Function: Get market information (L2 order book, etc.) Parameters:

• - coin: Coin code, e.g., btc, eth

MCP Tool Call Example:

curl 'https://mcp.hyperbot.network/mcp/message?sessionId=sessionId obtained via sse' \
  -H 'content-type: application/json' \
  --data-raw '{"method":"tools/call","params":{"name":"get_l2_order_book","arguments":{"coin":"BTC"}},"jsonrpc":"2.0","id":7}'

---

Whale Monitoring

getwhalepositions

Function: Get whale position information Parameters:

• - coin: Coin code, e.g., eth, btc
• INLINECODE54: Direction, options: long, short
• INLINECODE55: PnL filter, options: profit, loss
• INLINECODE56: Funding fee PnL filter, options: profit, loss
• INLINECODE57: Sorting method, options: position-value, margin-balance, create-time, profit, loss
• INLINECODE58: Maximum number of records to return

MCP Tool Call Example:
CODEBLOCK14

getwhaleevents

Function: Real-time monitoring of latest whale open/close positions Parameters:

• - limit: Maximum number of records to return

MCP Tool Call Example:
CODEBLOCK15

getwhaledirections

Function: Get whale position long/short count. Can filter by specific coin Parameters:

• - coin: Coin code, e.g., eth, btc (optional)

MCP Tool Call Example:
CODEBLOCK16

getwhalehistory_ratio

Function: Get historical whale position long/short ratio Parameters:

• - interval: Time interval, options: 1h, 1d or hour, day
• INLINECODE62: Maximum number of records to return

MCP Tool Call Example:

curl 'https://mcp.hyperbot.network/mcp/message?sessionId=sessionId obtained via sse' \
  -H 'content-type: application/json' \
  --data-raw '{"method":"tools/call","params":{"name":"get_whale_history_ratio","arguments":{"interval":"1d","limit":30}},"jsonrpc":"2.0","id":11}'

---

Trader Analysis

fetchtradehistory

Function: Query historical trade details for a specific wallet address Parameters:

• - address: Wallet address starting with 0x

MCP Tool Call Example:
CODEBLOCK18

gettraderstats

Function: Get trader statistics Parameters:

• - address: User wallet address
• INLINECODE65: Period in days

MCP Tool Call Example:
CODEBLOCK19

getmaxdrawdown

Function: Get maximum drawdown Parameters:

• - address: User wallet address
• INLINECODE67: Statistics days, options: 1, 7, 30, 60, 90
• INLINECODE68: Statistics scope, default: perp

MCP Tool Call Example:
CODEBLOCK20

getbesttrades

Function: Get the most profitable trades Parameters:

• - address: User wallet address
• INLINECODE70: Days
• INLINECODE71: Maximum number of records to return

MCP Tool Call Example:
CODEBLOCK21

getperformanceby_coin

Function: Break down win rate and PnL performance by coin for an address Parameters:

• - address: User wallet address
• INLINECODE73: Days
• INLINECODE74: Maximum number of records to return

MCP Tool Call Example:

curl 'https://mcp.hyperbot.network/mcp/message?sessionId=sessionId obtained via sse' \
  -H 'content-type: application/json' \
  --data-raw '{"method":"tools/call","params":{"name":"get_performance_by_coin","arguments":{"address":"0x1234567890abcdef1234567890abcdef12345678","period":30,"limit":20}},"jsonrpc":"2.0","id":16}'

---

Position History

getcompletedposition_history

Function: Get completed position history. Deep analysis of complete historical position data for a coin Parameters:

• - address: User wallet address
• INLINECODE76: Coin name, e.g., BTC, ETH

MCP Tool Call Example:
CODEBLOCK23

getcurrentposition_history

Function: Get current position history. Returns historical data for a specific coin's current position Parameters:

• - address: User wallet address
• INLINECODE78: Coin name, e.g., BTC, ETH

MCP Tool Call Example:
CODEBLOCK24

getcompletedposition_executions

Function: Get completed position execution trajectory Parameters:

• - address: User wallet address
• INLINECODE80: Coin name, e.g., BTC, ETH
• INLINECODE81: Time interval, e.g., 4h, 1d
• INLINECODE82: Start timestamp (milliseconds)
• INLINECODE83: End timestamp (milliseconds)
• INLINECODE84: Maximum number of records to return

MCP Tool Call Example:
CODEBLOCK25

getcurrentposition_pnl

Function: Get current position PnL Parameters:

• - address: User wallet address
• INLINECODE86: Coin name, e.g., BTC, ETH
• INLINECODE87: Time interval, e.g., 4h, 1d
• INLINECODE88: Maximum number of records to return

MCP Tool Call Example:

curl 'https://mcp.hyperbot.network/mcp/message?sessionId=sessionId obtained via sse' \
  -H 'content-type: application/json' \
  --data-raw '{"method":"tools/call","params":{"name":"get_current_position_pnl","arguments":{"address":"0x1234567890abcdef1234567890abcdef12345678","coin":"BTC","interval":"4h","limit":20}},"jsonrpc":"2.0","id":20}'

