{"id":"https://openalex.org/W7119860193","doi":"https://doi.org/10.36227/techrxiv.176799772.28164151/v1","title":"A Unified Evaluation and Governance Framework for Trustworthy LLM Agents","display_name":"A Unified Evaluation and Governance Framework for Trustworthy LLM Agents","publication_year":2026,"publication_date":"2026-01-09","ids":{"openalex":"https://openalex.org/W7119860193","doi":"https://doi.org/10.36227/techrxiv.176799772.28164151/v1"},"language":"en","primary_location":{"id":"doi:10.36227/techrxiv.176799772.28164151/v1","doi":"https://doi.org/10.36227/techrxiv.176799772.28164151/v1","is_oa":true,"landing_page_url":"https://doi.org/10.36227/techrxiv.176799772.28164151/v1","pdf_url":"https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.176799772.28164151/v1","source":null,"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":"posted-content"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.176799772.28164151/v1","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5120319780","display_name":"Vinay Bamil","orcid":"https://orcid.org/0009-0004-3784-0828","observed_orcids":["https://orcid.org/0009-0004-3784-0828"]},"institutions":[{"id":"https://openalex.org/I44829897","display_name":"Golden Gate University","ror":"https://ror.org/01wj7vt66","country_code":"US","type":"education","lineage":["https://openalex.org/I44829897"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Vinay Bamil","raw_affiliation_strings":["Doctoral Researcher , Golden Gate University (GGU)"],"raw_orcid":"https://orcid.org/0009-0004-3784-0828","affiliations":[{"raw_affiliation_string":"Doctoral Researcher , Golden Gate University (GGU)","institution_ids":["https://openalex.org/I44829897"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5122519550","display_name":"RAMA KRISHNA KUMAR LINGAMGUNTA","orcid":"https://orcid.org/0009-0001-6201-7620","observed_orcids":["https://orcid.org/0009-0001-6201-7620"]},"institutions":[{"id":"https://openalex.org/I44829897","display_name":"Golden Gate University","ror":"https://ror.org/01wj7vt66","country_code":"US","type":"education","lineage":["https://openalex.org/I44829897"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"RAMA KRISHNA KUMAR LINGAMGUNTA","raw_affiliation_strings":["Doctoral Researcher , Golden Gate University (GGU)"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Doctoral Researcher , Golden Gate University (GGU)","institution_ids":["https://openalex.org/I44829897"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I44829897"],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":3,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.5175999999046326,"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/T10883","display_name":"Ethics and Social Impacts of AI","score":0.5175999999046326,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.22380000352859497,"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/T10927","display_name":"Access Control and Trust","score":0.06120000034570694,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/llm-based-agents","display_name":"LLM agents","score":0.8050000071525574},{"id":"https://openalex.org/keywords/trustworthy-ai","display_name":"trustworthy AI","score":0.8009999990463257},{"id":"https://openalex.org/keywords/agent-evaluation","display_name":"agent evaluation","score":0.6200000047683716},{"id":"https://openalex.org/keywords/ai-governance","display_name":"AI governance","score":0.7509999871253967},{"id":"https://openalex.org/keywords/hallucination-mitigation","display_name":"hallucination mitigation","score":0.8209999799728394},{"id":"https://openalex.org/keywords/policy-alignment","display_name":"policy alignment","score":0.5270000100135803},{"id":"https://openalex.org/keywords/trust-metrics","display_name":"trust metrics","score":0.6119999885559082},{"id":"https://openalex.org/keywords/auditability","display_name":"auditability","score":0.5630000233650208},{"id":"https://openalex.org/keywords/continuous-assessment","display_name":"continuous