OSS agent harness auto-researches a C++ scheduler that beats expert solver on 29/30 problems
hyperparticle · x · 2026-10-21
Developer hyperparticle shows results from an open-source autoresearch harness: the agent accumulates knowledge even when idle, auto-researched a better C++ factory scheduler, and beat an expert-designed solver on 29/30 fresh problems — with the final solver running without any LLM.
In one case the agent's solver found a legal plan the reference never considered, cutting required trucks from 30 to 10. Two OSS tools ship with it: sagent (a self-mutating, hot-swapping, multi-provider coding-agent CLI and typed Python library with async tool calls and recursive spawn, 81 stars) and knowledge-graph tool trackinizer.
Related event: Open-Source Autoresearch Agent Beats Expert Solvers on 29 of 30 Problems(3 posts)→
More from coding & agent
- Decisions API hits GA as agents automate cloud video editing via feedback loops — dkundel · 2026-10-23
- A curated library of 50+ free AI agent resources: videos, repos, one Google Doc — ifioknkem · 2026-10-23
- The 'best free AI agent resource library on the internet,' in one Google Doc — ifioknkem · 2026-10-23
- Vercel adds no-account deploys: drag and drop a file to get a live URL — steren · 2026-10-23
- Dev runs 5 Claude Code sessions that coordinate via MCP Agent Mail to root his network cameras — doodlestein · 2026-10-23
- Vercel ships 'vercel deploy --temporary' for ephemeral agent-built deployments — cramforce · 2026-10-23