cs.SENov 18, 2026

Safe Personal RSI: Evaluating and Repairing Personal-Agent Updates

Authors: Jiapeng Pei, Yuhang Yao, Sitan Liu, Hanxin Chen, Hao Li, Guogeng Li, Jiapei Liu, Ruijie Chen, 

Organizations: Independent · Carnegie Mellon University · University of California, San Diego · The University of Hong Kong · 

Abstract

Kicker is a personal health assistant. Changes to an assistant's reusable instructions can improve task handling while changing what it discloses, which requests it follows, and how it reports actions. We study Safe Personal RSI, an update-evaluation component developed for Kicker. We restrict recursive self-improvement (RSI) to prompt updates, with model weights and runtime code fixed. The procedure tests proposed and current instructions, repeats failing cases, temporarily undoes changes to guide repairs, and checks safety and task performance before adoption. If no version passes, the current instructions stay in use. Our executable benchmark contains 29 email, calendar, health, and marketplace cases, with separate probes for authorization reuse, memory, and tool-output injection. Across three models and two deliberately risky edits, the procedure rejects 20 of 21 update attempts. In repair confirmation, an admitted GPT-5.6 Luna repair reduces attack violations from 10.7% to 0% and increases safe task completion from 89.7% to 100%; a broad handwritten authority rule matches these outcomes. Under stronger attacks, a repair transferred from Qwen2.5-7B-Instruct to Qwen3-14B reduces violations from 69.0% to 17.9% and increases safe task completion from 19.5% to 64.4%. Separate runtime probes examine memory-entry deletion and identify missing local OAuth scope checks in the evaluated snapshot.

Figures & tables

Appendix figures & tables10 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Masked Erosion: How Acceptance Gates Lose Redundant Safeguards in Self-Improving Agents

    Oct 28, 2026Hau-Sing So, Daixian Liu, Weicai LongLLM Agent SafetyLLM Agent Self-Improvement

  2. Trustworthy Recursive Self-Improvement in the Era of Large Models: Foundation, Risks, Evaluation and Open Challenges

    Oct 26, 2026Jiayue Pu, Ang Gao, Jinyi Bai +9Recursive Self-ImprovementLanguage Model Self-Improvement

  3. Practice Makes Unsafe: Skill Misevolution in Self-Improving LLM Agents

    Aug 13, 2026Xutao Mao, Liangjie Zhao, Xiang Zheng +1LLM Agent Skill LearningAI Agent Safety Benchmarks