Google open-sources RRSI, a framework that auto-evolves LLM agent harnesses
rohanpaul_ai · x · 2026-10-01
Google Research released RRSI (Regularized Recursive Self-Improvement of Agent Harnesses), an open-source framework plus paper that automatically evolves an LLM agent's harness — prompts, control flow, tools, memory and context management — around a frozen model.
Evolving a harness against a fixed evolve set overfits: in-distribution gains shrink or vanish out of distribution. RRSI keeps the edit space open and regularizes the search trajectory:
- Proposal side: an annealed budget caps bundled edits per candidate; the proposer is conditioned on full edit history so falsified hypotheses aren't redrawn; stalled runs are redirected toward unexercised components.
- Selection side: a critic screens candidates for suite-specific logic before evaluation, with noise-adaptive selection.
The repo is on GitHub with 1.1k stars.
More from coding & agent
- Dev warns: less code review is coming, but don't follow the loud AI cargo-culters — mgill25 · 2026-10-01
- 39 Agents Across 8 Departments: The 2026 Data Science Team Is a Folder — mdancho84 · 2026-10-01
- Dev builds his own AI-powered Substance Designer for a city sim game — D3VAUX · 2026-10-01
- Ethan Mollick: I underestimated AI's ability to self-organize, agents beat elaborate orchestration — emollick · 2026-10-01
- RegLLM: a diagnostic harness measures bounded autonomy in regulated agentic AI — Dipankar Sarkar · 2026-10-01
- Jelly launches as open-source local-first agent workspace for running your whole business on one VM — Scobleizer · 2026-10-01