Show HN: Self-improving sandboxed agent with memory and scheduling

Category: ai-ml

Tags: autonomous-agent, self-improving, sandboxed, browser-automation, rag

Score: 7.0/10 (Innovation: 7, Technical: 7, Documentation: 8, Utility: 6)

GrimmBot is an autonomous AI agent that runs in a Docker sandbox with desktop/browser control, capable of learning from errors, creating custom tools, and using persistent memory. It's interesting because it combines several advanced agentic concepts—self-improvement via adaptation rules, on-the-fly tool creation, and RAG memory—into a single, sandboxed system with a live web interface.

Target audience: ai researchers, developers experimenting with autonomous agents, devops engineers

Repository: https://github.com/Grimm67123/GrimmBot · Python · AGPL-3.0 · 2 stars

View on Hacker News