OpenHarness: open-source workbench for orchestrating coding agents beyond code
dee_hw · x · 2026-09-20
deehw shares why he keeps coming back to OpenHarness — describing it as "tmux/vim/fzf/zsh for managing my fleet of AI agents," and notably for non-coding uses, like driving Blender to design an aero bike helmet.
OpenHarness (autonomous-ai/openharness, 495 stars, 1200+ commits) is an open-source software and hardware project for building with coding agents like Claude Code and Codex: software, parts and circuit board design, data exploration, and video creation — with agents, domain tools, and live viewers in one place, and sessions persisting across every machine you own. Its pitch: coding agents absorb the "ten thousand hours" tax, so polymaths can build across disciplines.
More from coding & agent
- Claude Skill turns an image's buildings into resizable Three.js models — jasonkneen · 2026-09-20
- Cursor Pro users report credits draining fast and degraded Auto mode, weighing $25/month alternatives — ReporterUnusual9295 · 2026-09-20
- Fuzzing vs property-based testing: the terms are splitting — exploration vs assertion — blaizedsouza · 2026-09-20
- Software factory take: LLMs may generate, but never specify or verify — blaizedsouza · 2026-09-20
- Jev as an NSFW prompt filter: 93% on CSAM evals, sub-cent cost, and where thresholds bite — Murky_Ad8671 · 2026-09-20
- "A heuristic cannot certify a heuristic": why software-factory orchestration thinking is wrong — blaizedsouza · 2026-09-20