performing-deception-technology-deployment

Imported from mukul975/anthropic-cybersecurity-skills (skills/performing-deception-technology-deployment/SKILL.md). Install upstream with npx skills add mukul975/anthropic-cybersecurity-skills --skill performing-deception-technology-deployment. Copyright stays with the author (Apache-2.0).

Performing Deception Technology Deployment

When to Use

Use this skill when:

• SOC teams need high-fidelity detection of post-compromise lateral movement with near-zero false positives
• Existing detection tools miss advanced attackers who avoid triggering threshold-based alerts
• The organization wants to detect credential abuse by planting fake credentials as honeytokens
• Network segmentation gaps need compensating detection controls

Do not use as a replacement for fundamental security controls (patching, EDR, network segmentation) — deception is a detection layer, not a prevention mechanism.

Prerequisites

• Network segments identified for honeypot/decoy deployment (server VLANs, DMZ, OT networks)
• Deception platform (Thinkst Canary, Attivo/SentinelOne Hologram, or open-source alternatives)
• SIEM integration for deception alerts (any interaction with deception assets is suspicious)
• Active Directory access for honeytoken account and credential creation
• Network team coordination for IP allocation and traffic routing

Workflow

Step 1: Map Attack Surface for Deception Placement

Identify high-value network segments where attackers would traverse:

DECEPTION DEPLOYMENT MAP
━━━━━━━━━━━━━━━━━━━━━━━━
Segment              Decoy Type          Rationale
Server VLAN          Fake file server    Attackers enumerate SMB shares during recon
Database VLAN        Fake DB server      SQL scanning detected in past incidents
AD/DC Segment        Honeytoken account  Credential theft detection
Executive Subnet     Fake workstation    Targeted attacks pivot through exec systems
DMZ                  Honeypot web app    External attacker detection
OT Network           Fake PLC/HMI        Industrial threat detection
Cloud (AWS VPC)      Canary EC2 + S3     Cloud lateral movement detection

Step 2: Deploy Thinkst Canary Devices

Configure Canary devices mimicking real infrastructure:

Windows File Server Canary:

{
  "device_name": "FILESERVER-BK04",
  "personality": "windows-server-2019",
  "services": {
    "smb": {
      "enabled": true,
      "shares": ["Finance_Backup", "HR_Archive", "IT_Docs"],
      "files": [
        {"name": "Q4_Revenue_2024.xlsx", "alert_on": "read"},
        {"name": "employee_ssn_export.csv", "alert_on": "read"},
        {"name": "admin_passwords.kdbx", "alert_on": "read"}
      ]
    },
    "rdp": {"enabled": true},
    "http": {"enabled": false}
  },
  "network": {
    "ip": "10.0.5.200",
    "hostname": "FILESERVER-BK04",
    "domain": "company.local"
  },
  "alert_webhook": "https://soar.company.com/api/webhook/canary"
}

Database Server Canary:

{
  "device_name": "DB-ARCHIVE-02",
  "personality": "linux-mysql",
  "services": {
    "mysql": {
      "enabled": true,
      "port": 3306,
      "databases": ["customer_pii", "payment_archive"],
      "alert_on_login_attempt": true
    },
    "ssh": {
      "enabled": true,
      "port": 22,
      "alert_on_login_attempt": true
    }
  },
  "network": {
    "ip": "10.0.10.50",
    "hostname": "db-archive-02"
  }
}

Step 3: Deploy Honeytokens in Active Directory

Create fake privileged accounts that should never be used:

# Create honeytoken service account
New-ADUser -Name "svc_sql_backup" `
    -SamAccountName "svc_sql_backup" `
    -UserPrincipalName "svc_sql_backup@company.local" `
    -Description "SQL Backup Service Account - DO NOT DELETE" `
    -AccountPassword (ConvertTo-SecureString "FakeP@ssw0rd2024!" -AsPlainText -Force) `
    -Enabled $true `
    -PasswordNeverExpires $true `
    -CannotChangePassword $true

# Add to a group that looks attractive (but monitor for any use)
Add-ADGroupMember -Identity "Domain Admins" -Members "svc_sql_backup"

# Place cached credentials on decoy workstation
# (Mimikatz/credential dumping will find these)
cmdkey /add:fileserver-bk04.company.local /user:company\svc_sql_backup /pass:FakeP@ssw0rd2024!

