Project
openmind
Self-Improving Cognitive AI System
TypeScriptEffect-TSNext.jsWSL2CDPDrizzleBun
openmind is a self-improving cognitive system that learns across sessions, audits its own code, and gets better every cycle — built from scratch by a 14-year-old developer.
At its core lies a 19-service cognitive infrastructure spanning associative memory (TF-IDF + fuzzy + n-gram), a private Linux kernel execution environment via WSL2, autonomous browser control through Chrome DevTools Protocol, and a multi-agent security orchestration pipeline.
The system continuously evolves through 7 autonomous self-improvement subsystems: failure datalake, auto-learn (generates Semgrep rules from fixes), harden-on-write, parallel reasoning, fuzzing, codebase graphing, and sandboxed change validation.
Key Features
19 cognitive infrastructure services
Private Linux kernel execution (WSL2)
Autonomous browser control via Chrome CDP
Multi-agent security orchestration
Associative memory with TF-IDF + fuzzy + n-gram
Self-audit and self-healing pipeline
7 autonomous self-improvement systems