Skip to main content
Back to Projects
DEDSEC-AI2024 - Present

DedSec.ai

Scope-controlled automated adversary emulation and attack platform

#The Problem

Organizations struggle to continuously validate their security posture against evolving threats. Traditional penetration testing is expensive, point-in-time, and often fails to cover the full attack surface. Security teams need a way to safely and repeatedly test defenses without the risk of unauthorized actions.

#Key Constraints

01All operations must be scope-controlled and authorized
02Full audit trail and forensic logging required
03Must integrate with existing security infrastructure
04Repeatability and consistency across engagements

#The Solution

Built an AI-driven adversary emulation platform that uses autonomous agents to simulate realistic attack chains within strictly defined boundaries. The system employs a multi-agent architecture where a planning agent coordinates with specialized execution agents, all operating under a governance layer that enforces scope constraints. Each action is validated against the authorized scope before execution.

#Impact

Reduced time-to-validation by 70% compared to manual assessments
Enabled continuous security validation with weekly automated runs
Achieved 100% scope compliance across all engagements
During security competitions, acquired 60% of the total points

Metadata

ROLEFounder & Developer
TAGS
Red TeamAI AgentsSecurity

Visualization

LAT: 34.0522 N
LON: 118.2437 W
LIVE FEED
REC
SIGNAL STRENGTH
TARGET ACQUIRED

Technology Stack

PythonTypeScriptLangChainDockerKubernetesPostgreSQL