assessment","score":0.32100000977516174}],"concepts":[{"id":"https://openalex.org/C153701036","wikidata":"https://www.wikidata.org/wiki/Q659974","display_name":"Trustworthiness","level":2,"score":0.6151999831199646},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5389999747276306},{"id":"https://openalex.org/C39389867","wikidata":"https://www.wikidata.org/wiki/Q380767","display_name":"Corporate governance","level":2,"score":0.5020999908447266},{"id":"https://openalex.org/C195094911","wikidata":"https://www.wikidata.org/wiki/Q14167904","display_name":"Process management","level":1,"score":0.5019999742507935},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.44940000772476196},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.4359999895095825},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.40450000762939453},{"id":"https://openalex.org/C56739046","wikidata":"https://www.wikidata.org/wiki/Q192060","display_name":"Knowledge management","level":1,"score":0.3912999927997589},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.3790999948978424},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.34549999237060547},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.33340001106262207},{"id":"https://openalex.org/C539667460","wikidata":"https://www.wikidata.org/wiki/Q2414942","display_name":"Management science","level":1,"score":0.3156000077724457},{"id":"https://openalex.org/C2778137410","wikidata":"https://www.wikidata.org/wiki/Q2732820","display_name":"Government (linguistics)","level":2,"score":0.31369999051094055},{"id":"https://openalex.org/C2776007630","wikidata":"https://www.wikidata.org/wiki/Q2798912","display_name":"Accountability","level":2,"score":0.3037000000476837},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.2831999957561493},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2775000035762787},{"id":"https://openalex.org/C2780233690","wikidata":"https://www.wikidata.org/wiki/Q535347","display_name":"Transparency (behavior)","level":2,"score":0.2757999897003174},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.2678000032901764},{"id":"https://openalex.org/C189922023","wikidata":"https://www.wikidata.org/wiki/Q17056348","display_name":"Information governance","level":4,"score":0.2630000114440918},{"id":"https://openalex.org/C36503486","wikidata":"https://www.wikidata.org/wiki/Q11235244","display_name":"Domain (mathematical analysis)","level":2,"score":0.258899986743927},{"id":"https://openalex.org/C110354214","wikidata":"https://www.wikidata.org/wiki/Q6314146","display_name":"Engineering management","level":1,"score":0.2554999887943268},{"id":"https://openalex.org/C182306322","wikidata":"https://www.wikidata.org/wiki/Q1779371","display_name":"Order (exchange)","level":2,"score":0.251800000667572}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.36227/techrxiv.176799772.28164151/v1","doi":"https://doi.org/10.36227/techrxiv.176799772.28164151/v1","is_oa":true,"landing_page_url":"https://doi.org/10.36227/techrxiv.176799772.28164151/v1","pdf_url":"https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.176799772.28164151/v1","source":null,"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":"posted-content"}],"best_oa_location":{"id":"doi:10.36227/techrxiv.176799772.28164151/v1","doi":"https://doi.org/10.36227/techrxiv.176799772.28164151/v1","is_oa":true,"landing_page_url":"https://doi.org/10.36227/techrxiv.176799772.28164151/v1","pdf_url":"https://www.techrxiv.org/doi/pdf/10.36227/techrxiv.176799772.28164151/v1","source":null,"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":"posted-content"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7119860193.pdf","grobid_xml":"https://content.openalex.org/works/W7119860193.grobid-xml"},"referenced_works_count":8,"referenced_works":["https://openalex.org/W3210904070","https://openalex.org/W4296413526","https://openalex.org/W4304195432","https://openalex.org/W4311991106","https://openalex.org/W4327810158","https://openalex.org/W4377121468","https://openalex.org/W4386721614","https://openalex.org/W4402671286"],"related_works":["https://openalex.org/W7125105280