{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:5B7RV4TJQHTNC2EGAJENX3C4XO","short_pith_number":"pith:5B7RV4TJ","schema_version":"1.0","canonical_sha256":"e87f1af26981e6d168860248dbec5cbbbb94fa87b18e0dfca7d2d028f18e9109","source":{"kind":"arxiv","id":"2607.07405","version":1},"attestation_state":"computed","paper":{"title":"Reason Less, Verify More: Deterministic Gates Recover a Silent Policy-Violation Failure Mode in Tool-Using LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.AI","authors_text":"Abhishek Basu, Sumanth Reddy Challaram, Vikas Reddy","submitted_at":"2026-07-08T13:38:54Z","abstract_excerpt":"Tool-using LLM agents can violate the very policies they are deployed to enforce while appearing to complete the task successfully. In policy-permissive environments, a tool may execute any well-formed call even when the corresponding state transition is forbidden by domain policy. The result is a silent wrong state (a booking cancelled, a passenger count changed, a claim acted on without verification) that neither the tool nor the agent's self-report exposes.\n  We study this failure mode in the $\\tau^2$-bench airline domain. On a budget agent, 78% of observed failures are silent wrong-state f"},"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":"2607.07405","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-07-08T13:38:54Z","cross_cats_sorted":["cs.CR"],"title_canon_sha256":"fd9fc2c42a6b4e4945c7d7ae87d5bd8542a4db0bd83a3c316f7fd4e32d29a80b","abstract_canon_sha256":"6e27de801ff9cb83b4f0c60a6b972aadc1b92eea6f8b45b4428b2f8356dbf29a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-09T01:20:24.121751Z","signature_b64":"dYwJz5tH06qEywOHhs5g/xLCeOonpjVj1lURoFBqailT7Isjn+p94nqubsawMpXsP1J5fiLEtRdcOm3IhNfHAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e87f1af26981e6d168860248dbec5cbbbb94fa87b18e0dfca7d2d028f18e9109","last_reissued_at":"2026-07-09T01:20:24.121335Z","signature_status":"signed_v1","first_computed_at":"2026-07-09T01:20:24.121335Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reason Less, Verify More: Deterministic Gates Recover a Silent Policy-Violation Failure Mode in Tool-Using LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CR"],"primary_cat":"cs.AI","authors_text":"Abhishek Basu, Sumanth Reddy Challaram, Vikas Reddy","submitted_at":"2026-07-08T13:38:54Z","abstract_excerpt":"Tool-using LLM agents can violate the very policies they are deployed to enforce while appearing to complete the task successfully. In policy-permissive environments, a tool may execute any well-formed call even when the corresponding state transition is forbidden by domain policy. The result is a silent wrong state (a booking cancelled, a passenger count changed, a claim acted on without verification) that neither the tool nor the agent's self-report exposes.\n  We study this failure mode in the $\\tau^2$-bench airline domain. On a budget agent, 78% of observed failures are silent wrong-state f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.07405","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/2607.07405/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":"2607.07405","created_at":"2026-07-09T01:20:24.121398+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.07405v1","created_at":"2026-07-09T01:20:24.121398+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.07405","created_at":"2026-07-09T01:20:24.121398+00:00"},{"alias_kind":"pith_short_12","alias_value":"5B7RV4TJQHTN","created_at":"2026-07-09T01:20:24.121398+00:00"},{"alias_kind":"pith_short_16","alias_value":"5B7RV4TJQHTNC2EG","created_at":"2026-07-09T01:20:24.121398+00:00"},{"alias_kind":"pith_short_8","alias_value":"5B7RV4TJ","created_at":"2026-07-09T01:20:24.121398+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5B7RV4TJQHTNC2EGAJENX3C4XO","json":"https://pith.science/pith/5B7RV4TJQHTNC2EGAJENX3C4XO.json","graph_json":"https://pith.science/api/pith-number/5B7RV4TJQHTNC2EGAJENX3C4XO/graph.json","events_json":"https://pith.science/api/pith-number/5B7RV4TJQHTNC2EGAJENX3C4XO/events.json","paper":"https://pith.science/paper/5B7RV4TJ"},"agent_actions":{"view_html":"https://pith.science/pith/5B7RV4TJQHTNC2EGAJENX3C4XO","download_json":"https://pith.science/pith/5B7RV4TJQHTNC2EGAJENX3C4XO.json","view_paper":"https://pith.science/paper/5B7RV4TJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.07405&json=true","fetch_graph":"https://pith.science/api/pith-number/5B7RV4TJQHTNC2EGAJENX3C4XO/graph.json","fetch_events":"https://pith.science/api/pith-number/5B7RV4TJQHTNC2EGAJENX3C4XO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5B7RV4TJQHTNC2EGAJENX3C4XO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5B7RV4TJQHTNC2EGAJENX3C4XO/action/storage_attestation","attest_author":"https://pith.science/pith/5B7RV4TJQHTNC2EGAJENX3C4XO/action/author_attestation","sign_citation":"https://pith.science/pith/5B7RV4TJQHTNC2EGAJENX3C4XO/action/citation_signature","submit_replication":"https://pith.science/pith/5B7RV4TJQHTNC2EGAJENX3C4XO/action/replication_record"}},"created_at":"2026-07-09T01:20:24.121398+00:00","updated_at":"2026-07-09T01:20:24.121398+00:00"}