EvoUndo：受可恢复性约束的LLM智能体框架自进化

摘要

LLM智能体越来越频繁地在运行时修改自身的提示词、工具、中间件、资源和执行框架。这种自我进化可以提升能力，但一次成功的变异可能留下持久影响，而这些影响在与其产生时不同的状态下无法被安全逆转。我们提出EvoUndo——一个用于跨反事实状态表示、综合、诊断和独立验证模型生成的自我修改之可恢复性的框架。在600个未见过的单次自我进化任务中，我们识别出197个未通过可恢复性验证的能力提升变异。在原始恢复表示下，传统修复策略未能恢复这些自然失败中的任何一个（0/197）。确定性预言机分析在原始恢复语言L0下恢复了48/197，而扩展恢复演算将经验预言机恢复提升至191/197。随后，一种协议锁定的2×2接地-表达力干预分离出两个瓶颈：在原始语言足够的条件下，精确状态地址接地将成功恢复率从0/48提升至38/48（79.2%），而扩展恢复语言使得预言机定义的S1层中142/143（99.3%）的失败得以恢复。在主要gpt-oss-120b主干模型上，向更丰富的语言添加精确地址诊断将恢复率降至133/143（93.0%）；Qwen3.8-27B的重复实验重现了接地与表达力的效应，但未重现这一负面交互，表明后者具有模型依赖性。这些结果表明，可靠的智能体自我进化需要协同设计验证、状态接地、见证语义与恢复语言表达力，而非仅依赖迭代式提示。

English

LLM agents increasingly modify their own prompts, tools, middleware, resources, and execution harnesses at runtime. Such self-evolution can improve capability, but a successful mutation may leave persistent effects that cannot be safely reversed in states different from the one in which it was created. We introduce EvoUndo, a framework for representing, synthesizing, diagnosing, and independently verifying recoverability of model-generated self-modifications across counterfactual states. Across 600 unseen one-shot self-evolution tasks, we identify 197 capability-improving mutations that fail recoverability verification. Under the original recovery representation, conventional repair strategies recover 0/197 of these natural failures. Deterministic oracle analysis recovers 48/197 under the original recovery language L0, while the extended recovery calculus increases empirical oracle recovery to 191/197. A protocol-locked 2x2 grounding-by-expressivity intervention then separates two bottlenecks: exact state-address grounding increases successful recovery from 0/48 to 38/48 (79.2%) when the original language is sufficient, while extending the recovery language enables recovery on 142/143 (99.3%) failures in the oracle-defined S1 stratum. On the primary gpt-oss-120b backbone, adding exact-address diagnostics to the richer language reduces recovery to 133/143 (93.0%); a Qwen3.8-27B replication preserves the grounding and expressivity effects but not this negative interaction, indicating that the latter is model-dependent. These results indicate that reliable agent self-evolution requires co-designing verification, state grounding, witness semantics, and recovery-language expressivity rather than relying on iterative prompting alone.