{"paper":{"title":"Self-Evolving Agent Harness","arxiv_id":"2610.18813","primary_cat":"cs.AI","submitted_at":"2026-10-13T17:48:49+00:00","version_compiled":"v1"},"notice":"Every quoted sentence is the authors' text from the arXiv listing abstract. Labels and probabilities are the outputs of classifier-1.13.0 on the question bank abstract-p1-20260923, exactly as returned. No sentence in this record was written by a language model.","schema":"pith.compiled-record/v0","status":"compiled","abstract":{"source":"arXiv listing abstract v1","sentences":[{"i":0,"p":0.45,"id":"s0","own":false,"role":"contribution","text":"Harness self-evolution revises an agent's software scaffolding, including prompts, memory, workflows, and executable components, while keeping model weights fixed.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.33,"main_claim_p":0.1,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.0,"method":0.43,"result":0.0,"background":0.12,"limitation":0.0,"future_work":0.0,"contribution":0.45,"interpretation":0.0}},{"i":1,"p":0.63,"id":"s1","own":true,"role":"method","text":"We formalize a feedback-driven process in which accepted revisions persist across runs.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.88,"main_claim_p":0.07,"quantitative":false,"quantitative_p":0.03,"role_probabilities":{"gap":0.0,"method":0.63,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.37,"interpretation":0.0}},{"i":2,"p":0.82,"id":"s2","own":true,"role":"method","text":"The framework separates update targets from proposal and admission mechanisms.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.68,"main_claim_p":0.05,"quantitative":false,"quantitative_p":0.05,"role_probabilities":{"gap":0.0,"method":0.82,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.18,"interpretation":0.0}},{"i":3,"p":0.76,"id":"s3","own":true,"role":"method","text":"It distinguishes changes to interpreted state from revisions to tools and code, repository-level control logic, and plugin composition.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.58,"main_claim_p":0.04,"quantitative":false,"quantitative_p":0.07,"role_probabilities":{"gap":0.0,"method":0.76,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.24,"interpretation":0.0}},{"i":4,"p":0.55,"id":"s4","own":true,"role":"method","text":"We characterize search through evolutionary optimization and examine its reliability through continual PAC-learning.","hedge":{"p":0.97,"label":"asserted"},"own_p":0.81,"main_claim_p":0.07,"quantitative":false,"quantitative_p":0.06,"role_probabilities":{"gap":0.0,"method":0.55,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.45,"interpretation":0.0}},{"i":5,"p":0.83,"id":"s5","own":true,"role":"result","text":"Under explicit assumptions, the analysis yields an expected generalization bound, a retention bound conditional on historical regression budgets, and a high-probability admission rule using fresh, independent validation data.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.9,"main_claim_p":0.25,"quantitative":true,"quantitative_p":0.95,"role_probabilities":{"gap":0.0,"method":0.02,"result":0.83,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.15,"interpretation":0.0}},{"i":6,"p":0.87,"id":"s6","own":false,"role":"interpretation","text":"We examine how persistent harness revisions affect knowledge sharing and organizational workflows, with implications for scientific reliability, physical safety, and institutional accountability.","hedge":{"p":0.74,"label":"asserted"},"own_p":0.47,"main_claim_p":0.2,"quantitative":false,"quantitative_p":0.03,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.0,"limitation":0.0,"future_work":0.0,"contribution":0.13,"interpretation":0.87}},{"i":7,"p":0.83,"id":"s7","own":false,"role":"background","text":"Cases from prior work illustrate model--harness co-design, distributed execution, and plugin reuse.