{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:LIJY3BMTVDLRU7SA5BK6R2DMA7","short_pith_number":"pith:LIJY3BMT","schema_version":"1.0","canonical_sha256":"5a138d8593a8d71a7e40e855e8e86c07eab630ab0a357ec4104df16b740950a9","source":{"kind":"arxiv","id":"2507.14293","version":1},"attestation_state":"computed","paper":{"title":"WebGuard: Building a Generalizable Guardrail for Web Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.AI","authors_text":"Boyuan Zheng, Dawn Song, Huan Sun, Michael Lin, Qinyuan Zheng, Scott Salisbury, Xiang Deng, Yu Su, Zeyi Liao, Zeyuan Liu, Zifan Wang","submitted_at":"2025-07-18T18:06:27Z","abstract_excerpt":"The rapid development of autonomous web agents powered by Large Language Models (LLMs), while greatly elevating efficiency, exposes the frontier risk of taking unintended or harmful actions. This situation underscores an urgent need for effective safety measures, akin to access controls for human users. To address this critical challenge, we introduce WebGuard, the first comprehensive dataset designed to support the assessment of web agent action risks and facilitate the development of guardrails for real-world online environments. In doing so, WebGuard specifically focuses on predicting the o"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2507.14293","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-07-18T18:06:27Z","cross_cats_sorted":["cs.CL","cs.CV"],"title_canon_sha256":"053d41143a3a122f0c9662ea706df7ae58e12a382301aafbe6b305757b447161","abstract_canon_sha256":"6769a4ae6635528dbb727c4cb40736ff5926e790cc455b51efdf241dc51a8f20"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:40:03.837274Z","signature_b64":"mMOIcO+h0vd8DLVApBYoVLjzNjKaUneEQ44ZkYGJSWbg+WPKfqHAeXBB9iXoWJr1ZEed5aXfMCna0ZOYWxizCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5a138d8593a8d71a7e40e855e8e86c07eab630ab0a357ec4104df16b740950a9","last_reissued_at":"2026-07-05T11:40:03.836756Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:40:03.836756Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"WebGuard: Building a Generalizable Guardrail for Web Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.CV"],"primary_cat":"cs.AI","authors_text":"Boyuan Zheng, Dawn Song, Huan Sun, Michael Lin, Qinyuan Zheng, Scott Salisbury, Xiang Deng, Yu Su, Zeyi Liao, Zeyuan Liu, Zifan Wang","submitted_at":"2025-07-18T18:06:27Z","abstract_excerpt":"The rapid development of autonomous web agents powered by Large Language Models (LLMs), while greatly elevating efficiency, exposes the frontier risk of taking unintended or harmful actions. This situation underscores an urgent need for effective safety measures, akin to access controls for human users. To address this critical challenge, we introduce WebGuard, the first comprehensive dataset designed to support the assessment of web agent action risks and facilitate the development of guardrails for real-world online environments. In doing so, WebGuard specifically focuses on predicting the o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.14293","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2507.14293/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2507.14293","created_at":"2026-07-05T11:40:03.836823+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.14293v1","created_at":"2026-07-05T11:40:03.836823+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.14293","created_at":"2026-07-05T11:40:03.836823+00:00"},{"alias_kind":"pith_short_12","alias_value":"LIJY3BMTVDLR","created_at":"2026-07-05T11:40:03.836823+00:00"},{"alias_kind":"pith_short_16","alias_value":"LIJY3BMTVDLRU7SA","created_at":"2026-07-05T11:40:03.836823+00:00"},{"alias_kind":"pith_short_8","alias_value":"LIJY3BMT","created_at":"2026-07-05T11:40:03.836823+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08147","citing_title":"Prismata: Confining Cross-Site Prompt Injection in Web Agents","ref_index":100,"is_internal_anchor":true},{"citing_arxiv_id":"2606.14517","citing_title":"From Shield to Target: Denial-of-Service Attacks on LLM-Based Agent Guardrails","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.23989","citing_title":"Towards trustworthy agentic AI: a comprehensive survey of safety, robustness, privacy, and system security","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2510.13727","citing_title":"From Refusal to Recovery: A Control-Theoretic Approach to Generative AI Guardrails","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25562","citing_title":"SnapGuard: Lightweight Prompt Injection Detection for Screenshot-Based Web Agents","ref_index":58,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05229","citing_title":"From Governance Norms to Enforceable Controls: A Layered Translation Method for Runtime Guardrails in Agentic AI","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19818","citing_title":"Beyond Task Success: An Evidence-Synthesis Framework for Evaluating, Governing, and Orchestrating Agentic AI","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LIJY3BMTVDLRU7SA5BK6R2DMA7","json":"https://pith.science/pith/LIJY3BMTVDLRU7SA5BK6R2DMA7.json","graph_json":"https://pith.science/api/pith-number/LIJY3BMTVDLRU7SA5BK6R2DMA7/graph.json","events_json":"https://pith.science/api/pith-number/LIJY3BMTVDLRU7SA5BK6R2DMA7/events.json","paper":"https://pith.science/paper/LIJY3BMT"},"agent_actions":{"view_html":"https://pith.science/pith/LIJY3BMTVDLRU7SA5BK6R2DMA7","download_json":"https://pith.science/pith/LIJY3BMTVDLRU7SA5BK6R2DMA7.json","view_paper":"https://pith.science/paper/LIJY3BMT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.14293&json=true","fetch_graph":"https://pith.science/api/pith-number/LIJY3BMTVDLRU7SA5BK6R2DMA7/graph.json","fetch_events":"https://pith.science/api/pith-number/LIJY3BMTVDLRU7SA5BK6R2DMA7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LIJY3BMTVDLRU7SA5BK6R2DMA7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LIJY3BMTVDLRU7SA5BK6R2DMA7/action/storage_attestation","attest_author":"https://pith.science/pith/LIJY3BMTVDLRU7SA5BK6R2DMA7/action/author_attestation","sign_citation":"https://pith.science/pith/LIJY3BMTVDLRU7SA5BK6R2DMA7/action/citation_signature","submit_replication":"https://pith.science/pith/LIJY3BMTVDLRU7SA5BK6R2DMA7/action/replication_record"}},"created_at":"2026-07-05T11:40:03.836823+00:00","updated_at":"2026-07-05T11:40:03.836823+00:00"}