Harness-of-Harness lets coding agents run unsupervised for days, 52% avg gains
dair_ai · x · 2026-09-02
New research on Harness-of-Harness wraps whatever coding harness you already run and organizes executions into repeated planning-coding-testing increments, enabling agents to keep building for days without a human.
Key mechanics:
- Balances repair against capability growth
- Scopes work into small verifiable steps
- Keeps implementation-time testing separate from independent evaluation
- Constrains outputs rather than the workflow
Across GameCraft-Bench, FrontierSWE and ProgramBench with three harness-model pairs, it averages a 52.25% relative gain over standalone harnesses after three iterations, peaking at 82.86%.
Related event: Harness-of-Harness lets coding agents run unattended for days(3 posts)→
More from coding & agent
- LLM one-shots a Minecraft 3x3 piston door by building its own Linux test infra — Vjeux · 2026-10-18
- Dev builds Mumbai-inspired 3D manga background generator from OpenStreetMap data — CurieuxExplorer · 2026-10-18
- Solo founder builds a money-printing data service on Claude's Pi Durable harness — MicahBerkley · 2026-10-18
- Dev builds Arch-based OS with custom desktop and browser using Claude Code — Snoo-10493 · 2026-10-18
- Five SaaS agent workflows fail the same way — and it's never model intelligence — usehive · 2026-10-18
- Claude skill generates hand-drawn interactive diagrams that teach kids how things work — CatAstro_Piyush · 2026-10-18