{"id":"https://openalex.org/W4412889677","doi":"https://doi.org/10.18653/v1/2025.acl-long.1435","title":"The Task Shield: Enforcing Task Alignment to Defend Against Indirect Prompt Injection in LLM Agents","display_name":"The Task Shield: Enforcing Task Alignment to Defend Against Indirect Prompt Injection in LLM Agents","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W4412889677","doi":"https://doi.org/10.18653/v1/2025.acl-long.1435"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2025.acl-long.1435","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1435","pdf_url":"https://aclanthology.org/2025.acl-long.1435.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2025.acl-long.1435.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5047396940","display_name":"Feiran Jia","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Feiran Jia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104204525","display_name":"Tong Wu","orcid":"https://orcid.org/0000-0003-2370-4644"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tong Wu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050608316","display_name":"Qin Xin","orcid":"https://orcid.org/0000-0002-6178-8538"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xin Qin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5062519505","display_name":"Anna Squicciarini","orcid":"https://orcid.org/0000-0002-7396-1895"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Anna Squicciarini","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":26.2803,"has_fulltext":true,"cited_by_count":10,"citation_normalized_percentile":{"value":0.99716332,"is_in_top_1_percent":true,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":91,"max":100},"biblio":{"volume":null,"issue":null,"first_page":"29680","last_page":"29697"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10551","display_name":"Scheduling and Optimization Algorithms","score":0.9584000110626221,"subfield":{"id":"https://openalex.org/subfields/2209","display_name":"Industrial and Manufacturing Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10551","display_name":"Scheduling and Optimization Algorithms","score":0.9584000110626221,"subfield":{"id":"https://openalex.org/subfields/2209","display_name":"Industrial and Manufacturing Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10703","display_name":"Business Process Modeling and Analysis","score":0.9544000029563904,"subfield":{"id":"https://openalex.org/subfields/1404","display_name":"Management Information Systems"},"field":{"id":"https://openalex.org/fields/14","display_name":"Business, Management and Accounting"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10142","display_name":"Formal Methods in Verification","score":0.9502999782562256,"subfield":{"id":"https://openalex.org/subfields/1703","display_name":"Computational Theory and Mathematics"},"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/task","display_name":"Task (project management)","score":0.7867745161056519},{"id":"https://openalex.org/keywords/shield","display_name":"Shield","score":0.6750355362892151},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6401895880699158},{"id":"https://openalex.org/keywords/computer-security","display_name":"Computer security","score":0.4433761239051819},{"id":"https://openalex.org/keywords/human\u2013computer-interaction","display_name":"Human\u2013computer interaction","score":0.3227064907550812},{"id":"https://openalex.org/keywords/engineering","display_name":"Engineering","score":0.1824277937412262},{"id":"https://openalex.org/keywords/systems-engineering","display_name":"Systems engineering","score":0.12272235751152039},{"id":"https://openalex.org/keywords/geology","display_name":"Geology","score":0.08519566059112549}],"concepts":[{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.7867745161056519},{"id":"https://openalex.org/C138081364","wikidata":"https://www.wikidata.org/wiki/Q852013","display_name":"Shield","level":2,"score":0.6750355362892151},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6401895880699158},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.4433761239051819},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.3227064907550812},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.1824277937412262},{"id":"https://openalex.org/C201995342","wikidata":"https://www.wikidata.org/wiki/Q682496","display_name":"Systems engineering","level":1,"score":0.12272235751152039},{"id":"https://openalex.org/C127313418","wikidata":"https://www.wikidata.org/wiki/Q1069","display_name":"Geology","level":0,"score":0.08519566059112549},{"id":"https://openalex.org/C5900021","wikidata":"https://www.wikidata.org/wiki/Q163082","display_name":"Petrology","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2025.acl-long.1435","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1435","pdf_url":"https://aclanthology.org/2025.acl-long.1435.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2025.acl-long.1435","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2025.acl-long.1435","pdf_url":"https://aclanthology.org/2025.acl-long.1435.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/5","display_name":"Gender equality","score":0.5099999904632568}],"awards":[],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4412889677.pdf","grobid_xml":"https://content.openalex.org/works/W4412889677.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W191020423","https://openalex.org/W212028160","https://openalex.org/W4388071457","https://openalex.org/W2384581482","https://openalex.org/W4379056371","https://openalex.org/W3196817267","https://openalex.org/W1976600725"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Model":[2],"(LLM)":[3],"agents":[4,58],"are":[5],"increasingly":[6],"being":[7],"deployed":[8],"as":[9],"conversational":[10],"assistants":[11],"capable":[12],"of":[13],"performing":[14],"complex":[15],"realworld":[16],"tasks":[17],"through":[18],"tool":[19,122],"integration.This":[20],"enhanced":[21],"ability":[22],"to":[23,59,70,92,100,125],"interact":[24],"with":[25],"external":[26,53],"systems":[27],"and":[28,81,121],"process":[29],"various":[30,156],"data":[31,54],"sources,":[32],"while":[33,74,143],"powerful,":[34],"introduces":[35],"significant":[36],"security":[37,73,87],"vulnerabilities.In":[38],"particular,":[39],"indirect":[40],"prompt":[41],"injection":[42],"attacks":[43],"pose":[44],"a":[45,79,111],"critical":[46],"threat,":[47],"where":[48],"malicious":[49],"instructions":[50],"embedded":[51],"within":[52],"sources":[55],"can":[56],"manipulate":[57],"deviate":[60],"from":[61,88],"user":[62,102],"intentions.While":[63],"existing":[64,153],"defenses":[65,154],"show":[66],"promise,":[67],"they":[68],"struggle":[69],"maintain":[71],"robust":[72],"preserving":[75],"task":[76,94,146],"functionality.We":[77],"propose":[78],"novel":[80],"orthogonal":[82],"perspective":[83],"that":[84,115,135],"reframes":[85],"agent":[86,98],"preventing":[89],"harmful":[90],"actions":[91],"ensuring":[93],"alignment,":[95],"requiring":[96],"every":[97],"action":[99],"serve":[101],"objectives.Based":[103],"on":[104,129,149],"this":[105],"insight,":[106],"we":[107,133],"develop":[108],"Task":[109,136],"Shield,":[110],"test-time":[112],"defense":[113],"mechanism":[114],"systematically":[116],"verifies":[117],"whether":[118],"each":[119],"instruction":[120],"call":[123],"contributes":[124],"user-specified":[126],"goals.Through":[127],"experiments":[128],"the":[130],"Agent-Dojo":[131],"benchmark,":[132],"demonstrate":[134],"Shield":[137],"reduces":[138],"attack":[139],"success":[140],"rates":[141],"(2.07%)":[142],"maintaining":[144],"high":[145],"utility":[147],"(69.79%)":[148],"GPT-4o,":[150],"significantly":[151],"outperforming":[152],"in":[155],"real-world":[157],"scenarios.":[158]},"counts_by_year":[{"year":2026,"cited_by_count":9},{"year":2025,"cited_by_count":1}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}
