Packetscope Tracer

PacketScope Tracer API Skill & MCP Server

This skill enables LLM to interact with the PacketScope Tracer module API for route tracing, risk analysis, and history queries.

Overview

Tracer is a network path analysis module that provides:

• Real-time traceroute with ICMP and TCP protocols
• Geographic and ASN enrichment for each hop
• Route anomaly detection and risk scoring
• Historical route tracking and comparison
• Malicious IP detection against threat intelligence feeds

MCP Server Tools

The Tracer MCP Server provides the following tools for LLM interaction:

Route Tracing Tools

• trace_target - Trace a network target and return hop-by-hop results

  • Parameters: target (string), use_cache (bool, default: True), protocol ("icmp" or "tcp", default: "icmp"), port (int, optional, required for TCP)
  • Returns: Target, resolved IP, source (cache/live), hop list with geo/ASN info

• get_trace_detail - Get detailed information about a specific hop

  • Parameters: target (string), hop_index (int, 0-based)
  • Returns: Detailed hop info including IP, latency, jitter, packet loss, location, ASN, ISP, geo coordinates

• compare_routes - Compare current route against historical routes

  • Parameters: target (string)
  • Returns: New IPs, removed IPs, latency changes between current and historical path

Risk Analysis Tools

• analyze_target - Run anomaly analysis and calculate risk score
  • Parameters: target (string), cache (bool, default: True)
  • Returns: Anomalies list, alerts list, riskScore (0-100), riskLevel (low/medium/high)

History Tools

• get_history - Get traceroute history records
  • Parameters: target (string, optional), limit (int, default: 20)
  • Returns: Dictionary of history records keyed by target IP

Status Check Tools

• health_check - Check server health and readiness

  • Parameters: None
  • Returns: Health status, readiness, timestamp

• server_capabilities - Get server capabilities and tool usage examples

  • Parameters: None
  • Returns: Server capabilities, tool list, natural language examples

API Endpoints

Base URL

http://localhost:8000

Route Tracing

Trace Target

• Endpoint: GET /api/trace
• Description: Run a traceroute to the target
• Query Parameters:
  • target: IP address or domain (required)
  • use_cache: Use cached results if available (default: "true")
  • protocol: "icmp" or "tcp" (default: "icmp")
  • port: Port number, required when protocol=tcp (1-65535)

Response (streaming NDJSON, one hop per line):

{"hop": 1, "ip": "192.168.1.1", "latency": 1.23, "jitter": 0.5, "packet_loss": "0.0%", "bandwidth_mbps": 47.17, "location": "Beijing, China", "asn": "4134", "isp": "China Telecom", "geo": {"lat": 39.9, "lon": 116.4, "radius_km": 50, "timezone": "Asia/Shanghai"}}

Risk Analysis

Analyze Target

• Endpoint: GET /api/analyze
• Description: Run anomaly analysis and risk scoring
• Query Parameters:
  • target: IP address or domain (required)
  • cache: Use cached trace results (default: "true")

Response:

{
  "anomalies": [
    {"type": "PathDeviation", "detail": "跳点 5 出现新IP 10.0.0.1"},
    {"type": "HighLatency", "detail": "跳点 5 (10.0.0.1) 延迟过高 250ms"}
  ],
  "alerts": [
    "跳点 10.0.0.1 被列为恶意IP: Spamhaus DROP list"
  ],
  "riskScore": 50
}

History

Get History

• Endpoint: GET /api/history
• Description: Get traceroute history records
• Query Parameters:
  • target: IP or domain filter (optional)

Response:

{
  "8.8.8.8": [
    {
      "timestamp": "20240101-120000",
      "protocol": "icmp",
      "result": [...]
    }
  ]
}

Status Check

Readiness Check

• Endpoint: GET /api/ready
• Description: Check if the Tracer service is ready

Response:

{
  "ready": true,
  "timestamp": "2024-01-01T12:00:00.000000"
}

Usage Examples

Python Client Example

from tracer_client import TracerClient

# Create client
client = TracerClient("http://localhost:8000")

# ICMP trace
result = client.trace("8.8.8.8")
print(f"Hops: {len(result.hops)}, Source: {result.source}")

# TCP trace
result = client.trace("1.1.1.1", protocol="tcp", port=443)
print(f"Hops: {len(result.hops)}")

# Risk analysis
analysis = client.analyze("8.8.8.8")
print(f"Risk Score: {analysis.risk_score}, Anomalies: {len(analysis.anomalies)}")

# History
history = client.get_history(target="8.8.8.8")
print(f"Records: {len(history.get('8.8.8.8', []))}")

Risk Score Levels

Anomaly Types