","https://openalex.org/W7163194659","https://openalex.org/W7171349057","https://openalex.org/W7204787896","https://openalex.org/W7212178473","https://openalex.org/W7221783937","https://openalex.org/W4414973799","https://openalex.org/W7124421359","https://openalex.org/W7131893979","https://openalex.org/W4412877234"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Model":[2],"(LLM)":[3],"agents":[4,28,255],"are":[5],"increasingly":[6],"deployed":[7],"as":[8],"autonomous":[9],"decision-support":[10],"and":[11,18,43,55,75,90,100,129,141,166,169,195,231,245,252],"workflow-automation":[12],"components":[13],"in":[14,23,81,223,228,256],"enterprise,":[15],"healthcare,":[16],"finance,":[17],"public-sector":[19],"environments.":[20,259],"Despite":[21],"advances":[22],"model":[24],"capability,":[25],"current":[26],"LLM":[27,107,254],"still":[29],"exhibit":[30],"persistent":[31],"weaknesses,":[32],"including":[33,217],"hallucinated":[34],"reasoning":[35],"steps,":[36],"unverifiable":[37],"factual":[38],"assertions,":[39],"opaque":[40],"decision":[41,102],"traces,":[42],"inconsistent":[44],"adherence":[45],"to":[46,219],"operational":[47,258],"policies.":[48],"Existing":[49],"evaluation":[50,89],"methodologies":[51],"emphasize":[52],"offline":[53],"benchmarks":[54],"text-generation":[56],"metrics,":[57],"whereas":[58],"governance":[59,91],"mechanisms":[60],"typically":[61],"rely":[62],"on":[63],"post-hoc":[64],"filtering.":[65],"This":[66,84],"separation":[67],"creates":[68],"a":[69,87,147,152,156,160,164,170,225,243],"structural":[70],"gap":[71],"between":[72],"agent":[73,76,108],"capability":[74],"controllability,":[77],"hindering":[78],"safe":[79],"deployment":[80],"mission-critical":[82],"settings.":[83],"paper":[85],"introduces":[86],"unified":[88],"framework":[92,182,241],"that":[93,174,238],"embeds":[94],"multi-layer":[95],"verification,":[96],"retrievalaugmented":[97],"grounding":[98,229],"assessment,":[99],"policy-aligned":[101],"controls":[103],"directly":[104],"within":[105],"the":[106,180,204,239],"inference":[109],"loop.":[110],"We":[111,144,178],"define":[112],"four":[113],"quantitative":[114],"trust":[115],"metrics\u2014the":[116],"Agent":[117],"Reliability":[118],"Score":[119,132],"(ARS),":[120],"RAG":[121,157],"Grounding":[122,158],"Confidence":[123],"(RGC),":[124],"Attribution":[125],"Completeness":[126],"Ratio":[127],"(ACR),":[128],"Policy-Aligned":[130],"Action":[131],"(PAAS)\u2014to":[133],"jointly":[134],"measure":[135],"end-to-end":[136],"correctness,":[137],"evidential":[138],"grounding,":[139],"transparency,":[140],"policy":[142],"compliance.":[143],"further":[145],"propose":[146],"modular":[148],"architecture":[149],"consisting":[150],"of":[151],"structured":[153],"Reasoning":[154],"Engine,":[155],"Optimizer,":[159],"Multi-Layer":[161],"Verification":[162],"Module,":[163],"Governance":[165],"Audit":[167],"Layer,":[168],"Continuous":[171],"Evaluation":[172],"Loop":[173],"adaptively":[175],"regulates":[176],"autonomy.":[177],"evaluate":[179],"proposed":[181,240],"through":[183],"two":[184],"real-world":[185],"case":[186],"studies:":[187],"(i)":[188],"enterprise":[189],"workflow":[190],"automation":[191],"using":[192,203],"Salesforce":[193],"Agentforce,":[194],"(ii)":[196],"causal":[197],"evidence":[198],"synthesis":[199],"for":[200,248],"global":[201],"development":[202],"World":[205],"Bank":[206],"DIME":[207],"ImpactAI":[208],"platform.":[209],"Results":[210],"show":[211],"substantial":[212],"gains":[213],"over":[214],"baseline":[215],"agents,":[216],"up":[218],"an":[220],"88%":[221],"reduction":[222],"hallucinations,":[224],"20\u201330%":[226],"improvement":[227],"confidence,":[230],"near-perfect":[232],"policy-alignment":[233],"scores.":[234],"These":[235],"findings":[236],"indicate":[237],"provides":[242],"practical":[244],"principled":[246],"foundation":[247],"deploying":[249],"trustworthy,":[250],"auditable,":[251],"governable":[253],"high-stakes":[257]},"counts_by_year":[{"year":2026,"cited_by_count":3}],"updated_date":"2026-11-09T13:22:03.289127","created_date":"2026-01-10T00:00:00"}