Monitor honeytoken usage in Splunk:

index=wineventlog sourcetype="WinEventLog:Security"
(EventCode=4624 OR EventCode=4625 OR EventCode=4648 OR EventCode=4768 OR EventCode=4769)
TargetUserName="svc_sql_backup"
| eval alert_severity = "CRITICAL"
| eval alert_message = "HONEYTOKEN ACCOUNT USED — Likely credential theft detected"
| table _time, EventCode, src_ip, ComputerName, TargetUserName, Logon_Type, alert_message

Step 4: Deploy Canary Files and Documents

Plant tracked documents that beacon when opened:

Canary Document (Word doc with tracking):

# Using Thinkst Canary API to create a canary token document
import requests

response = requests.post(
    "https://YOURCOMPANY.canary.tools/api/v1/canarytoken/create",
    data={
        "auth_token": "YOUR_API_TOKEN",
        "kind": "doc-msword",
        "memo": "Finance backup folder canary document",
        "flock_id": "flock:default"
    }
)
token = response.json()
download_url = token["canarytoken"]["canarytoken_url"]
print(f"Download canary doc: {download_url}")
# Place this document in honeypot SMB shares and sensitive directories

AWS Canary Token (S3 access key):

# Create AWS canary token — alerts when access key is used
response = requests.post(
    "https://YOURCOMPANY.canary.tools/api/v1/canarytoken/create",
    data={
        "auth_token": "YOUR_API_TOKEN",
        "kind": "aws-id",
        "memo": "Canary AWS key in developer laptop .aws/credentials"
    }
)
aws_keys = response.json()
print(f"Access Key: {aws_keys['canarytoken']['access_key_id']}")
print(f"Secret Key: {aws_keys['canarytoken']['secret_access_key']}")
# Plant in .aws/credentials on developer workstations

Step 5: Integrate Deception Alerts with SIEM/SOAR

All deception alerts are high-fidelity — any interaction is suspicious:

Splunk Alert for Canary Triggers:

index=canary sourcetype="canary:alerts"
| eval severity = "CRITICAL"
| eval confidence = "HIGH — Deception asset triggered, zero false positive expected"
| table _time, canary_name, alert_type, source_ip, service, details
| sendalert create_notable param.rule_title="Deception Alert — Canary Triggered"
  param.severity="critical" param.drilldown_search="index=canary source_ip=$source_ip$"

SOAR Automated Response:

def canary_triggered(container):
    """Auto-response for deception alerts — high confidence, no approval needed"""
    source_ip = container["artifacts"][0]["cef"]["sourceAddress"]

    # Immediately isolate the source
    phantom.act("quarantine device",
                parameters=[{"ip_hostname": source_ip}],
                assets=["crowdstrike_prod"],
                name="isolate_attacker_host")

    # Block at firewall
    phantom.act("block ip",
                parameters=[{"ip": source_ip, "direction": "both"}],
                assets=["palo_alto_prod"],
                name="block_attacker_ip")

    # Create high-priority incident
    phantom.act("create ticket",
                parameters=[{
                    "short_description": f"DECEPTION ALERT: Canary triggered from {source_ip}",
                    "urgency": "1",
                    "impact": "1"
                }],
                assets=["servicenow_prod"])

    phantom.set_severity(container, "critical")

Step 6: Maintain Deception Realism

Regularly update decoys to maintain believability:

• Rotate honeytoken passwords quarterly (update cached credentials on decoy workstations)
• Update canary file modification dates to appear recently accessed
• Add realistic network traffic to honeypots (scheduled SMB enumeration, DNS lookups)
• Register honeypot hostnames in DNS and Active Directory to appear in network scans
• Update canary document contents to match current business context

Key Concepts

Tools & Systems

• Thinkst Canary: Commercial deception platform offering hardware/virtual canaries and canary tokens
• Canarytokens.org: Free honeytoken generation service (DNS, HTTP, AWS keys, Word docs, SQL queries)
• Attivo Networks (SentinelOne): Enterprise deception platform with AD decoys and endpoint breadcrumbs
• HoneyDB: Community honeypot data aggregation platform for threat intelligence sharing
• T-Pot: Open-source multi-honeypot platform combining 20+ honeypot types in a Docker deployment

Common Scenarios

• Lateral Movement Detection: Attacker enumerates SMB shares and accesses honeypot file server — immediate high-fidelity alert
• Credential Theft Discovery: Mimikatz dumps honeytoken cached credentials — usage of fake account triggers alert
• Cloud Key Compromise: Stolen AWS canary token used from external IP — detects supply chain or insider compromise
• Ransomware Early Warning: Ransomware encrypts canary files on honeypot shares — early detection before production systems affected
• Insider Threat Signal: Employee accesses honeypot "salary database" — indicates unauthorized data exploration

Output Format

DECEPTION ALERT — CRITICAL
━━━━━━━━━━━━━━━━━━━━━━━━━━
Time:         2024-03-15 14:23:07 UTC
Canary:       FILESERVER-BK04 (10.0.5.200)
Service:      SMB — File share "Finance_Backup" accessed
Source:       192.168.1.105 (WORKSTATION-042, Finance Dept)
User:         company\jsmith
File Accessed: Q4_Revenue_2024.xlsx (canary document)

Alert Confidence: HIGH — No legitimate reason to access deception asset
False Positive Likelihood: <1%

Automated Response:
  [DONE] WORKSTATION-042 isolated via CrowdStrike
  [DONE] 192.168.1.105 blocked at firewall (bidirectional)
  [DONE] Incident INC0012567 created (P1 — Critical)
  [PENDING] Tier 2 investigation — determine if workstation compromised or insider threat