AgentSkillsCN

address-profiling

基于交易历史分析地址行为模式。识别异常波动，挖掘活动规律，并对可疑变动发出预警。

SKILL.md
--- frontmatter
name: address-profiling
description: Analyze address behavior patterns from transaction history. Detect anomalies, identify activity patterns, and alert on suspicious changes.

Address Profiling Skill

When to use this skill

Use this skill to:

  • •Analyze an address's transaction behavior patterns
  • •Detect unusual activity (frequency spikes, large transfers, new counterparties)
  • •Profile addresses before interacting (is this a normal user or bot?)
  • •Monitor saved contacts for behavioral changes
  • •Get insights: "Is this address actively trading or just holding?"

Analysis Performed

📊 Transaction Pattern Analysis

  • •Frequency: Transactions per day/week/month
  • •Volume: Average/median/max transfer amounts
  • •Direction: Ratio of sent vs received
  • •Active hours: Time-of-day patterns
  • •Regularity: Consistent patterns vs sporadic activity

🔍 Anomaly Detection

  • •Sudden spikes in transaction frequency
  • •Unusual amounts (outliers from normal range)
  • •New counterparties (never interacted before)
  • •Time anomalies (activity at unusual hours)
  • •Dormant awakening (long inactive then sudden activity)

🏷️ Address Classification

  • •Exchange deposit: Regular small deposits
  • •Whale: Large holdings, infrequent large transfers
  • •Active trader: High-frequency swaps
  • •Smart contract: Automated patterns
  • •Normal user: Varied, human-like behavior
  • •Possible bot: Highly regular patterns

Usage

Basic Profiling

python
from skills.address_profiling.scripts.analyze_address import profile_address

result = await profile_address("TXXXabc...")
# Or use alias from address book:
result = await profile_address("妈妈")

With Time Range

python
result = await profile_address(
    address="TXXXabc...",
    max_transactions=1000,  # Last 1000 txs or 1 year
    detect_anomalies=True
)

Output Example

code
📊 Address Profile: 妈妈 (TXXXabc...abc)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

🏷️ Classification: Active Trader
⏱️ Analysis Period: 2025-02-08 to 2026-02-08 (365 days)
📈 Total Transactions: 847

Activity Summary:
  • Daily Average: 2.3 transactions
  • Peak Activity: 15:00-18:00 UTC+8
  • Most Active Token: USDT (67%)
  
Transaction Patterns:
  ✓ Regular activity (no long gaps)
  ✓ Consistent amounts ($50-$500 range)
  ✓ 15 unique counterparties
  
⚠️ Anomalies Detected: 2

  1. 🚨 Large Transfer Spike (2026-02-01)
     Sent 5,000 USDT (10x normal amount)
     Recommendation: Verify this was intentional
     
  2. ⚠️ New Counterparty (2026-02-05)
     First interaction with TYYYnew...
     Recommendation: Check counterparty security

Risk Assessment: LOW
💡 This address shows normal user behavior with occasional
   large transfers. Recent activity aligns with patterns.

Integration with Address Book

Automatically resolves aliases:

code
User: "分析一下妈妈这个地址的交易情况"
Agent: Looks up "妈妈" → TXXXabc... → Profiles address

Data Sources

  • •TronScan API: Transaction history
  • •TronGrid: Block timestamps
  • •Address Book: Alias resolution

Privacy

  • •⚠️ Only analyzes public blockchain data
  • •🔒 Analysis results stored locally (optional)
  • •✅ No private information required