{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:MYRJS5WUJEITROGSNXDZF6MIBN","short_pith_number":"pith:MYRJS5WU","schema_version":"1.0","canonical_sha256":"66229976d4491138b8d26dc792f9880b7b2f1995b035d49a8028c5884ffbc060","source":{"kind":"arxiv","id":"2606.01815","version":1},"attestation_state":"computed","paper":{"title":"CRAB-Bench: Evaluating LLM Agents under Complex Task Dependencies and Human-aligned User Simulation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Akshay Sivaraman, Danqing Wang, Lei Li","submitted_at":"2026-06-01T07:31:04Z","abstract_excerpt":"Evaluating LLM agents in realistic service scenarios requires complex task dependencies, imperfect user behavior, and an evaluation that accommodates multiple valid solutions. We introduce CRAB-Bench (Constraint-based Realistic Agent Benchmark) and RUSE (Realistic User Simulation Engine) to address this gap. CRAB-Bench generates tasks via a constraint graph over multiple interdependent entities with structured distractors, requiring agents to reason carefully over thousands of misleading candidates where only a tiny fraction of solutions are valid. RUSE replaces cooperative, template-like simu"},"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":"2606.01815","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-01T07:31:04Z","cross_cats_sorted":[],"title_canon_sha256":"3128847a9766e66252fde59dc3811dac7269c0ef4fdc4d97bf789581de4b711c","abstract_canon_sha256":"29e3c5da16ad9e219cc438271f210717bfbc689554c473914ee0a0655a9e885a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-02T02:04:57.619291Z","signature_b64":"l6pcuFYzO1nWt8oaoV4J1Z0VkuWEThqD8xgvG+Bb+k658YQ9fqSRLLasKTtWkUjZ3PUJ92dPsDlvKLDT7tZ2Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"66229976d4491138b8d26dc792f9880b7b2f1995b035d49a8028c5884ffbc060","last_reissued_at":"2026-06-02T02:04:57.618835Z","signature_status":"signed_v1","first_computed_at":"2026-06-02T02:04:57.618835Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CRAB-Bench: Evaluating LLM Agents under Complex Task Dependencies and Human-aligned User Simulation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Akshay Sivaraman, Danqing Wang, Lei Li","submitted_at":"2026-06-01T07:31:04Z","abstract_excerpt":"Evaluating LLM agents in realistic service scenarios requires complex task dependencies, imperfect user behavior, and an evaluation that accommodates multiple valid solutions. We introduce CRAB-Bench (Constraint-based Realistic Agent Benchmark) and RUSE (Realistic User Simulation Engine) to address this gap. CRAB-Bench generates tasks via a constraint graph over multiple interdependent entities with structured distractors, requiring agents to reason carefully over thousands of misleading candidates where only a tiny fraction of solutions are valid. RUSE replaces cooperative, template-like simu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.01815","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/2606.01815/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":"2606.01815","created_at":"2026-06-02T02:04:57.618902+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.01815v1","created_at":"2026-06-02T02:04:57.618902+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.01815","created_at":"2026-06-02T02:04:57.618902+00:00"},{"alias_kind":"pith_short_12","alias_value":"MYRJS5WUJEIT","created_at":"2026-06-02T02:04:57.618902+00:00"},{"alias_kind":"pith_short_16","alias_value":"MYRJS5WUJEITROGS","created_at":"2026-06-02T02:04:57.618902+00:00"},{"alias_kind":"pith_short_8","alias_value":"MYRJS5WU","created_at":"2026-06-02T02:04:57.618902+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/MYRJS5WUJEITROGSNXDZF6MIBN","json":"https://pith.science/pith/MYRJS5WUJEITROGSNXDZF6MIBN.json","graph_json":"https://pith.science/api/pith-number/MYRJS5WUJEITROGSNXDZF6MIBN/graph.json","events_json":"https://pith.science/api/pith-number/MYRJS5WUJEITROGSNXDZF6MIBN/events.json","paper":"https://pith.science/paper/MYRJS5WU"},"agent_actions":{"view_html":"https://pith.science/pith/MYRJS5WUJEITROGSNXDZF6MIBN","download_json":"https://pith.science/pith/MYRJS5WUJEITROGSNXDZF6MIBN.json","view_paper":"https://pith.science/paper/MYRJS5WU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.01815&json=true","fetch_graph":"https://pith.science/api/pith-number/MYRJS5WUJEITROGSNXDZF6MIBN/graph.json","fetch_events":"https://pith.science/api/pith-number/MYRJS5WUJEITROGSNXDZF6MIBN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MYRJS5WUJEITROGSNXDZF6MIBN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MYRJS5WUJEITROGSNXDZF6MIBN/action/storage_attestation","attest_author":"https://pith.science/pith/MYRJS5WUJEITROGSNXDZF6MIBN/action/author_attestation","sign_citation":"https://pith.science/pith/MYRJS5WUJEITROGSNXDZF6MIBN/action/citation_signature","submit_replication":"https://pith.science/pith/MYRJS5WUJEITROGSNXDZF6MIBN/action/replication_record"}},"created_at":"2026-06-02T02:04:57.618902+00:00","updated_at":"2026-06-02T02:04:57.618902+00:00"}