Discover and explore top open-source AI tools and projects—updated daily.
yv1ingMulti-agent workbench for authorized security assessment and research
Top 73.2% on SourcePulse
A controlled multi-agent workbench designed for authorized security assessments, code auditing, internal reviews, and research. It empowers security professionals and researchers by orchestrating specialized agents within a governed workflow, ensuring traceable and secure operations. The system facilitates planning, evidence collection, validation, and reporting through clear role boundaries and sandboxed execution.
How It Works
Z3r0 employs a layered architecture with a React frontend, FastAPI backend, and a robust agent runtime orchestration. It coordinates a Chief Security Officer (CSO) agent with domain specialists (audit, intelligence, penetration, reverse engineering, crypto) to decompose and execute tasks. Core to its design are Docker-backed sandboxes for controlled execution of commands and tools, interrupt-driven task handling for atomicity, and a durable timeline event log for session replay and auditing. This approach ensures clear separation of concerns, traceable execution, and secure, isolated environments for sensitive operations.
Quick Start & Requirements
.z3r0/config.json.example to .z3r0/config.json and configure settings, then run docker compose -f docker-compose.prod.yml up -d --build. Access the workbench at http://127.0.0.1:8000.QUICKSTART.md (mentioned for setup).Highlighted Details
Maintenance & Community
The project acknowledges support from the Linux.do website and its community. No specific details on active maintainers, sponsorships, or dedicated community channels (like Discord/Slack) are provided in the README.
Licensing & Compatibility
This project is licensed under the MIT License, which is generally permissive for commercial use and integration into closed-source projects.
Limitations & Caveats
Z3r0 is strictly intended for authorized security assessment, code auditing, internal review, and controlled research or training environments. It explicitly prohibits unauthorized or unlawful use, including testing third-party systems without explicit permission. Users are solely responsible for obtaining authorization, defining scope, and complying with all applicable laws and contracts. High-privilege assets like the Docker socket and model credentials necessitate trusted, isolated environments.
1 day ago
Inactive