{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:D3EHK3FAKT2IFQVUJBTQJC45WK","short_pith_number":"pith:D3EHK3FA","canonical_record":{"source":{"id":"2602.01664","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-02T05:30:42Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"0817d5e021e1bf3cb926aaf87f4e894860bc8b29984b687d85ea6731d61e69e8","abstract_canon_sha256":"5ebd31a8260a59d823164b2d97d01f3f54e8ac6edbd62761b4d7815a797a9872"},"schema_version":"1.0"},"canonical_sha256":"1ec8756ca054f482c2b44867048b9db2b3ff2c6935893b22ff762378c381dacc","source":{"kind":"arxiv","id":"2602.01664","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.01664","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"arxiv_version","alias_value":"2602.01664v4","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.01664","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"pith_short_12","alias_value":"D3EHK3FAKT2I","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"D3EHK3FAKT2IFQVU","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"D3EHK3FA","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:D3EHK3FAKT2IFQVUJBTQJC45WK","target":"record","payload":{"canonical_record":{"source":{"id":"2602.01664","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-02T05:30:42Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"0817d5e021e1bf3cb926aaf87f4e894860bc8b29984b687d85ea6731d61e69e8","abstract_canon_sha256":"5ebd31a8260a59d823164b2d97d01f3f54e8ac6edbd62761b4d7815a797a9872"},"schema_version":"1.0"},"canonical_sha256":"1ec8756ca054f482c2b44867048b9db2b3ff2c6935893b22ff762378c381dacc","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:16.457464Z","signature_b64":"R3TUlFeydp9u6o/AQZmlTY/3Tfj4j4hNCjDJdPHiLw+AqlOsrYJwd469kBLQapp+D6Tmc8C2VNRDnOSINahMDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1ec8756ca054f482c2b44867048b9db2b3ff2c6935893b22ff762378c381dacc","last_reissued_at":"2026-05-17T23:39:16.456876Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:16.456876Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2602.01664","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:39:16Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NlPpU4khaz1YfMk/FcjGNmDCunH6B4R9+28Ohqdl37ZZLMbTc1MJv8rzqdIGOgpnsdX+ZY8ARGbgmPUkK8blCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-05T17:50:29.821215Z"},"content_sha256":"3ae8e5aa2aff419099c67441bd6fca52705ccd0f9a0fded4151155ceab6f3ee2","schema_version":"1.0","event_id":"sha256:3ae8e5aa2aff419099c67441bd6fca52705ccd0f9a0fded4151155ceab6f3ee2"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:D3EHK3FAKT2IFQVUJBTQJC45WK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"FlowSteer: Towards Agents Designing Agentic Workflows via Reinforced Progressive Canvas Editing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"A single agent can design complete agentic workflows end-to-end by making sequential edits to an executable canvas that supplies real-time syntax-checked feedback.","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Erik Cambria, Haoran Luo, Mingda Zhang, Qika Lin, Rui Mao, Tiesunlong Shen, Wenjin Liu, Xiaoying Tang","submitted_at":"2026-02-02T05:30:42Z","abstract_excerpt":"In recent years, agentic workflows have been widely applied to solve complex human tasks. However, existing workflow construction still faces key challenges, including human-dependent workflow construction, the lack of graph-level execution feedback, and the inability to repair errors in-loop during long-horizon construction. To address these challenges, we propose FlowSteer, a new paradigm of Agent Designing Agentic Workflows - a single agent itself end-to-end designs the workflow that a downstream executor runs. To support this paradigm, we introduce the Workflow Canvas, a novel executable g"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experimental results on twelve datasets show that FlowSteer significantly outperforms baselines across various tasks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That real-time syntax-checked execution feedback from the Workflow Canvas is sufficient to train a policy agent that can reliably repair errors during long-horizon workflow construction without external human guidance.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A reinforcement learning policy agent designs executable agentic workflows by issuing atomic edits to a feedback-providing Workflow Canvas environment.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A single agent can design complete agentic workflows end-to-end by making sequential edits to an executable canvas that supplies real-time syntax-checked feedback.