{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:SAJNS5FRJF5ND6WYNBKXIXP7NI","short_pith_number":"pith:SAJNS5FR","schema_version":"1.0","canonical_sha256":"9012d974b1497ad1fad86855745dff6a1614d94c0c8b16d25a643af780ad3128","source":{"kind":"arxiv","id":"2402.01030","version":4},"attestation_state":"computed","paper":{"title":"Executable Code Actions Elicit Better LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Hao Peng, Heng Ji, Lifan Yuan, Xingyao Wang, Yangyi Chen, Yizhe Zhang, Yunzhu Li","submitted_at":"2024-02-01T21:38:58Z","abstract_excerpt":"Large Language Model (LLM) agents, capable of performing a broad range of actions, such as invoking tools and controlling robots, show great potential in tackling real-world challenges. LLM agents are typically prompted to produce actions by generating JSON or text in a pre-defined format, which is usually limited by constrained action space (e.g., the scope of pre-defined tools) and restricted flexibility (e.g., inability to compose multiple tools). This work proposes to use executable Python code to consolidate LLM agents' actions into a unified action space (CodeAct). Integrated with a Pyth"},"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":"2402.01030","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-01T21:38:58Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"12ec3260d8776aedca9f7be90c61e59784df94c714383f23c2718ae973e1bb48","abstract_canon_sha256":"73d4d874db4de98d090d551002fa3a791bff8af6fd05bf58a36ee71f48adfcf8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:37.593494Z","signature_b64":"ugUcoBZmVtZXMcb75tzxcjTu7nAimJw979V7zySGxZJ1VqB4owkKlkg83uzwyqsdEe9kSF/8VswjiOlcuJpnCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9012d974b1497ad1fad86855745dff6a1614d94c0c8b16d25a643af780ad3128","last_reissued_at":"2026-07-05T08:28:37.593013Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:37.593013Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Executable Code Actions Elicit Better LLM Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Hao Peng, Heng Ji, Lifan Yuan, Xingyao Wang, Yangyi Chen, Yizhe Zhang, Yunzhu Li","submitted_at":"2024-02-01T21:38:58Z","abstract_excerpt":"Large Language Model (LLM) agents, capable of performing a broad range of actions, such as invoking tools and controlling robots, show great potential in tackling real-world challenges. LLM agents are typically prompted to produce actions by generating JSON or text in a pre-defined format, which is usually limited by constrained action space (e.g., the scope of pre-defined tools) and restricted flexibility (e.g., inability to compose multiple tools). This work proposes to use executable Python code to consolidate LLM agents' actions into a unified action space (CodeAct). Integrated with a Pyth"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.01030","kind":"arxiv","version":4},"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/2402.01030/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":"2402.01030","created_at":"2026-07-05T08:28:37.593071+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.01030v4","created_at":"2026-07-05T08:28:37.593071+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.01030","created_at":"2026-07-05T08:28:37.593071+00:00"},{"alias_kind":"pith_short_12","alias_value":"SAJNS5FRJF5N","created_at":"2026-07-05T08:28:37.593071+00:00"},{"alias_kind":"pith_short_16","alias_value":"SAJNS5FRJF5ND6WY","created_at":"2026-07-05T08:28:37.593071+00:00"},{"alias_kind":"pith_short_8","alias_value":"SAJNS5FR","created_at":"2026-07-05T08:28:37.593071+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":34,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26669","citing_title":"SKILL-DISCO: Distilling and Compiling Agent Traces into Reusable Procedural Skills","ref_index":58,"is_internal_anchor":false},{"citing_arxiv_id":"2606.17929","citing_title":"PreAct: Computer-Using Agents that Get Faster on Repeated Tasks","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2606.17454","citing_title":"Dissecting model behavior through agent trajectories","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2606.15874","citing_title":"LLM-as-Code: Agentic Programming for Agent Harness","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2606.13643","citing_title":"Recursive Agent Harnesses","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2606.08542","citing_title":"When Video Misreads: Closed-Loop Distillation of Reading Heuristics for