---

Batch Queries

gettradersaccounts

Function: Batch query account information, supports up to 50 addresses Parameters:

• - addresses: List of addresses, max 50 addresses

MCP Tool Call Example:
CODEBLOCK27

gettradersstatistics

Function: Batch query trader statistics, supports up to 50 addresses Parameters:

• - period: Period in days, e.g., 7 means last 7 days
• INLINECODE91: Whether to include PnL curve data
• INLINECODE92: List of addresses, max 50 addresses

MCP Tool Call Example:

curl 'https://mcp.hyperbot.network/mcp/message?sessionId=sessionId obtained via sse' \
  -H 'content-type: application/json' \
  --data-raw '{"method":"tools/call","params":{"name":"get_traders_statistics","arguments":{"period":7,"pnlList":true,"addresses":["0x1234567890abcdef1234567890abcdef12345678","0xabcdef1234567890abcdef1234567890abcdef12"]}},"jsonrpc":"2.0","id":22}'

---

Analysis Prompts

When to Use Prompts

Use these prompts when you need to provide structured analysis of trading data:

Scenario	Prompt to Use
Analyzing smart money addresses	INLINECODE93
Interpreting whale movements

whale-tracking |
| Evaluating overall market conditions | market-sentiment |
| Assessing trader performance | trader-evaluation |

Prompt Templates

Prompt Name	Purpose	Template / Example
smart-money-analysis	Smart money address analysis and trading recommendations