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"2b56aa3efe74f45787fd0c5caf250bc3e668b37b905735c3a90ac8a8c94ba93c"},"source":{"id":"2602.01664","kind":"arxiv","version":4},"verdict":{"id":"cf8b4ae0-4740-41ec-b9cc-09e4164b647b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T08:45:54.061811Z","strongest_claim":"Experimental results on twelve datasets show that FlowSteer significantly outperforms baselines across various tasks.","one_line_summary":"A reinforcement learning policy agent designs executable agentic workflows by issuing atomic edits to a feedback-providing Workflow Canvas environment.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That real-time syntax-checked execution feedback from the Workflow Canvas is sufficient to train a policy agent that can reliably repair errors during long-horizon workflow construction without external human guidance.","pith_extraction_headline":"A single agent can design complete agentic workflows end-to-end by making sequential edits to an executable canvas that supplies real-time syntax-checked feedback."},"references":{"count":21,"sample":[{"doi":"","year":null,"title":"Greedy always picks largest coin≤N, which is locally optimal but not always globally optimal","work_id":"6358b90a-6834-4909-b15d-fe15e73c2df4","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Greedy fails when using fewer large coins plus more medium coins yields fewer total coins (e.g., N=30: greedy=25+5×1=6 coins, optimal=3×10=3 coins)","work_id":"8a4dbd32-6020-4b30-bf94-22ec19bb3fa1","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Dynamic programming guarantees finding the true minimum coin count","work_id":"ca908b1c-1bbd-4c05-abaa-571c89dadfa0","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"The coin system {1,10,25} is NOT canonical (unlike {1,5,10,25} US coins), so greedy can fail. Plan:","work_id":"83aaf664-afdc-4282-ab47-6414d4268a43","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Implementgreedy_coins(n)that iteratively subtracts largest possible coin","work_id":"b8e9206a-1666-40e7-8b90-b93558734947","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":21,"snapshot_sha256":"d2b3a863eeefa54eefea25d50c3acddc941a82b3c63456488e6c6641ff07bf5a","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"},"verdict_id":"cf8b4ae0-4740-41ec-b9cc-09e4164b647b"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-17T23:39:16Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"sxD+NYJTQgGAUmyl7XGAAtAAMzF7xONP4s7Mnjdn2iJA0ZBBe8pLRcJMteSf4mbj6PgELo5Z+JGfAKJrGBlnCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-05T17:50:29.821867Z"},"content_sha256":"425159f78a02881f02620eabdc9bb9e603b43db0059b0f5d95ccaeabbec3d633","schema_version":"1.0","event_id":"sha256:425159f78a02881f02620eabdc9bb9e603b43db0059b0f5d95ccaeabbec3d633"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/D3EHK3FAKT2IFQVUJBTQJC45WK/bundle.json","state_url":"https://pith.science/pith/D3EHK3FAKT2IFQVUJBTQJC45WK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/D3EHK3FAKT2IFQVUJBTQJC45WK/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-06-05T17:50:29Z","links":{"resolver":"https://pith.science/pith/D3EHK3FAKT2IFQVUJBTQJC45WK","bundle":"https://pith.science/pith/D3EHK3FAKT2IFQVUJBTQJC45WK/bundle.json","state":"https://pith.science/pith/D3EHK3FAKT2IFQVUJBTQJC45WK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/D3EHK3FAKT2IFQVUJBTQJC45WK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:D3EHK3FAKT2IFQVUJBTQJC45WK","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"5ebd31a8260a59d823164b2d97d01f3f54e8ac6edbd62761b4d7815a797a9872","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-02T05:30:42Z","title_canon_sha256":"0817d5e021e1bf3cb926aaf87f4e894860bc8b29984b687d85ea6731d61e69e8"},"schema_version":"1.0","source":{"id":"2602.01664","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.01664","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"arxiv_version","alias_value":"2602.01664v4","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.01664","created_at":"2026-05-17T23:39:16Z"},{"alias_kind":"pith_short_12","alias_value":"D3EHK3FAKT2I","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"D3EHK3FAKT2IFQVU","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"D3EHK3FA","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:425159f78a02881f02620eabdc9bb9e603b43db0059b0f5d95ccaeabbec3d633","target":"graph","created_at":"2026-05-17T23:39:16Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Experimental results on twelve datasets show that FlowSteer significantly outperforms baselines across various tasks."