Safety
// Skill Misevolution in Self-Improving LLM Agents // Self-improving agents write their successes down as reusable skills. An unsafe success…
// Skill Misevolution in Self-Improving LLM Agents // Self-improving agents write their successes down as reusable skills. An unsafe success becomes reusable policy long after the input that triggered
// Skill Misevolution in Self-Improving LLM Agents // Self-improving agents write their successes down as reusable skills. An unsafe success becomes reusable policy long after the input that triggered it is gone. SkillMisevo-Gym versions skill state across agent frameworks so risk can be attributed separately to authoring, retrieval, and later execution. Across 25 agent-method configurations covering 525 tasks each, all 21 evolved configurations author unsafe artifacts. Only 15 produce harm in a fresh session, which means authoring risk and execution risk are separate problems. Three malicious tasks raise carryover attack success from 16.0% to 35.3%. Their SafeEvolve wrapper cuts unsafe retrieval by 26.7 points and fresh-session harm by 17.3 points while benign utility moves 0.4 points. Paper: https://arxiv.org/abs/2608.12851 Track more trending AI papers in our academy: https://academy.dair.ai/
Related
- Practice Makes Unsafe: Skill Misevolution in Self-Improving LLM Agents
- AlignEvoSkill: Towards Knowledge-Aware and Task-Aligned Agent Skill Evolution
- Who Grades the Grader? Co-Evolving Evaluation Metrics and Skills for Self-Improving LLM Agents
Source: DAIR.AI (X) | 2026-08-14