{"id":"https://openalex.org/W7204276245","doi":"https://doi.org/10.48550/arxiv.2608.24275","title":"RePolicy: Learning to Select and Apply Safety Policies for Agent Safeguards","display_name":"RePolicy: Learning to Select and Apply Safety Policies for Agent Safeguards","publication_year":2026,"publication_date":"2026-08-25","ids":{"openalex":"https://openalex.org/W7204276245","doi":"https://doi.org/10.48550/arxiv.2608.24275"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2608.24275","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24275","pdf_url":"https://arxiv.org/pdf/2608.24275","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2608.24275","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128104215","display_name":"Houcheng Jiang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Jiang, Houcheng","raw_affiliation_strings":["University of Science and Technology of China ,","Zhongguancun Academy ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Science and Technology of China ,","institution_ids":["https://openalex.org/I126520041"]},{"raw_affiliation_string":"Zhongguancun Academy ,","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5148351731","display_name":"Boxuan Zhang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhang, Boxuan","raw_affiliation_strings":["University of Science and Technology of China ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Science and Technology of China ,","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051197073","display_name":"Qiyong Zhong","orcid":"https://orcid.org/0009-0005-7576-9556","observed_orcids":["https://orcid.org/0009-0005-7576-9556"]},"institutions":[{"id":"https://openalex.org/I76130692","display_name":"Zhejiang University","ror":"https://ror.org/00a2xv884","country_code":"CN","type":"education","lineage":["https://openalex.org/I76130692"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhong, Qiyong","raw_affiliation_strings":["Zhejiang University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Zhejiang University","institution_ids":["https://openalex.org/I76130692"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126089354","display_name":"Junfeng Fang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I165932596","display_name":"National University of Singapore","ror":"https://ror.org/01tgyzw49","country_code":"SG","type":"education","lineage":["https://openalex.org/I165932596"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Fang, Junfeng","raw_affiliation_strings":["National University of Singapore ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"National University of Singapore ,","institution_ids":["https://openalex.org/I165932596"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135856890","display_name":"Xiang Wang","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wang, Xiang","raw_affiliation_strings":["University of Science and Technology of China ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Science and Technology of China ,","institution_ids":["https://openalex.org/I126520041"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5127314158","display_name":"Xiangnan He","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I126520041","display_name":"University of Science and Technology of China","ror":"https://ror.org/04c4dkn09","country_code":"CN","type":"education","lineage":["https://openalex.org/I126520041","https://openalex.org/I19820366"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"He, Xiangnan","raw_affiliation_strings":["University of Science and Technology of China ,","Zhongguancun Academy ,"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"University of Science and Technology of China ,","institution_ids":["https://openalex.org/I126520041"]},{"raw_affiliation_string":"Zhongguancun Academy ,","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.4636000096797943,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.4636000096797943,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10883","display_name":"Ethics and Social Impacts of AI","score":0.19900000095367432,"subfield":{"id":"https://openalex.org/subfields/3311","display_name":"Safety Research"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T13295","display_name":"Safety Systems Engineering in Autonomy","score":0.07580000162124634,"subfield":{"id":"https://openalex.org/subfields/2213","display_name":"Safety, Risk, Reliability and Quality"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/language-model-agents","display_name":"language model agents","score":0.7870000004768372},{"id":"https://openalex.org/keywords/agent-safety","display_name":"agent safety","score":0.7639999985694885},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"reinforcement learning","score":0.7289999723434448},{"id":"https://openalex.org/keywords/group-relative-policy-optimization","display_name":"Group Relative Policy Optimization","score":0.6299999952316284}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7260000109672546},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7003999948501587},{"id":"https://openalex.org/C2776527387","wikidata":"https://www.wikidata.org/wiki/Q1671839","display_name":"Invocation","level":2,"score":0.6985999941825867},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.6549000144004822},{"id":"https://openalex.org/C2776743756","wikidata":"https://www.wikidata.org/wiki/Q5097921","display_name":"Safeguarding","level":2,"score":0.5952000021934509},{"id":"https://openalex.org/C2780771206","wikidata":"https://www.wikidata.org/wiki/Q3271761","display_name":"Safeguard","level":2,"score":0.5447999835014343},{"id":"https://openalex.org/C2779436431","wikidata":"https://www.wikidata.org/wiki/Q30672407","display_name":"Policy