Exploratory Manipulation Trace QA","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2606.05548","citing_title":"ADK Arena: Evaluating Agent Development Kits via LLM-as-a-Developer","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07358","citing_title":"A Comprehensive Survey on Agent Skills: Taxonomy, Techniques, and Applications","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28146","citing_title":"Cybersecurity AI (CAI) Dataset","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2605.29163","citing_title":"BCER Agent: Reliable Long-Horizon MRI Workflow Execution via Compilation, Artifact Binding, and Bounded Local Recovery","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27632","citing_title":"Yuvion LLM: An Adversarially-Aware Large Language Model for Content And AI Safety","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10106","citing_title":"What makes a harness a harness: necessary and sufficient conditions for an agent harness","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02688","citing_title":"MatClaw: An Autonomous Code-First LLM Agent for End-to-End Materials Exploration","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2503.21460","citing_title":"Large Language Model Agent: A Survey on Methodology, Applications and Challenges","ref_index":161,"is_internal_anchor":false},{"citing_arxiv_id":"2507.21035","citing_title":"GenoMAS: A Multi-Agent Framework for Scientific Discovery via Code-Driven Gene Expression Analysis","ref_index":123,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07358","citing_title":"A Comprehensive Survey on Agent Skills: Taxonomy, Techniques, and Applications","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09730","citing_title":"RubricRefine: Improving Tool-Use Agent Reliability with Training-Free Pre-Execution Refinement","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19677","citing_title":"Agentic Discovery of Cryomicroneedle Formulations","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2512.10371","citing_title":"AgentProg: Empowering Long-Horizon GUI Agents with Program-Guided Context Management","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09730","citing_title":"RubricRefine: Improving Tool-Use Agent Reliability with Training-Free Pre-Execution Refinement","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2602.20867","citing_title":"SoK: Agentic Skills -- Beyond Tool Use in LLM Agents","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2406.00515","citing_title":"A Survey on Large Language Models for Code Generation","ref_index":270,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02688","citing_title":"MatClaw: An Autonomous Code-First LLM Agent for End-to-End Materials Exploration","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08386","citing_title":"SkillLens: Adaptive Multi-Granularity Skill Reuse for Cost-Efficient LLM Agents","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09730","citing_title":"RubricRefine: Improving Tool-Use Agent Reliability with Training-Free Pre-Execution Refinement","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SAJNS5FRJF5ND6WYNBKXIXP7NI","json":"https://pith.science/pith/SAJNS5FRJF5ND6WYNBKXIXP7NI.json","graph_json":"https://pith.science/api/pith-number/SAJNS5FRJF5ND6WYNBKXIXP7NI/graph.json","events_json":"https://pith.science/api/pith-number/SAJNS5FRJF5ND6WYNBKXIXP7NI/events.json","paper":"https://pith.science/paper/SAJNS5FR"},"agent_actions":{"view_html":"https://pith.science/pith/SAJNS5FRJF5ND6WYNBKXIXP7NI","download_json":"https://pith.science/pith/SAJNS5FRJF5ND6WYNBKXIXP7NI.json","view_paper":"https://pith.science/paper/SAJNS5FR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.01030&json=true","fetch_graph":"https://pith.science/api/pith-number/SAJNS5FRJF5ND6WYNBKXIXP7NI/graph.json","fetch_events":"https://pith.science/api/pith-number/SAJNS5FRJF5ND6WYNBKXIXP7NI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SAJNS5FRJF5ND6WYNBKXIXP7NI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SAJNS5FRJF5ND6WYNBKXIXP7NI/action/storage_attestation","attest_author":"https://pith.science/pith/SAJNS5FRJF5ND6WYNBKXIXP7NI/action/author_attestation","sign_citation":"https://pith.science/pith/SAJNS5FRJF5ND6WYNBKXIXP7NI/action/citation_signature","submit_replication":"https://pith.science/pith/SAJNS5FRJF5ND6WYNBKXIXP7NI/action/replication_record"}},"created_at":"2026-07-05T08:28:37.593071+00:00","updated_at":"2026-07-05T08:28:37.593071+00:00"}