Preprint claims self-evolving skill graphs lift agent retrieval

A preprint, SE-GoS, claims that maintaining a retrieval graph from execution traces alone improves skill-library retrieval for LLM agents.

The paper by Dawei Fu and four coauthors was updated on arXiv on October 1. The method treats the skill retrieval graph as an index, adjusting its connections and node descriptions automatically from execution records — no model training, no retrieval-algorithm changes, and no separate model judging which skills relate.

The authors' own tests show one evolution round lifting average reward on SkillsBench from 52.4% to 59.4%, above full-library loading and vector retrieval baselines; on a held-out split it never saw, from 52.9% to 58.3%, at about two-thirds the input tokens of loading the full library. Repeating the round adds nothing; the abstract does not explain why.