# EvoSkill v2 preprint reports gains from agents rewriting task instructions

_Published Monday, October 26, 2026 at 10:07 AM EDT · AI, Science · Latest · Tier 2 — Notable_

An EvoSkill v2 preprint reports that agents improved held-out task performance by rewriting their instructions while leaving the underlying model and harness unchanged. With fine-grained ground truth, the system exceeded the stronger of SkillOpt and SkillBoost by 7.7 percentage points on SpreadsheetBench v2 and 5.4 percentage points on a banking benchmark.

A persistent coding agent reviews training attempts and revises instructions each epoch. Across both benchmarks and feedback settings, its optimizer cost less per percentage point of held-out improvement than either baseline. The results concern benchmark tasks graded with supplied feedback.

## Sources

- [cs.AI updates on arXiv.org](https://arxiv.org/abs/2610.25931)

---
Canonical: https://techandbusiness.org/newswire/bHgumqtYWzIH5Jg2xk4-Wv
Published: 2026-10-26T14:07:30.315Z
Story chronology: 2026-10-26T04:00:00.000Z
Retrieved: 2026-10-26T17:12:18.753Z
Publisher: Tech & Business (techandbusiness.org)