[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-81787-en":3,"doc-seo-81787-105":30,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":13,"seo_description":14,"update_tm":28,"read_time":29},81787,549758252649,"Ivy","https://ap-avatar.wpscdn.com/avatar/8000253669c5317157?_k=1778319167496531819",8,"Research & Report","Self-Evolving Agents with Anytime-Valid Certificates","Self-evolving agents violate the premises of many learning-theoretic guarantees because the data, evaluator, components, and hypothesis space are generated by the policy being updated. The SEA architecture confines self-modification to a small steering adapter and a versioned harness around a frozen base model. Every modification passes an anytime-valid gate that issues auditable certificates under a fixed error budget. Five loop controllers and verifier-in-the-loop mechanisms provide grader-free dense signals from issue text, and experiments on a SWE-bench verified subset show dominant base capability and controlled suite contributions, with event logs tracking regressions prevention.","arXiv :2607 .0087 1v 1 [ cs .AI] 1 Jul 2026  \nSelf-Evolving Agents with Anytime-Valid Certificates  \nBiswa Sengupta  \nLLM Suite Team, JPMorgan Chase & Co.  \n[biswa.sengupta@jpmorgan.com](biswa.sengupta@jpmorgan.com)  \nAbstract  \nSelf-evolving agents violate the assumption behind most learning-theoretic guarantees: the data, evaluator, components, and hypothesis space are produced by the policy being updated. We present SEA, an architecture that confines selfmodification to a small steering adapter and a versioned harness around a frozen base model and admits each modification only through an anytime-valid gate that emits an auditable certificate against a fixed error budget. Five loop controllers compose published guarantees; because such gates can only select among behaviors the frozen base already produces, five verifier-in-the-loop mechanisms—best-of-N , micro-step search, self-authored reproduction oracles, search-layer control, and self-repair—supply the dense, grader-free signal the gates require, computed from the issue text alone. On a 52-instance SWE-bench Verified subset across four base models, base capability is the dominant, confound-free effect, and on two strong base models a deliberate no-op-composite control isolates the suite’s contribution at +4 and +5 (GLM 5.2 24 → 28 ; GPT 29 → 34, the 65% best), with event logs confirming that its mechanisms fire and prevent regressions. Results are single-run on expensive evaluations; confirming run-to-run variance and adapting the per-task algorithm mix are future work.  \nDisclaimer: This paper was prepared for informational purposes by the LLM Suite group of JP Morgan Chase and its affiliates (‘JPMC’) and is not a product of the Research Department of JP Morgan. JPMorgan makes no representation, warranty or undertaking whatsoever and disclaims all liability for the completeness, accuracy or reliability of the information contained herein. This document is not intended as investment research or investment advice, or a recommendation, offer or solicitation for the purchase or sale of any security, financial instrument, financial product or service, or to be used in anyway for evaluating the merits of participating in any transaction, and shall not constitute a solicitation under any jurisdiction or to any person, if such solicitation under such jurisdiction or to such person would be unlawful.  \n1 Introduction  \nA self-evolving agent improves its own future behavior using data, evaluations, components, and a hypothesis space that it itself produces—rewriting prompts and tools, distilling its outputs, learning its reward models, growing skill libraries. The guarantees one would invoke for such systems, however, were proven for exogenous environments: continual-learning forgetting bounds [Farajtabaret al., 2020, Chugg et al., 2023], convergence of preference optimization [Tiapkin et al., 2025, Wang et al., 2025], unbiasedness of policy-gradient estimators [Meulemans et al., 2023], safe policy improvement [Thomas et al., 2015], and library-learning optimality [Bowers et al., 2023] each assume a task stream, evaluator, MDP, or program library fixed independently of the learner. We call the violation the endogenous-loop failure mode: the evolving policy generates the data it trains on, the evaluator it is judged by, the components it is built from, and the hypothesis space it searches. The name is a shorthand, not a precise mathematical category, and taken literally it overstates the problem: violating a theorem’s hypotheses voids its certificate but does not negate its conclusion. The  \nPreprint. Under review.  \n2026 JP Morgan Chase & Co. All rights reserved  \nguarantee simply ceases to be certified—the bound may still hold, may degrade gracefully, or may break, depending on the problem—and performative-prediction theory shows the loop can in fact still contract when the policy-induced distribution shift is small enough [Perdomo et al., 2020] . We use the term to mark where ","cbCaioivW28DAFE8","https://ap.wps.com/l/cbCaioivW28DAFE8","pdf",618467,4,1,30,"English","en",105,"# Abstract\n# Introduction\n## Endogenous-loop failure mode\n## Architecture principles and anytime-valid gates\n# Contributions\n## Reference architecture\n## Loop controllers","[{\"question\":\"What problem does SEA target in self-evolving agents?\",\"answer\":\"SEA targets the endogenous-loop failure mode, where the evolving policy produces the data, evaluator, components, and hypothesis space it trains and is judged on, which can invalidate classical learning-theoretic assumptions.\"},{\"question\":\"How does SEA keep guarantees valid during self-modification?\",\"answer\":\"SEA routes every self-modification through an anytime-valid gate that emits an auditable certificate against a fixed error budget, ensuring each modification preserves the intended guarantee under closed-loop operation.\"},{\"question\":\"What role do the verifier-in-the-loop mechanisms play?\",\"answer\":\"They supply dense, grader-free signals computed from the issue text alone, allowing five loop controllers to realize the published guarantees in practice when gating can only select behaviors the frozen base already produces.\"}]",1784176142,76,{"code":4,"msg":31,"data":32},"ok",{"site_id":25,"language":24,"slug":33,"title":13,"keywords":34,"description":14,"schema_data":35,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":28},"self-evolving-agents-with-anytime-valid-certificates","",{"@graph":36,"@context":85},[37,53,68],{"@type":38,"itemListElement":39},"BreadcrumbList",[40,44,48,51],{"item":41,"name":42,"@type":43,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":45,"name":46,"@type":43,"position":47},"https://docshare.wps.com/document/","Document",2,{"item":49,"name":12,"@type":43,"position":50},"https://docshare.wps.com/document/research-report/",3,{"item":52,"name":13,"@type":43,"position":20},"https://docshare.wps.com/document/self-evolving-agents-with-anytime-valid-certificates/81787/",{"url":52,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":24,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":41,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-26","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What problem does SEA target in self-evolving agents?","Question",{"text":75,"@type":76},"SEA targets the endogenous-loop failure mode, where the evolving policy produces the data, evaluator, components, and hypothesis space it trains and is judged on, which can invalidate classical learning-theoretic assumptions.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does SEA keep guarantees valid during self-modification?",{"text":80,"@type":76},"SEA routes every self-modification through an anytime-valid gate that emits an auditable certificate against a fixed error budget, ensuring each modification preserves the intended guarantee under closed-loop operation.",{"name":82,"@type":73,"acceptedAnswer":83},"What role do the verifier-in-the-loop mechanisms play?",{"text":84,"@type":76},"They supply dense, grader-free signals computed from the issue text alone, allowing five loop controllers to realize the published guarantees in practice when gating can only select behaviors the frozen base already produces.","https://schema.org",{"og:url":52,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,122,127,130,134],{"id":21,"doc_module":4,"doc_module_name":46,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":47,"doc_module":4,"doc_module_name":46,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":20,"doc_module":4,"doc_module_name":46,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":46,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":46,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":46,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":46,"category_name":12,"show_sort_weight":22,"slug":121},"research-report",{"id":123,"doc_module":4,"doc_module_name":46,"category_name":124,"show_sort_weight":125,"slug":126},9,"Religion & Spirituality",20,"religion-spirituality",{"id":125,"doc_module":4,"doc_module_name":46,"category_name":128,"show_sort_weight":125,"slug":129},"World Cup","world-cup",{"id":131,"doc_module":4,"doc_module_name":46,"category_name":132,"show_sort_weight":131,"slug":133},10,"Lifestyle","lifestyle",{"id":135,"doc_module":4,"doc_module_name":46,"category_name":136,"show_sort_weight":106,"slug":137},19,"General","general"]