learning","level":2,"score":0.5166000127792358},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.501800000667572},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.4894999861717224},{"id":"https://openalex.org/C85847156","wikidata":"https://www.wikidata.org/wiki/Q59015987","display_name":"Verifiable secret sharing","level":3,"score":0.4327000081539154},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.4171999990940094},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.40849998593330383},{"id":"https://openalex.org/C154908896","wikidata":"https://www.wikidata.org/wiki/Q2167404","display_name":"Security policy","level":2,"score":0.3479999899864197},{"id":"https://openalex.org/C138673069","wikidata":"https://www.wikidata.org/wiki/Q322229","display_name":"Tracing","level":2,"score":0.3127000033855438},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.3124000132083893},{"id":"https://openalex.org/C527821871","wikidata":"https://www.wikidata.org/wiki/Q228502","display_name":"Access control","level":2,"score":0.3073999881744385},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.30079999566078186},{"id":"https://openalex.org/C154690210","wikidata":"https://www.wikidata.org/wiki/Q1668499","display_name":"Rewriting","level":2,"score":0.29280000925064087},{"id":"https://openalex.org/C2775924081","wikidata":"https://www.wikidata.org/wiki/Q55608371","display_name":"Control (management)","level":2,"score":0.28619998693466187},{"id":"https://openalex.org/C41550386","wikidata":"https://www.wikidata.org/wiki/Q529909","display_name":"Multi-agent system","level":2,"score":0.27810001373291016},{"id":"https://openalex.org/C74072328","wikidata":"https://www.wikidata.org/wiki/Q1142726","display_name":"Intelligent agent","level":2,"score":0.273499995470047},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.27300000190734863},{"id":"https://openalex.org/C113843644","wikidata":"https://www.wikidata.org/wiki/Q901882","display_name":"Interface (matter)","level":4,"score":0.26809999346733093},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.26649999618530273},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.26579999923706055},{"id":"https://openalex.org/C137703981","wikidata":"https://www.wikidata.org/wiki/Q4692093","display_name":"Agent architecture","level":3,"score":0.2565999925136566}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2608.24275","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24275","pdf_url":"https://arxiv.org/pdf/2608.24275","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2608.24275","doi":"https://doi.org/10.48550/arxiv.2608.24275","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2608.24275","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2608.24275","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2608.24275","pdf_url":"https://arxiv.org/pdf/2608.24275","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7204276245.pdf","grobid_xml":"https://content.openalex.org/works/W7204276245.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7214751197","https://openalex.org/W4417068164","https://openalex.org/W4414922625","https://openalex.org/W4399695705","https://openalex.org/W4415336612","https://openalex.org/W4404348650","https://openalex.org/W7166098801","https://openalex.org/W4406806966","https://openalex.org/W7223250153","https://openalex.org/W2744647809"],"abstract_inverted_index":{"Agent":[0],"safeguards":[1,39,183],"must":[2],"determine":[3],"whether":[4,117],"an":[5,70,76,94],"agent's":[6],"behavior":[7,119],"violates":[8],"the":[9,13,30,35,49,105,110,118],"safety":[10,78,91,137,187],"rules":[11,21],"of":[12,100,127],"current":[14],"task,":[15],"environment,":[16],"and":[17,90,97,115,139,156,166,175],"user.":[18],"Because":[19],"these":[20],"vary":[22],"across":[23],"contexts,":[24],"a":[25,42,63,66,83,98,125],"safeguard":[26,84],"needs":[27],"to":[28,61,75,158,189],"select":[29,62],"relevant":[31],"policy":[32,51,64,88,106,153],"before":[33],"judging":[34],"behavior.":[36,194],"Existing":[37],"policy-aware":[38],"can":[40,73,184],"apply":[41],"supplied":[43],"policy,":[44],"but":[45],"generally":[46],"assume":[47],"that":[48,85,107,182],"correct":[50],"is":[52,120],"already":[53],"known.":[54],"They":[55],"therefore":[56],"do":[57],"not":[58],"learn":[59,185],"how":[60],"from":[65],"library,":[67],"even":[68],"though":[69],"incorrect":[71,77],"choice":[72],"lead":[74],"judgment.":[79],"We":[80,122,141],"introduce":[81],"RePolicy,":[82],"jointly":[86],"learns":[87],"selection":[89],"assessment.":[92],"Given":[93],"agent":[95,131,193],"trajectory":[96],"set":[99],"candidate":[101],"policies,":[102,136],"RePolicy":[103,143],"selects":[104],"best":[108],"matches":[109],"situation,":[111],"applies":[112],"its":[113],"requirements,":[114],"predicts":[116],"safe.":[121],"construct":[123],"PolicyTraj-20K,":[124],"dataset":[126],"more":[128],"than":[129],"20,000":[130],"trajectories":[132],"annotated":[133],"with":[134,144],"applicable":[135],"labels,":[138],"explanations.":[140],"train":[142],"supervised":[145],"demonstrations":[146],"followed":[147],"by":[148],"reinforcement":[149],"learning,":[150],"while":[151],"randomizing":[152],"order,":[154],"identifiers,":[155],"distractors":[157],"reduce":[159],"shortcut":[160],"learning.":[161],"Across":[162],"six":[163],"agent-safety":[164],"benchmarks":[165],"19":[167],"baselines,":[168],"RePolicy-4B":[169],"achieves":[170],"leading":[171],"overall":[172],"safety-detection":[173],"performance":[174],"improves":[176],"policy-selection":[177],"accuracy.":[178],"These":[179],"results":[180],"show":[181],"which":[186],"requirements":[188],"use":[190],"for":[191],"assessing":[192]},"counts_by_year":[],"updated_date":"2026-10-31T12:09:11.096373","created_date":"2026-08-27T00:00:00"}
