{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:LOFD2KDGRM5K3NOWRQU2OTLPO6","short_pith_number":"pith:LOFD2KDG","schema_version":"1.0","canonical_sha256":"5b8a3d28668b3aadb5d68c29a74d6f77ad6e0b8891def81e5fa25de87baea3db","source":{"kind":"arxiv","id":"2606.28061","version":1},"attestation_state":"computed","paper":{"title":"ToolPrivacyBench: Benchmarking Purpose-Bound Privacy in Tool-Using LLM Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Liang Liu, Shijing Hu, Zhicheng Zhao, Zhu Meng","submitted_at":"2026-06-26T13:08:46Z","abstract_excerpt":"Large language models (LLMs) have increasingly moved from standalone text generation systems to agents that invoke external tools, access environments, and execute multi-step tasks. However, conventional function-calling benchmarks mainly evaluate task completion and API correctness, while privacy evaluation benchmarks typically focus on final responses or privacy judgments. Neither perspective captures purpose-bound information flow across an executed multi-tool trajectory. Motivated by this limitation in current agent evaluation, ToolPrivacyBench audits whether task-private atoms are routed "},"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":"2606.28061","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2026-06-26T13:08:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"253fc6f5faccb1dcbae786c3eaa02ded1e8ec4612bb6f76c56405806d4468152","abstract_canon_sha256":"4c79850a4db12770252c0f483b851fd78e30115137203994fa66fd72dc583fef"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-29T01:14:56.607267Z","signature_b64":"ZjOU20CbYceY1EysgpBcJTckvgwritiRGF5RuFSxopjSZNZm+WirAGADhJdz6I9LYb5dCqm5b0jTwG0YtDntAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5b8a3d28668b3aadb5d68c29a74d6f77ad6e0b8891def81e5fa25de87baea3db","last_reissued_at":"2026-06-29T01:14:56.606874Z","signature_status":"signed_v1","first_computed_at":"2026-06-29T01:14:56.606874Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ToolPrivacyBench: Benchmarking Purpose-Bound Privacy in Tool-Using LLM Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Liang Liu, Shijing Hu, Zhicheng Zhao, Zhu Meng","submitted_at":"2026-06-26T13:08:46Z","abstract_excerpt":"Large language models (LLMs) have increasingly moved from standalone text generation systems to agents that invoke external tools, access environments, and execute multi-step tasks. However, conventional function-calling benchmarks mainly evaluate task completion and API correctness, while privacy evaluation benchmarks typically focus on final responses or privacy judgments. Neither perspective captures purpose-bound information flow across an executed multi-tool trajectory. Motivated by this limitation in current agent evaluation, ToolPrivacyBench audits whether task-private atoms are routed "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.28061","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/2606.28061/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":"2606.28061","created_at":"2026-06-29T01:14:56.606932+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.28061v1","created_at":"2026-06-29T01:14:56.606932+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.28061","created_at":"2026-06-29T01:14:56.606932+00:00"},{"alias_kind":"pith_short_12","alias_value":"LOFD2KDGRM5K","created_at":"2026-06-29T01:14:56.606932+00:00"},{"alias_kind":"pith_short_16","alias_value":"LOFD2KDGRM5K3NOW","created_at":"2026-06-29T01:14:56.606932+00:00"},{"alias_kind":"pith_short_8","alias_value":"LOFD2KDG","created_at":"2026-06-29T01:14:56.606932+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/LOFD2KDGRM5K3NOWRQU2OTLPO6","json":"https://pith.science/pith/LOFD2KDGRM5K3NOWRQU2OTLPO6.json","graph_json":"https://pith.science/api/pith-number/LOFD2KDGRM5K3NOWRQU2OTLPO6/graph.json","events_json":"https://pith.science/api/pith-number/LOFD2KDGRM5K3NOWRQU2OTLPO6/events.json","paper":"https://pith.science/paper/LOFD2KDG"},"agent_actions":{"view_html":"https://pith.science/pith/LOFD2KDGRM5K3NOWRQU2OTLPO6","download_json":"https://pith.science/pith/LOFD2KDGRM5K3NOWRQU2OTLPO6.json","view_paper":"https://pith.science/paper/LOFD2KDG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.28061&json=true","fetch_graph":"https://pith.science/api/pith-number/LOFD2KDGRM5K3NOWRQU2OTLPO6/graph.json","fetch_events":"https://pith.science/api/pith-number/LOFD2KDGRM5K3NOWRQU2OTLPO6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LOFD2KDGRM5K3NOWRQU2OTLPO6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LOFD2KDGRM5K3NOWRQU2OTLPO6/action/storage_attestation","attest_author":"https://pith.science/pith/LOFD2KDGRM5K3NOWRQU2OTLPO6/action/author_attestation","sign_citation":"https://pith.science/pith/LOFD2KDGRM5K3NOWRQU2OTLPO6/action/citation_signature","submit_replication":"https://pith.science/pith/LOFD2KDGRM5K3NOWRQU2OTLPO6/action/replication_record"}},"created_at":"2026-06-29T01:14:56.606932+00:00","updated_at":"2026-06-29T01:14:56.606932+00:00"}