Decepticon Red Team

Decepticon Red Team

Overview

Decepticon is an autonomous red team agent that executes full attack chains end-to-end. It operates as an AI-driven adversary simulation platform with domain specialists for Active Directory, cloud, smart contracts, reverse engineering, and analysis. Engagement packages structure operations around RoE, ConOps, and OPPLAN with MITRE ATT&CK mapping.

Source: https://github.com/PurpleAILAB/Decepticon Architecture: Modular kill chain with domain-specialist agents License: Open source

When to Use

Trigger phrases:

• "decepticon red team"

• "Autonomous red teaming of infrastructure"

• "Full attack chain simulation from recon to post-exploitation"

• "Engagement planning with structured packages (RoE, ConOps, OPPLAN)"

• Autonomous red teaming of infrastructure

• Full attack chain simulation from recon to post-exploitation

• Engagement planning with structured packages (RoE, ConOps, OPPLAN)

• MITRE ATT&CK technique mapping and adversary emulation

• Multi-domain assessments (AD + cloud + application)

Kill Chain Phases

This section covers kill chain phases for decepticon red team.

• Ensure all prerequisites are met before proceeding
• Follow the documented workflow steps in sequence
• Record results and any anomalies encountered during this phase

1. Orchestration

Central planning engine that coordinates all phases. Determines attack strategy, allocates resources across domain specialists, and adapts the plan based on findings at each phase.

2. Reconnaissance

Automated target discovery and profiling:

• Passive OSINT (DNS, WHOIS, certificate transparency, social media)
• Active enumeration (subdomains, ports, services, technology fingerprinting)
• Attack surface mapping and priority ranking
• Credential harvesting from public breach databases

3. Exploitation

Vulnerability exploitation and initial access:

• Web application exploitation (OWASP Top 10)
• Network service exploitation (known CVEs, misconfigurations)
• Phishing payload delivery and social engineering
• Credential-based access (password spraying, default creds)

4. Post-Exploitation

Establishing persistence and expanding access:

• Privilege escalation (local and domain)
• Lateral movement across network segments
• Credential harvesting from memory and storage
• Data staging for exfiltration
• Persistence mechanism installation

5. Vulnerability Research

Deep-dive vulnerability discovery:

• Custom fuzzing based on discovered services
• Zero-day research on proprietary applications
• Binary analysis and reverse engineering
• Smart contract auditing for DeFi targets

6. Domain Specialists

Active Directory Specialist

• Kerberoasting and AS-REP roasting
• DCSync and DCShadow attacks
• Delegation abuse (constrained, unconstrained, RBCD)
• Certificate abuse (ESC1-ESC8)
• GPO exploitation and ACL abuse

Cloud Specialist

• AWS/GCP/Azure misconfiguration exploitation
• IAM privilege escalation paths
• Metadata service abuse
• Storage bucket enumeration and access
• Serverless function exploitation

Smart Contract Specialist

• Solidity vulnerability detection
• Flash loan attack simulation
• Reentrancy, overflow, access control flaws
• DeFi protocol attack path analysis
• Token economics manipulation

Reverse Engineering Specialist

• Binary analysis and vulnerability identification
• Malware analysis and capability extraction
• Firmware extraction and analysis
• Protocol reverse engineering
• Obfuscation and packing defeat

Analyst Specialist

• Finding correlation and impact assessment
• MITRE ATT&CK technique mapping
• Risk scoring and prioritization
• Detection gap analysis (what would blue team catch?)
• Executive summary generation

Engagement Packages

This section covers engagement packages for decepticon red team.

• Ensure all prerequisites are met before proceeding
• Follow the documented workflow steps in sequence
• Record results and any anomalies encountered during this phase

Rules of Engagement (RoE)

Structured boundaries for the operation:

• Authorized targets and IP ranges
• Prohibited actions and data handling rules
• Communication channels and escalation paths
• Working hours and emergency contacts
• Legal authorization references

Concept of Operations (ConOps)

High-level operational approach:

• Attack strategy and phases
• Resource allocation across specialists
• Timeline and milestones
• Success criteria and abort conditions
• Coordination with blue team (if purple team)

Operational Plan (OPPLAN)

Detailed execution plan with MITRE ATT&CK mapping:

Installation

# Install required dependencies
sudo apt-get update && sudo apt-get install -y <tool-name>

Docker

git clone https://github.com/PurpleAILAB/Decepticon.git
cd Decepticon
docker build -t decepticon .
docker run -it decepticon

Pip

git clone https://github.com/PurpleAILAB/Decepticon.git
cd Decepticon
pip install -r requirements.txt
python main.py

Workflow Example

1. Define engagement: scope, rules, objectives
2. Generate RoE + ConOps + OPPLAN packages
3. Run reconnaissance phase (automated)
4. Review recon findings, adjust OPPLAN
5. Execute exploitation phase (specialist-driven)
6. Post-exploitation and lateral movement
7. Analyst correlates findings with ATT&CK
8. Generate final report with detection gaps

When NOT to Use

• Task is outside your authorization scope
• You need to implement controls (use implementing-* skills)
• Task is about analysis, not action (use analyzing-* skills)
• You don't have access to target systems
• Task requires compliance expertise (consult professionals)
• Task is about defense, not offense (use defensive skills)

Red Flags

• Autonomous execution requires strict RoE enforcement
• Always verify scope boundaries before exploitation phases
• Smart contract testing must use testnets unless authorized
• Real credentials and data handling must follow engagement rules
• Ensure kill switch / abort mechanism is available at all phases

Verification

• All steps executed successfully against a test environment before production use
• Output documented with screenshots or logs demonstrating expected behavior
• All exploited vulnerabilities documented with reproduction steps
• Scope boundaries confirmed — only authorized targets were tested
• Remediation recommendations included for every finding

Process

1. Analyze the task requirements
2. Apply domain expertise
3. Verify output quality

Anti-Rationalization Table