{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:26NESNOYIHFOTLFTRGQ35U74AR","short_pith_number":"pith:26NESNOY","schema_version":"1.0","canonical_sha256":"d79a4935d841cae9acb389a1bed3fc047e1205a30925628da8c4cd8df10a6d67","source":{"kind":"arxiv","id":"2509.07315","version":1},"attestation_state":"computed","paper":{"title":"SafeToolBench: Pioneering a Prospective Benchmark to Evaluating Tool Utilization Safety in LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SE"],"primary_cat":"cs.CR","authors_text":"HaiFeng Wang, Hongfei Xia, Hongru Wang, Qian Yu, Yuhang Guo, Zeming Liu","submitted_at":"2025-09-09T01:31:25Z","abstract_excerpt":"Large Language Models (LLMs) have exhibited great performance in autonomously calling various tools in external environments, leading to better problem solving and task automation capabilities. However, these external tools also amplify potential risks such as financial loss or privacy leakage with ambiguous or malicious user instructions. Compared to previous studies, which mainly assess the safety awareness of LLMs after obtaining the tool execution results (i.e., retrospective evaluation), this paper focuses on prospective ways to assess the safety of LLM tool utilization, aiming to avoid i"},"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":"2509.07315","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2025-09-09T01:31:25Z","cross_cats_sorted":["cs.SE"],"title_canon_sha256":"8a4530834189173a189b050ea05b5914d89b3ae996b06140a3904f5ecf5cefba","abstract_canon_sha256":"a490b1c69874ef1c1657317b547c1262f1f5610fffc1953851bdac1fbe46f872"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:07:19.434901Z","signature_b64":"KexW/4sMpDLJTeCq5rrdB7kjuoxMn5dANwZZ5VB+AK901EmVhofuJNGgY6sZG2SZzEwMMN+A/Dy1z3dq34PRBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d79a4935d841cae9acb389a1bed3fc047e1205a30925628da8c4cd8df10a6d67","last_reissued_at":"2026-07-05T12:07:19.434351Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:07:19.434351Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SafeToolBench: Pioneering a Prospective Benchmark to Evaluating Tool Utilization Safety in LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.SE"],"primary_cat":"cs.CR","authors_text":"HaiFeng Wang, Hongfei Xia, Hongru Wang, Qian Yu, Yuhang Guo, Zeming Liu","submitted_at":"2025-09-09T01:31:25Z","abstract_excerpt":"Large Language Models (LLMs) have exhibited great performance in autonomously calling various tools in external environments, leading to better problem solving and task automation capabilities. However, these external tools also amplify potential risks such as financial loss or privacy leakage with ambiguous or malicious user instructions. Compared to previous studies, which mainly assess the safety awareness of LLMs after obtaining the tool execution results (i.e., retrospective evaluation), this paper focuses on prospective ways to assess the safety of LLM tool utilization, aiming to avoid i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.07315","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/2509.07315/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":"2509.07315","created_at":"2026-07-05T12:07:19.434429+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.07315v1","created_at":"2026-07-05T12:07:19.434429+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.07315","created_at":"2026-07-05T12:07:19.434429+00:00"},{"alias_kind":"pith_short_12","alias_value":"26NESNOYIHFO","created_at":"2026-07-05T12:07:19.434429+00:00"},{"alias_kind":"pith_short_16","alias_value":"26NESNOYIHFOTLFT","created_at":"2026-07-05T12:07:19.434429+00:00"},{"alias_kind":"pith_short_8","alias_value":"26NESNOY","created_at":"2026-07-05T12:07:19.434429+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.08531","citing_title":"VESTA: A Fully Automated Scenario Generation and Safety Evaluation Framework for LLM Agents","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24941","citing_title":"Memory-Induced Tool-Drift in LLM Agents","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16282","citing_title":"Taxonomy and Consistency Analysis of Safety Benchmarks for AI Agents","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02022","citing_title":"ATBench: A Diverse and Realistic Agent Trajectory Benchmark for Safety Evaluation and Diagnosis","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02022","citing_title":"ATBench: A Diverse and Realistic Agent Trajectory Benchmark for Safety Evaluation and Diagnosis","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04978","citing_title":"Measuring the Permission Gate: A Stress-Test Evaluation of Claude Code's Auto Mode","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/26NESNOYIHFOTLFTRGQ35U74AR","json":"https://pith.science/pith/26NESNOYIHFOTLFTRGQ35U74AR.json","graph_json":"https://pith.science/api/pith-number/26NESNOYIHFOTLFTRGQ35U74AR/graph.json","events_json":"https://pith.science/api/pith-number/26NESNOYIHFOTLFTRGQ35U74AR/events.json","paper":"https://pith.science/paper/26NESNOY"},"agent_actions":{"view_html":"https://pith.science/pith/26NESNOYIHFOTLFTRGQ35U74AR","download_json":"https://pith.science/pith/26NESNOYIHFOTLFTRGQ35U74AR.json","view_paper":"https://pith.science/paper/26NESNOY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.07315&json=true","fetch_graph":"https://pith.science/api/pith-number/26NESNOYIHFOTLFTRGQ35U74AR/graph.json","fetch_events":"https://pith.science/api/pith-number/26NESNOYIHFOTLFTRGQ35U74AR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/26NESNOYIHFOTLFTRGQ35U74AR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/26NESNOYIHFOTLFTRGQ35U74AR/action/storage_attestation","attest_author":"https://pith.science/pith/26NESNOYIHFOTLFTRGQ35U74AR/action/author_attestation","sign_citation":"https://pith.science/pith/26NESNOYIHFOTLFTRGQ35U74AR/action/citation_signature","submit_replication":"https://pith.science/pith/26NESNOYIHFOTLFTRGQ35U74AR/action/replication_record"}},"created_at":"2026-07-05T12:07:19.434429+00:00","updated_at":"2026-07-05T12:07:19.434429+00:00"}