{"ID":6138279,"CreatedAt":"2026-07-09T01:07:32.349475501Z","UpdatedAt":"2026-07-11T14:03:15.775363119Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.07436","arxiv_id":"2607.07436","title":"The Blind Curator: How a Biased Judge Silently Disables Skill Retirement in Self-Evolving Agents","abstract":"A self-evolving agent retires its bad skills by watching them fail, so what happens when the judge cannot see the failures? Skill retirement is the structural constraint that keeps a growing library from drifting below the no-skill baseline, but its guarantee assumes an unbiased reward, which is false for the LLM judges that reference-free tasks force upon us. We show that a biased judge does not merely add noise; it \\emph{silently switches off the curator}. We make this precise with a corrupted-reward analysis and, isolating the causal channel by injecting corruption on top of a deterministic reward, a behavioral study on a reference-free report-writing testbed with a code-generation cross-check. Symmetric noise leaves retirement intact, but \\emph{false-pass} bias (failures slipping through as passes) disables contribution-based retirement past a sharp threshold that no amount of data can cross. Separating genuine retirement from cap-eviction churn shows this \\emph{mechanism} failure is universal, holding across domains and failure rates and sparing only near-zero-false-pass, verifier-like graders. The downstream \\emph{outcome}, though, is regime-dependent: eval quality degrades only where the same corruption also starves skill synthesis, and otherwise holds steady, so the disabled curator is \\emph{silent}, surfacing in no aggregate metric. The contribution is a behavioral safety result, not a performance one. A cheap defect-injection audit then tells an operator, before deployment, which side of the threshold their judge occupies.","short_abstract":"A self-evolving agent retires its bad skills by watching them fail, so what happens when the judge cannot see the failures? Skill retirement is the structural constraint that keeps a growing library from drifting below the no-skill baseline, but its guarantee assumes an unbiased reward, which is false for the LLM judge...","url_abs":"https://arxiv.org/abs/2607.07436","url_pdf":"https://arxiv.org/pdf/2607.07436v1","authors":"[\"Xing Zhang\",\"Yanwei Cui\",\"Guanghui Wang\",\"Ziyuan Li\",\"Wei Qiu\",\"Bing Zhu\",\"Peiyang He\"]","published":"2026-07-08T14:08:04Z","proceeding":"cs.AI","tasks":"[\"cs.AI\",\"cs.CL\",\"cs.CR\"]","methods":"[\"Large Language Model\"]","has_code":false}