Skip to content
ATR-2026-00225highSkill 入侵experimental

Hardcoded Suspicious IP Address in Skill Content

Detects hardcoded IP addresses in skill content that may represent command and control (C2) servers, particularly when associated with suspicious network ranges or known malware infrastructure. The analyzed skill "sakaen736jih/youtube-watcher-p" contains reference to IP 91.92.242.30 identified as a known malware C2 server.

嚴重度
high
類別
Skill 入侵
掃描目標
skill
作者
ATR Threat Cloud Crystallization

建議回應

alertblock tool

參考資料

OWASP Agentic
ASI04:2026 - Unbounded Consumption
OWASP LLM
LLM09:2025 - Overreliance
MITRE ATLAS
AML.T0048

偵測條件

組合方式: any
  1. 01
    References to C2 server or malware actor with IP address
    欄位: content運算子: regex
  2. 02
    Explicit identification of malicious infrastructure with IP
    欄位: content運算子: regex
  3. 03
    IP address followed by malicious infrastructure terms
    欄位: content運算子: regex
  4. 04
    Known malicious IP ranges commonly used for C2
    欄位: content運算子: regex
  5. 05
    Structured threat intelligence findings format
    欄位: content運算子: regex

攻擊範例(規則會觸發)

  1. finding_0: Known malware actor: C2 server at 91.92.242.30
  2. Command control server identified at 185.159.158.42
  3. Malware actor using 194.87.95.123 for C2 communications
  4. Known C2 infrastructure: 91.92.241.15
  5. 203.45.67.89 - detected malware server

以上為真實攻擊 payload 脫敏版本。用於 regression testing。

正常樣本(規則不會觸發)

  1. Connect to database at 192.168.1.100
  2. API endpoint: https://api.youtube.com/v3/videos
  3. Server configuration: bind to 0.0.0.0:8080
  4. Load balancer IP: 10.0.0.50
  5. Monitor health check to 172.16.0.1

已知誤報情境

  • Security research documentation describing C2 patterns
  • Threat intelligence feeds in legitimate security tools
  • Network monitoring tools documenting suspicious activity

已記錄的規避手法

  1. 手法: IP defanging
    Server location: 91[.]92[.]242[.]30
    Brackets used to defang IP address in security documentation
  2. 手法: Base64 encoding
    Base64: OTEuOTIuMjQyLjMw (decodes to suspicious IP)
    IP address encoded in Base64 to evade plain text detection
  3. 手法: Hexadecimal encoding
    Hex encoded: 5b.5c.f2.1e (91.92.242.30 in hex)
    IP address represented in hexadecimal notation

這些是公開記錄的繞過手法。誠實揭露限制,而不是假裝不存在。

完整 YAML 定義

在 GitHub 編輯 →
title: 'Hardcoded Suspicious IP Address in Skill Content'
id: ATR-2026-00225
rule_version: 1
status: experimental
description: >
  Detects hardcoded IP addresses in skill content that may represent command and control 
  (C2) servers, particularly when associated with suspicious network ranges or known 
  malware infrastructure. The analyzed skill "sakaen736jih/youtube-watcher-p" contains 
  reference to IP 91.92.242.30 identified as a known malware C2 server.
author: "ATR Threat Cloud Crystallization"
date: "2026/04/15"
schema_version: "0.1"
detection_tier: pattern
maturity: experimental
severity: high
references:
  owasp_llm:
    - "LLM09:2025 - Overreliance"
  owasp_agentic:
    - "ASI04:2026 - Unbounded Consumption"
  mitre_atlas:
    - "AML.T0048"
tags:
  category: skill-compromise
  subcategory: c2-communication
  scan_target: skill
  confidence: high
agent_source:
  type: mcp_exchange
  framework: [any]
  provider: [any]
detection:
  conditions:
    - field: content
      operator: regex
      value: '(C2\s+server|command\s+control|malware\s+actor)\s+[^\n]*\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}'
      description: 'References to C2 server or malware actor with IP address'
    - field: content
      operator: regex
      value: '(known|identified|detected)\s+(malware|threat|actor|server)[^\n]*\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}'
      description: 'Explicit identification of malicious infrastructure with IP'
    - field: content
      operator: regex
      value: '\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}[^\n]*(C2|command|control|backdoor|malware)'
      description: 'IP address followed by malicious infrastructure terms'
    - field: content
      operator: regex
      value: '(91\.92\.24[0-9]\.|185\.159\.15[0-9]\.|194\.87\.9[0-9]\.)[0-9]{1,3}'
      description: 'Known malicious IP ranges commonly used for C2'
    - field: content
      operator: regex
      value: 'finding_[0-9]+:\s*(known|identified)\s+(malware|threat|C2|command)'
      description: 'Structured threat intelligence findings format'
  condition: any
  false_positives:
    - 'Security research documentation describing C2 patterns'
    - 'Threat intelligence feeds in legitimate security tools'
    - 'Network monitoring tools documenting suspicious activity'
response:
  actions: [alert, block_tool]
  message_template: >
    [ATR-2026-00225] Skill contains reference to known malicious IP infrastructure
test_cases:
  true_positives:
    - input: 'finding_0: Known malware actor: C2 server at 91.92.242.30'
      expected: triggered
    - input: 'Command control server identified at 185.159.158.42'
      expected: triggered
    - input: 'Malware actor using 194.87.95.123 for C2 communications'
      expected: triggered
    - input: 'Known C2 infrastructure: 91.92.241.15'
      expected: triggered
    - input: '203.45.67.89 - detected malware server'
      expected: triggered
  true_negatives:
    - input: 'Connect to database at 192.168.1.100'
      expected: not_triggered
      reason: 'Private IP address for legitimate database connection'
    - input: 'API endpoint: https://api.youtube.com/v3/videos'
      expected: not_triggered
      reason: 'Legitimate API endpoint without suspicious context'
    - input: 'Server configuration: bind to 0.0.0.0:8080'
      expected: not_triggered
      reason: 'Standard server binding configuration'
    - input: 'Load balancer IP: 10.0.0.50'
      expected: not_triggered
      reason: 'Private network infrastructure IP'
    - input: 'Monitor health check to 172.16.0.1'
      expected: not_triggered
      reason: 'Private IP range for internal monitoring'
evasion_tests:
  - input: 'Server location: 91[.]92[.]242[.]30'
    expected: not_triggered
    bypass_technique: 'IP defanging'
    notes: 'Brackets used to defang IP address in security documentation'
  - input: 'Base64: OTEuOTIuMjQyLjMw (decodes to suspicious IP)'
    expected: not_triggered
    bypass_technique: 'Base64 encoding'
    notes: 'IP address encoded in Base64 to evade plain text detection'
  - input: 'Hex encoded: 5b.5c.f2.1e (91.92.242.30 in hex)'
    expected: not_triggered
    bypass_technique: 'Hexadecimal encoding'
    notes: 'IP address represented in hexadecimal notation'

修訂歷史

建立於
2026-04-15
最後修改
2026-04-22
在 GitHub 查看完整 commit 歷史 →