Securing agentic and AI applications —
together.
An open, collaborative community dedicated to securing agentic enterprise applications and accelerating the responsible use of AI in security operations.
Open source
Flagship Projects
Reference implementations and shared infrastructure, built in the open.
Agent behavior verifier. Compares an AI agent's declared policy against the available evidence, and reports where observed behavior diverges from declared intent.
Visit Praxen → Available now
An open-source observability and telemetry framework designed for AI and agentic systems, providing deep visibility into agent behavior, execution flows, runtime events, and inter-agent communication.
Visit Observra →Beyond the flagships: Promptfall in production, and new work incubating on GitHub — see how we classify project maturity →
Get involved
Come build with us
Two simple ways in: follow along and join the conversation on LinkedIn, or jump straight into the code on GitHub.
News & discussion
Read the blog for release announcements and updates, and join the community on LinkedIn for the conversation on securing agentic AI.
Code & issues
Contribute pull requests and file issues on our open-source projects — from flagship to incubator — right on GitHub.
Why we exist
Mission & vision
A neutral place to build the shared tools, patterns, and guidance that make agentic AI safe to adopt.
Mission
To build an open, collaborative community dedicated to securing agentic enterprise applications and accelerating the responsible use of AI in security operations.
The community brings together security practitioners, AI builders, researchers, engineers, and defenders to create shared tools, patterns, integrations, and guidance that help organizations safely develop, deploy, monitor, and govern AI agents and extensions in real-world enterprise environments.
Our goal is to make it easier for security and SOC teams to adopt agentic technologies with confidence by advancing practical, open, interoperable, and community-driven approaches to AI security.
Vision
We envision a future where agentic AI systems are secure by design, transparent in operation, and trusted by the people and organizations that depend on them.
As enterprises increasingly adopt AI agents to automate workflows, augment analysts, and extend security platforms, the community will serve as a neutral place to define best practices, share reusable components, identify emerging risks, and build technologies that help defenders move faster without compromising safety, control, or accountability.
Our purpose
This community exists to
Concrete, practical work the community takes on together.
- Help organizations securely build and operate AI agents and enterprise AI extensions.
- Provide open source tools, reference implementations, and reusable components for security and SOC use cases.
- Create practical guidance for agent security, governance, permissions, identity, observability, testing, and incident response.
- Encourage collaboration between security practitioners, developers, researchers, and enterprise teams.
- Promote interoperability across security tools, AI platforms, models, data sources, and automation frameworks.
- Reduce duplication of effort by enabling the community to share patterns, connectors, evaluation methods, and lessons learned.