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.08,"main_claim_p":0.03,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.04,"result":0.01,"background":0.83,"limitation":0.0,"future_work":0.0,"contribution":0.12,"interpretation":0.0}},{"i":8,"p":0.76,"id":"s8","own":false,"role":"interpretation","text":"Reliable deployment requires continual validation as task distributions and evaluators change, with independent safeguards for learning, runtime privileges, and consequential actions.","hedge":{"p":1.0,"label":"asserted"},"own_p":0.19,"main_claim_p":0.28,"quantitative":false,"quantitative_p":0.04,"role_probabilities":{"gap":0.0,"method":0.01,"result":0.0,"background":0.01,"limitation":0.04,"future_work":0.18,"contribution":0.0,"interpretation":0.76}},{"i":9,"p":0.98,"id":"s9","own":true,"role":"contribution","text":"We provide a comprehensive 66-page discussion of related work, theory, algorithms, evaluation, applications, and the broader impact of Harness RSI, together with continuously updated code and supplementary materials, at https://github.com/zgcagi/Harness-RSI.","hedge":{"p":0.99,"label":"asserted"},"own_p":0.81,"main_claim_p":0.04,"quantitative":true,"quantitative_p":0.97,"role_probabilities":{"gap":0.0,"method":0.0,"result":0.0,"background":0.01,"limitation":0.0,"future_work":0.01,"contribution":0.98,"interpretation":0.0}}]},"compiled":{"main_claim":{"p":0.89,"text":"Reliable deployment requires continual validation as task distributions and evaluators change, with independent safeguards for learning, runtime privileges, and consequential actions.","location":"abstract s8","hedge_strength":{"p":1.0,"label":"asserted"}},"provenance":{"route":{"via":"TypeSafe API","model":"classifier-1.13.0","endpoint":"https://api.typesafe.ai/v1/systemone","provider":"TypeSafe"},"packing":"one request per paper","requests":1,"input_tokens":7458,"cost_usd_reported":0.0},"compiled_at":"2026-10-14","bank_version":"abstract-p1-20260923","model_version":"classifier-1.13.0","abstract_states":{"claims_first":0.05,"states_limitation":0.71,"human_participants":0.12,"controlled_comparison":0.06,"randomized_assignment":0.02,"states_quantitative_result":0.71,"compares_with_prior_results":0.25,"mentions_code_or_data_availability":0.98},"evidential_basis":{"p":0.94,"value":"proof or derivation","probabilities":{"argument":0.04,"literature":0.01,"proof_or_derivation":0.94,"analysis_of_existing_data":0.0,"observation_or_experiment":0.0,"simulation_or_computation":0.0}},"contribution_type":{"p":0.99,"value":"method or tool","probabilities":{"review":0.0,"method_or_tool":0.99,"empirical_study":0.0,"simulation_study":0.0,"theoretical_result":0.01,"dataset_or_benchmark":0.0,"position_or_commentary":0.0}},"contribution_statement":{"p":0.69,"text":"We formalize a feedback-driven process in which accepted revisions persist across runs.","location":"abstract s1"}},"coverage":{"tier":"abstract","levels_run":["P abstract"],"not_examined":["full text","figures and tables","references and citing sentences","relations to other papers"],"sources_read":["arXiv listing abstract v1"],"levels_not_run":["S sections","T body sentences","C citations","R cross-paper relations"],"sentences_examined":10,"sentences_in_abstract":10},"measured_accuracy":{"url":"https://pith.science/accuracy","measured_on":"2026-10-02","method":"two independent model families labelled a fresh random sample blind; the figure is how often the published label matches their consensus","sentence_role_agreement":0.92,"main_claim_agreement":0.908,"typed_citation_agreement":0.933,"typed_citation_agreement_95_interval":[0.917,0.945],"review_comments_accurate":0.758,"review_comments_accurate_95_interval":[0.684,0.82],"review_caught_admitted_error_found_or_related":0.448,"review_comments_judged":{"judged_by":["gpt-5.6-luna","grok-4.7"],"judged_against":"Pith's own text copy of each paper: text extracted from the PDF, the first 90,000 characters","measured_on":"2026-10-04","accurate":113,"decided":149,"sampled_reviews":200,"split_by_lane":false,"by_lane":"https://pith.science/accuracy#by-lane"}}}