vxcontrol/pentagi
Pentagi
Pentagi is an autonomous security testing framework and agent orchestrator designed to plan and execute end-to-end security assessments. It utilizes a coordination engine to decompose complex goals into actionable subtasks, performing automated penetration testing and vulnerability research within isolated container environments.
The system distinguishes itself through a temporal knowledge graph that tracks semantic relationships between entities and vulnerabilities to reuse intelligence across projects. It includes a web intelligence reconnaissance tool for automated data gathering and agentic loop monitoring to detect inefficient tool usage patterns and trigger corrective guidance.
The platform provides capabilities for human-in-the-loop steering to redirect active investigations in real-time, alongside provider-agnostic integration for various artificial intelligence models. It further supports session-scoped file management and the generation of detailed vulnerability reports and exploitation guides.
Access to programmatic workflows is secured via token-based authentication and external identity providers using OAuth.
Features
- Agent Orchestrators - Provides a coordination engine that decomposes complex security goals into actionable subtasks and manages agent workflows.
- AI Agent Orchestrators - Coordinates specialized AI agents using structured workflows to execute end-to-end security assessments.
- Human-in-the-Loop Steering - Allows real-time redirection and priority updates to active autonomous investigations without restarting the process.
- Knowledge Graphs - Implements a temporal knowledge graph to track semantic relationships between entities and vulnerabilities for AI-driven analysis.
- Task Decomposition Systems - Employs prompt-driven logic to decompose high-level security goals into a sequence of actionable subtasks.
- Temporal Knowledge Graphs - Builds a temporal knowledge graph of entities and vulnerabilities to reuse intelligence across projects.
- Temporal - Implements a temporal graph database to track semantic relationships between entities and vulnerabilities over time.
- Sandboxed Execution Environments - Provides sandboxed computing environments for safely running security tools and performing file operations.
- Isolated Execution Sandboxes - Provides secure, resource-constrained environments for running autonomous security tools and tests.
- Container-Based Sandboxes - Uses ephemeral container environments to isolate security testing operations and prevent host system compromise.
- Execution Loop Monitors - Implements a monitoring system to detect repetitive tool usage patterns and steer autonomous agent behavior.
- Autonomous Penetration Testing - Ships an autonomous framework that plans and executes end-to-end security assessments and penetration tests.
- Agent Monitoring - Monitors AI agent behavior to detect inefficient tool usage patterns and trigger corrective guidance.
- AI Provider Integrations - Provides a configuration layer for integrating various external AI providers for reasoning and embedding tasks.
- Human-in-the-Loop Workflows - Allows real-time human intervention to redirect active autonomous investigation flows and update priorities.
- LLM Provider Integrations - Offers configuration and authentication adapters for connecting to external large language model services.
- Web Search Tools - Integrates search APIs and browser capabilities to allow autonomous agents to gather real-time web intelligence.
- Automated Exploitation Guides - Generates detailed vulnerability reports and exploitation guides based on autonomous assessment findings.
- Vulnerability Research - Implements systematic processes for gathering web intelligence and investigating previously unknown security weaknesses.
- Reconnaissance Tools - Includes tools for automated web reconnaissance to gather intelligence and map targets for security workflows.
- AI Red Teaming - Autonomous AI agent system for complex penetration testing tasks.
