{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ST5NVMWLMDGVYLT6WEID55T3A7","short_pith_number":"pith:ST5NVMWL","schema_version":"1.0","canonical_sha256":"94fadab2cb60cd5c2e7eb1103ef67b07db66c0e5a664ba1ae6fbb353e5b86ed2","source":{"kind":"arxiv","id":"2505.22967","version":1},"attestation_state":"computed","paper":{"title":"MermaidFlow: Redefining Agentic Workflow Generation via Safety-Constrained Evolutionary Programming","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.MA"],"primary_cat":"cs.LG","authors_text":"Chengqi Zheng, Haiyan Yin, HaoPeng Zhang, Ivor Tsang, Jianda Chen, Wen Zheng Terence Ng, Yew-Soon Ong, Yueming Lyu","submitted_at":"2025-05-29T01:08:36Z","abstract_excerpt":"Despite the promise of autonomous agentic reasoning, existing workflow generation methods frequently produce fragile, unexecutable plans due to unconstrained LLM-driven construction. We introduce MermaidFlow, a framework that redefines the agentic search space through safety-constrained graph evolution. At its core, MermaidFlow represent workflows as a verifiable intermediate representation using Mermaid, a structured and human-interpretable graph language. We formulate domain-aware evolutionary operators, i.e., crossover, mutation, insertion, and deletion, to preserve semantic correctness whi"},"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":"2505.22967","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-29T01:08:36Z","cross_cats_sorted":["cs.MA"],"title_canon_sha256":"64dfec7b1df1a99b3f1711797d9bb83642500c584ab6c8f4913a7bd64fa1a15e","abstract_canon_sha256":"a4ad09db81766a8266405767c4cddde1ceef39b1cc93408fb70229d40ed67c36"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:11:52.903151Z","signature_b64":"+wsXUuxwIe1YjRKucF4pFMkjWSefE551Z9gAS9vW47H5zDrtSEArFfg1AduSsFTORAHaQNchvThHYRhnWOnEDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"94fadab2cb60cd5c2e7eb1103ef67b07db66c0e5a664ba1ae6fbb353e5b86ed2","last_reissued_at":"2026-07-05T11:11:52.902645Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:11:52.902645Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MermaidFlow: Redefining Agentic Workflow Generation via Safety-Constrained Evolutionary Programming","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.MA"],"primary_cat":"cs.LG","authors_text":"Chengqi Zheng, Haiyan Yin, HaoPeng Zhang, Ivor Tsang, Jianda Chen, Wen Zheng Terence Ng, Yew-Soon Ong, Yueming Lyu","submitted_at":"2025-05-29T01:08:36Z","abstract_excerpt":"Despite the promise of autonomous agentic reasoning, existing workflow generation methods frequently produce fragile, unexecutable plans due to unconstrained LLM-driven construction. We introduce MermaidFlow, a framework that redefines the agentic search space through safety-constrained graph evolution. At its core, MermaidFlow represent workflows as a verifiable intermediate representation using Mermaid, a structured and human-interpretable graph language. We formulate domain-aware evolutionary operators, i.e., crossover, mutation, insertion, and deletion, to preserve semantic correctness whi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.22967","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/2505.22967/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":"2505.22967","created_at":"2026-07-05T11:11:52.902707+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.22967v1","created_at":"2026-07-05T11:11:52.902707+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.22967","created_at":"2026-07-05T11:11:52.902707+00:00"},{"alias_kind":"pith_short_12","alias_value":"ST5NVMWLMDGV","created_at":"2026-07-05T11:11:52.902707+00:00"},{"alias_kind":"pith_short_16","alias_value":"ST5NVMWLMDGVYLT6","created_at":"2026-07-05T11:11:52.902707+00:00"},{"alias_kind":"pith_short_8","alias_value":"ST5NVMWL","created_at":"2026-07-05T11:11:52.902707+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22385","citing_title":"MetaPS: Adaptive Programmatic Strategy Selection for Market Agents","ref_index":136,"is_internal_anchor":false},{"citing_arxiv_id":"2508.07407","citing_title":"A Comprehensive Survey of Self-Evolving AI Agents: A New Paradigm Bridging Foundation Models and Lifelong Agentic Systems","ref_index":122,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01293","citing_title":"Lifting Traces to Logic: Programmatic Skill Induction with Neuro-Symbolic Learning for Long-Horizon Agentic Tasks","ref_index":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ST5NVMWLMDGVYLT6WEID55T3A7","json":"https://pith.science/pith/ST5NVMWLMDGVYLT6WEID55T3A7.json","graph_json":"https://pith.science/api/pith-number/ST5NVMWLMDGVYLT6WEID55T3A7/graph.json","events_json":"https://pith.science/api/pith-number/ST5NVMWLMDGVYLT6WEID55T3A7/events.json","paper":"https://pith.science/paper/ST5NVMWL"},"agent_actions":{"view_html":"https://pith.science/pith/ST5NVMWLMDGVYLT6WEID55T3A7","download_json":"https://pith.science/pith/ST5NVMWLMDGVYLT6WEID55T3A7.json","view_paper":"https://pith.science/paper/ST5NVMWL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.22967&json=true","fetch_graph":"https://pith.science/api/pith-number/ST5NVMWLMDGVYLT6WEID55T3A7/graph.json","fetch_events":"https://pith.science/api/pith-number/ST5NVMWLMDGVYLT6WEID55T3A7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ST5NVMWLMDGVYLT6WEID55T3A7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ST5NVMWLMDGVYLT6WEID55T3A7/action/storage_attestation","attest_author":"https://pith.science/pith/ST5NVMWLMDGVYLT6WEID55T3A7/action/author_attestation","sign_citation":"https://pith.science/pith/ST5NVMWLMDGVYLT6WEID55T3A7/action/citation_signature","submit_replication":"https://pith.science/pith/ST5NVMWLMDGVYLT6WEID55T3A7/action/replication_record"}},"created_at":"2026-07-05T11:11:52.902707+00:00","updated_at":"2026-07-05T11:11:52.902707+00:00"}