Garry Tan, President and CEO of Y Combinator, has announced the release of GBrain v0.10.0, a significant update to his open-source personal AI agent knowledge system. The new version introduces 24 "fat skills" with robust testing, refined core configuration files, and crucial Access Control Lists (ACLs) for multi-user brain access.
"GBrain v0.10.0 is a big one," Tan stated in a recent tweet. "My personal OpenClaw setup and brain can now be yours. I've perfected my RESOLVER.md, my SOUL.md and ACLs for multi-user brain access. Now there are 24 distinct fat skills with fat code, fully tested with e2e tests, evals and unit tests."
GBrain functions as a persistent, searchable long-term memory for AI agents, primarily designed for OpenClaw and Hermes Agents. It stores knowledge in Markdown files, enabling agents to have "perfect total recall" of vast amounts of information. The system aims to transform an agent into a stateful entity by providing a local-first, entity-aware knowledge base.
The SOUL.md file defines an agent's personality, values, tone, and behavioral boundaries, acting as its "character sheet." Meanwhile, RESOLVER.md files, located within each knowledge domain directory, guide the agent on how to categorize and disambiguate new information, ensuring a structured and consistent knowledge base. The introduction of ACLs in v0.10.0 is expected to enhance collaboration and security for agents operating within shared knowledge environments.
The "24 distinct fat skills with fat code" represent specialized functionalities, rigorously validated through end-to-end (e2e) tests, evaluations (evals), and unit tests. This comprehensive testing suite underscores a commitment to reliability and performance, ensuring these skills operate effectively within the GBrain ecosystem. These skills likely cover a range of tasks, from information retrieval and summarization to complex decision-making processes.
This update reinforces GBrain's role as a foundational tool for building sophisticated personal AI agents. By open-sourcing his own production setup, Tan continues to provide developers with the infrastructure to create advanced AI agents capable of managing extensive personal and professional data.
