Conduzindo Lições Aprendidas Pós-Incidente

Quando Usar

• Após qualquer incidente de segurança ter sido totalmente resolvido e a recuperação concluída
• Após exercícios de simulação (tabletop exercises) ou simulações de resposta a incidentes (IR)
• Após eventos de "quase acidente" significativos
• Revisão trimestral de tendências acumuladas de incidentes
• Quando os playbooks de resposta a incidentes precisam ser atualizados com base na experiência real

Pré-requisitos

• Incidente totalmente resolvido (contenção, erradicação e recuperação completas)
• Cronograma e documentação do incidente coletados
• Todos os respondedores de incidentes disponíveis para a sessão de revisão
• Espaço de reunião para discussão colaborativa
• Dados do sistema de tickets de incidentes para análise de métricas

Fluxo de Trabalho

Etapa 1: Coletar Dados do Incidente

# Export incident timeline from ticketing system
curl -s "https://thehive.local/api/v1/case/$CASE_ID/timeline" \
  -H "Authorization: Bearer $THEHIVE_API_KEY" | jq '.' > incident_timeline.json

# Extract detection and response metrics from SIEM
index=notable incident_id="IR-2024-042"
| stats min(_time) as first_alert, max(_time) as last_alert,
  count as total_alerts, dc(src) as unique_sources

# Compile all responder actions and timestamps
grep -E "timestamp|action|analyst" /var/log/ir/IR-2024-042/*.json | \
  python3 -m json.tool > compiled_actions.json

Etapa 2: Realizar Reunião de Pós-Mortem sem Culpa

Structured Agenda (90 minutes):
1. Incident summary (5 min) - Factual overview
2. Timeline walkthrough (20 min) - Chronological events
3. What worked well (15 min) - Positive outcomes
4. What needs improvement (15 min) - Gaps and failures
5. Root cause analysis (15 min) - 5 Whys or fishbone
6. Action items (10 min) - Specific improvements with owners
7. Playbook updates (10 min) - Changes to IR procedures

Blameless Principles:
- Focus on systems and processes, not individuals
- Assume best intentions with available information
- Seek to understand, not to blame

Etapa 3: Realizar Análise da Causa Raiz

# 5 Whys analysis example:
# Why 1: Why did ransomware encrypt production servers?
#   Answer: Attacker had domain admin credentials
# Why 2: Why did attacker have domain admin credentials?
#   Answer: Kerberoasted a service account and cracked it
# Why 3: Why was the service account password crackable?
#   Answer: Used a 12-character dictionary-based password
# Why 4: Why was the service account password weak?
#   Answer: No enforcement of service account password policy
# Why 5: Why was there no service account password policy?
#   Answer: PAM was not implemented for service accounts
# ROOT CAUSE: Lack of privileged access management

Etapa 4: Calcular Métricas de Resposta

from datetime import datetime
events = {
    'compromise': '2024-01-10 14:00:00',
    'detection': '2024-01-15 08:30:00',
    'triage': '2024-01-15 08:45:00',
    'containment': '2024-01-15 09:30:00',
    'eradication': '2024-01-16 14:00:00',
    'recovery': '2024-01-18 16:00:00',
    'closure': '2024-01-25 10:00:00',
}
fmt = '%Y-%m-%d %H:%M:%S'
times = {k: datetime.strptime(v, fmt) for k, v in events.items()}
print(f"Dwell Time: {times['detection'] - times['compromise']}")
print(f"MTTD: {times['triage'] - times['detection']}")
print(f"MTTC: {times['containment'] - times['detection']}")
print(f"MTTR: {times['recovery'] - times['eradication']}")
print(f"Total Duration: {times['closure'] - times['detection']}")

Etapa 5: Documentar Descobertas e Criar Itens de Ação

# Create tracked action items in project management
curl -X POST "https://jira.local/rest/api/2/issue" \
  -H "Authorization: Bearer $JIRA_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "fields": {
      "project": {"key": "SEC"},
      "summary": "Implement PAM for service accounts (IR-2024-042)",
      "issuetype": {"name": "Task"},
      "priority": {"name": "High"},
      "assignee": {"name": "security_engineer"},
      "duedate": "2024-03-15"
    }
  }'

Etapa 6: Atualizar Playbooks e Regras de Detecção

# New Sigma detection rule based on incident learnings
title: Kerberoasting Activity Detected
status: stable
description: Detects Kerberoasting based on IR-2024-042 lessons
logsource:
  product: windows
  service: security
detection:
  selection:
    EventID: 4769
    TicketEncryptionType: '0x17'
  condition: selection
level: high
tags:
  - attack.credential_access
  - attack.t1558.003

Conceitos Chave

…