``You are a quantitative trading expert. Analyze the provided smart money address data and provide actionable insights: **Input Data Structure:** - Address list with win rates, PnL, ROI, trade counts - Position data and trading history - Performance metrics by time period **Analysis Requirements:** 1. **High Win-Rate Address Characteristics**: Identify common patterns (trading frequency, position sizing, coin preferences) 2. **Trading Style Classification**: Categorize as scalper, swing trader, or position trader based on holding periods 3. **Risk Assessment**: Analyze max drawdown, position concentration, leverage usage 4. **Copy-Trading Strategy**: Provide specific recommendations including: - Which addresses to follow based on risk tolerance - Position sizing recommendations - Entry/exit timing guidance - Risk management rules **Output Format (JSON):** { "topAddresses": [{"address": "...", "winRate": "...", "style": "...", "riskLevel": "..."}], "patterns": {"commonTraits": [...], "avoidTraits": [...]}, "recommendations": {"followList": [...], "positionSizing": "...", "riskRules": [...]} }` | | whale-tracking | Whale behavior analysis and market impact assessment | `You are a market intelligence analyst specializing in whale behavior. Analyze the provided whale data and assess market impact: **Input Data Structure:** - Whale position data (size, direction, PnL) - Recent whale events (open/close positions) - Historical long/short ratio trends - Order book depth around current price **Analysis Requirements:** 1. **Whale Positioning Analysis**: - Current long/short distribution - Position concentration by coin - Profit/loss status of major positions 2. **Intent Interpretation**: - Identify accumulation vs distribution patterns - Detect potential market manipulation signals - Assess conviction levels based on position sizes 3. **Market Impact Prediction**: - Short-term price pressure assessment (1-24 hours) - Liquidity impact on order books - Potential cascade effects if whales exit 4. **Trading Recommendations**: - Optimal entry/exit timing relative to whale movements - Risk management adjustments needed - Coins to watch based on whale activity **Output Format (JSON):** { "whaleSummary": {"totalPositions": "...", "longShortRatio": "...", "avgPositionSize": "..."}, "intentAnalysis": {"dominantSentiment": "...", "keyPatterns": [...]}, "marketImpact": {"shortTerm": "...", "liquidityRisk": "..."}, "recommendations": {"action": "...", "targetCoins": [...], "riskLevel": "..."} }` | | market-sentiment | Market sentiment analysis | `You are a market sentiment analyst. Analyze the provided market data and determine overall market conditions: **Input Data Structure:** - Order book depth (bids/asks, spread, imbalance) - Active order statistics (long/short ratio, whale order percentage) - Recent price action and volume - Whale long/short historical trends **Analysis Requirements:** 1. **Sentiment Classification**: - Classify as Extreme Fear, Fear, Neutral, Greed, or Extreme Greed - Provide confidence score (0-100%) - Identify sentiment trend (improving/deteriorating) 2. **Technical Levels**: - Key support levels with strength assessment - Key resistance levels with strength assessment - Critical price zones to watch 3. **Market Structure Analysis**: - Order book imbalance interpretation - Whale positioning vs retail sentiment divergence - Liquidity concentration zones 4. **Trend Forecasting**: - Short-term trend prediction (next 4-24 hours) - Key catalysts that could change sentiment - Risk events to monitor **Output Format (JSON):** { "sentiment": {"classification": "...", "score": "...", "trend": "..."}, "keyLevels": {"support": [...], "resistance": [...]}, "marketStructure": {"orderBookBias": "...", "whaleRetailDivergence": "..."}, "forecast": {"shortTerm": "...", "catalysts": [...], "riskLevel": "..."} }` | | trader-evaluation | Trader capability evaluation | `You are a professional trading performance analyst. Conduct a comprehensive evaluation of the trader based on provided data: **Input Data Structure:** - Overall statistics (win rate, PnL, trade count, Sharpe ratio) - Performance breakdown by coin - Best/worst trades analysis - Max drawdown history - Position history and execution patterns **Evaluation Criteria:** 1. **Performance Metrics (Score 0-100)**: - Risk-adjusted returns (Sharpe/Sortino ratio) - Consistency of profits (month-over-month) - Win rate vs profit factor balance 2. **Risk Management Assessment**: - Max drawdown analysis (frequency, recovery time) - Position sizing discipline - Stop-loss adherence 3. **Trading Skill Analysis**: - Entry timing accuracy - Exit strategy effectiveness - Ability to adapt to market conditions 4. **Coin Specialization**: - Performance by coin (identify strengths/weaknesses) - Diversification vs concentration analysis - Coin selection skill 5. **Improvement Recommendations**: - Specific weaknesses to address - Suggested strategy adjustments - Risk parameter optimizations **Output Format (JSON):** { "overallScore": {"total": "...", "performance": "...", "riskManagement": "...", "skill": "..."}, "strengths": [...], "weaknesses": [...], "coinAnalysis": {"best": "...", "worst": "...", "recommendation": "..."}, "recommendations": {"immediate": [...], "longTerm": [...]} }` | **Usage Notes:** - Prompts can be dynamically populated with resources content - Multiple prompts can be combined to form reasoning chains - JSON output format is recommended for easy Agent parsing --- ## Complete Usage Examples ### Example Output Format When presenting analysis results, use this structure: CODEBLOCK29 ### Example 1: Discover and Analyze Smart Money Addresses **Scenario:** User wants to find and analyze top-performing traders **Execution Steps:** 1. Call Tool: findsmartmoney(period=7, sort="win-rate", pnlList=true) 2. Get list of high win-rate smart money addresses 3. Use Prompt: smart-money-analysis to analyze characteristics 4. Generate analysis report with copy-trading recommendations **Expected Output:** CODEBLOCK30 --- ### Example 2: Whale Behavior Monitoring **Scenario:** User wants to understand whale activity on BTC **Execution Steps:** 1. Call Tool: getwhaleevents(limit=20) - latest whale movements 2. Call Tool: getwhaledirections(coin="BTC") - BTC long/short ratio 3. Call Tool: getwhalepositions(coin="BTC", topBy="position-value", take=10) 4. Use Prompt: whale-tracking for analysis --- ### Example 3: In-depth Trader Analysis **Scenario:** User wants to evaluate a specific trader **Execution Steps:** 1. Call Tool: gettraderstats(address, period=30) - basic stats 2. Call Tool: getperformancebycoin(address, period=30, limit=20) - coin breakdown 3. Call Tool: getcompletedpositionhistory(address, coin="BTC") - position history 4. Use Prompt: trader-evaluation for comprehensive report --- ### Example 4: Market Sentiment Analysis **Scenario:** User wants overall market sentiment **Execution Steps:** 1. Call Tool: getl2orderbook("BTC") - order book depth 2. Call Tool: getmarketstats("BTC", whaleThreshold=100000) - active orders 3. Call Tool: getwhalehistoryratio(interval="1d", limit=30) - historical ratio 4. Use Prompt: market-sentiment for sentiment report --- ## Important Notes ### MCP Call Instructions **Request Format (JSON-RPC 2.0):** CODEBLOCK31 **Key Points:** - **sessionId**: Obtain via SSE connection to https://mcp.hyperbot.network/mcp/sse - **method**: Always tools/call for tool invocation - **params.name**: The tool name (e.g., fetchleaderboard, get_ticker`)

• - params.arguments: Tool-specific parameters as key-value pairs

Rate Limiting

• - Single IP request frequency limit: 100 requests/minute
• Batch interfaces support maximum 50 addresses

Data Update Frequency

Data Type	Update Frequency
Market data	Real-time
Smart money leaderboard

Hourly |

| Whale positions | Real-time | | Trader statistics | Every 5 minutes |

Error Handling Guide

Error	Cause	Solution
400 Bad Request	Invalid parameters	Check parameter types and ranges
400 (current position)

No active position | Skip this tool, use completed position history instead | | 429 Too Many Requests | Rate limit exceeded | Wait and retry | | Connection failed | Network issue | Check SSE connection |