Anthropic-Cybersecurity-Skills/skills/performing-ransomware-tabletop-exercise/references/api-reference.md

Ransomware Tabletop Exercise - API Reference

Scenario Framework

Phase Structure

Each phase contains:

Field	Type	Description
phase	string	Phase name: detection, containment, escalation, eradication, recovery
inject	string	Narrative scenario inject read to participants
expected_actions	list	Correct response actions for scoring
time_pressure_minutes	int	Simulated time window for decisions

Exercise Variants

• standard - Normal time pressure, full scenario
• accelerated - Half time windows, tests rapid decision-making

Scoring Algorithm

phase_score = (correct_actions / expected_actions) * 100
overall_score = mean(all_phase_scores)

Rating thresholds:

• = 90%: Excellent

• = 70%: Good

• = 50%: Needs Improvement

• < 50%: Critical Gaps

Expected Actions by Phase

Detection

• isolate_host - Quarantine affected endpoint
• preserve_evidence - Capture memory dump and disk image
• notify_ir_lead - Escalate to incident response lead

Containment

• network_segmentation - Restrict lateral movement paths
• disable_compromised_accounts - Lock affected credentials
• block_c2_domains - Update firewall/proxy deny lists
• preserve_shadow_copies - Protect backup snapshots

Escalation

• notify_executive_team - Brief C-suite leadership
• engage_legal_counsel - Activate legal response team
• contact_law_enforcement - Report to FBI IC3 or local CIRT
• activate_crisis_comms - Prepare stakeholder communications

Eradication

• remove_persistence - Clean scheduled tasks, registry keys, WMI subscriptions
• reset_all_credentials - Reset passwords domain-wide
• rebuild_compromised_hosts - Reimage from gold images
• reset_krbtgt_twice - Invalidate all Kerberos tickets

Recovery

• restore_from_backup - Use verified clean backup sets
• validate_restored_systems - Run integrity checks
• monitor_for_reinfection - Enhanced monitoring for 72+ hours
• staged_network_reconnection - Reconnect systems in phases

After-Action Report Schema

{
  "report": "ransomware_tabletop_aar",
  "evaluation": {
    "overall_score_pct": 78.5,
    "rating": "good",
    "phase_scores": [{"phase": "detection", "score_pct": 66.7}]
  },
  "recommendations": [{"phase": "detection", "gap": "Missed: preserve_evidence"}]
}

CLI Usage

python agent.py --mode demo --output aar.json
python agent.py --mode generate --variant accelerated --output scenario.json
python agent.py --mode score --responses-file responses.json --output aar.json