{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YV7G722P4I5GOX7G6V6SL622ZL","short_pith_number":"pith:YV7G722P","schema_version":"1.0","canonical_sha256":"c57e6feb4fe23a675fe6f57d25fb5acaf7b018aa1444e33fda8b6a25d4521642","source":{"kind":"arxiv","id":"2504.02623","version":3},"attestation_state":"computed","paper":{"title":"Multi-Mission Tool Bench: Assessing the Robustness of LLM based Agents through Related and Dynamic Missions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Feng Zhang, Haorui Wang, Jinjian Li, Peijie Yu, Xiao Feng, Yifan Yang, Zelong Zhang","submitted_at":"2025-04-03T14:21:33Z","abstract_excerpt":"Large language models (LLMs) demonstrate strong potential as agents for tool invocation due to their advanced comprehension and planning capabilities. Users increasingly rely on LLM-based agents to solve complex missions through iterative interactions. However, existing benchmarks predominantly access agents in single-mission scenarios, failing to capture real-world complexity. To bridge this gap, we propose the Multi-Mission Tool Bench. In the benchmark, each test case comprises multiple interrelated missions. This design requires agents to dynamically adapt to evolving demands. Moreover, the"},"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":"2504.02623","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-04-03T14:21:33Z","cross_cats_sorted":[],"title_canon_sha256":"2b23e94220afe263eedd8fd666439a830bbd5de9a3b3dcb98801570016018462","abstract_canon_sha256":"83c7fdd7fb192ed4de5ff6c4700f3c35a5e5a0802110f1760e5609302563217b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:49:52.774571Z","signature_b64":"mQslRD8q9K7mXGA4M2mGv2r/qq/fF0LxiRZM/kIh58UPFJBFT7/t23HHKj/BUfMdfx0hjvGJby5IC666q+iaAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c57e6feb4fe23a675fe6f57d25fb5acaf7b018aa1444e33fda8b6a25d4521642","last_reissued_at":"2026-07-05T10:49:52.774077Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:49:52.774077Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multi-Mission Tool Bench: Assessing the Robustness of LLM based Agents through Related and Dynamic Missions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Feng Zhang, Haorui Wang, Jinjian Li, Peijie Yu, Xiao Feng, Yifan Yang, Zelong Zhang","submitted_at":"2025-04-03T14:21:33Z","abstract_excerpt":"Large language models (LLMs) demonstrate strong potential as agents for tool invocation due to their advanced comprehension and planning capabilities. Users increasingly rely on LLM-based agents to solve complex missions through iterative interactions. However, existing benchmarks predominantly access agents in single-mission scenarios, failing to capture real-world complexity. To bridge this gap, we propose the Multi-Mission Tool Bench. In the benchmark, each test case comprises multiple interrelated missions. This design requires agents to dynamically adapt to evolving demands. Moreover, the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.02623","kind":"arxiv","version":3},"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/2504.02623/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":"2504.02623","created_at":"2026-07-05T10:49:52.774134+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.02623v3","created_at":"2026-07-05T10:49:52.774134+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.02623","created_at":"2026-07-05T10:49:52.774134+00:00"},{"alias_kind":"pith_short_12","alias_value":"YV7G722P4I5G","created_at":"2026-07-05T10:49:52.774134+00:00"},{"alias_kind":"pith_short_16","alias_value":"YV7G722P4I5GOX7G","created_at":"2026-07-05T10:49:52.774134+00:00"},{"alias_kind":"pith_short_8","alias_value":"YV7G722P","created_at":"2026-07-05T10:49:52.774134+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24428","citing_title":"Escaping the Self-Confirmation Trap: An Execute-Distill-Verify Paradigm for Agentic Experience Learning","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YV7G722P4I5GOX7G6V6SL622ZL","json":"https://pith.science/pith/YV7G722P4I5GOX7G6V6SL622ZL.json","graph_json":"https://pith.science/api/pith-number/YV7G722P4I5GOX7G6V6SL622ZL/graph.json","events_json":"https://pith.science/api/pith-number/YV7G722P4I5GOX7G6V6SL622ZL/events.json","paper":"https://pith.science/paper/YV7G722P"},"agent_actions":{"view_html":"https://pith.science/pith/YV7G722P4I5GOX7G6V6SL622ZL","download_json":"https://pith.science/pith/YV7G722P4I5GOX7G6V6SL622ZL.json","view_paper":"https://pith.science/paper/YV7G722P","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.02623&json=true","fetch_graph":"https://pith.science/api/pith-number/YV7G722P4I5GOX7G6V6SL622ZL/graph.json","fetch_events":"https://pith.science/api/pith-number/YV7G722P4I5GOX7G6V6SL622ZL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YV7G722P4I5GOX7G6V6SL622ZL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YV7G722P4I5GOX7G6V6SL622ZL/action/storage_attestation","attest_author":"https://pith.science/pith/YV7G722P4I5GOX7G6V6SL622ZL/action/author_attestation","sign_citation":"https://pith.science/pith/YV7G722P4I5GOX7G6V6SL622ZL/action/citation_signature","submit_replication":"https://pith.science/pith/YV7G722P4I5GOX7G6V6SL622ZL/action/replication_record"}},"created_at":"2026-07-05T10:49:52.774134+00:00","updated_at":"2026-07-05T10:49:52.774134+00:00"}