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That real-time syntax-checked execution feedback from the Workflow Canvas is sufficient to train a policy agent that can reliably repair errors during long-horizon workflow construction without external human guidance."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"A reinforcement learning policy agent designs executable agentic workflows by issuing atomic edits to a feedback-providing Workflow Canvas environment."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A single agent can design complete agentic workflows end-to-end by making sequential edits to an executable canvas that supplies real-time syntax-checked feedback."}],"snapshot_sha256":"2b56aa3efe74f45787fd0c5caf250bc3e668b37b905735c3a90ac8a8c94ba93c"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"In recent years, agentic workflows have been widely applied to solve complex human tasks. However, existing workflow construction still faces key challenges, including human-dependent workflow construction, the lack of graph-level execution feedback, and the inability to repair errors in-loop during long-horizon construction. To address these challenges, we propose FlowSteer, a new paradigm of Agent Designing Agentic Workflows - a single agent itself end-to-end designs the workflow that a downstream executor runs. To support this paradigm, we introduce the Workflow Canvas, a novel executable g","authors_text":"Erik Cambria, Haoran Luo, Mingda Zhang, Qika Lin, Rui Mao, Tiesunlong Shen, Wenjin Liu, Xiaoying Tang","cross_cats":["cs.LG"],"headline":"A single agent can design complete agentic workflows end-to-end by making sequential edits to an executable canvas that supplies real-time syntax-checked feedback.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-02T05:30:42Z","title":"FlowSteer: Towards Agents Designing Agentic Workflows via Reinforced Progressive Canvas Editing"},"references":{"count":21,"internal_anchors":0,"resolved_work":21,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Greedy always picks largest coin≤N, which is locally optimal but not always globally optimal","work_id":"6358b90a-6834-4909-b15d-fe15e73c2df4","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Greedy fails when using fewer large coins plus more medium coins yields fewer total coins (e.g., N=30: greedy=25+5×1=6 coins, optimal=3×10=3 coins)","work_id":"8a4dbd32-6020-4b30-bf94-22ec19bb3fa1","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Dynamic programming guarantees finding the true minimum coin count","work_id":"ca908b1c-1bbd-4c05-abaa-571c89dadfa0","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"The coin system {1,10,25} is NOT canonical (unlike {1,5,10,25} US coins), so greedy can fail. Plan:","work_id":"83aaf664-afdc-4282-ab47-6414d4268a43","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Implementgreedy_coins(n)that iteratively subtracts largest possible coin","work_id":"b8e9206a-1666-40e7-8b90-b93558734947","year":null}],"snapshot_sha256":"d2b3a863eeefa54eefea25d50c3acddc941a82b3c63456488e6c6641ff07bf5a"},"source":{"id":"2602.01664","kind":"arxiv","version":4},"verdict":{"created_at":"2026-05-16T08:45:54.061811Z","id":"cf8b4ae0-4740-41ec-b9cc-09e4164b647b","model_set":{"reader":"grok-4.3"},"one_line_summary":"A reinforcement learning policy agent designs executable agentic workflows by issuing atomic edits to a feedback-providing Workflow Canvas environment.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A single agent can design complete agentic workflows end-to-end by making sequential edits to an executable canvas that supplies real-time syntax-checked feedback.","strongest_claim":"Experimental results on twelve datasets show that FlowSteer significantly outperforms baselines across various tasks.","weakest_assumption":"That real-time syntax-checked execution feedback from the Workflow Canvas is sufficient to train a policy agent that can reliably repair errors during long-horizon workflow construction without external human guidance."}},"verdict_id":"cf8b4ae0-4740-41ec-b9cc-09e4164b647b"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:3ae8e5aa2aff419099c67441bd6fca52705ccd0f9a0fded4151155ceab6f3ee2","target":"record","created_at":"2026-05-17T23:39:16Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"5ebd31a8260a59d823164b2d97d01f3f54e8ac6edbd62761b4d7815a797a9872","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-02-02T05:30:42Z","title_canon_sha256":"0817d5e021e1bf3cb926aaf87f4e894860bc8b29984b687d85ea6731d61e69e8"},"schema_version":"1.0","source":{"id":"2602.01664","kind":"arxiv","version":4}},"canonical_sha256":"1ec8756ca054f482c2b44867048b9db2b3ff2c6935893b22ff762378c381dacc","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1ec8756ca054f482c2b44867048b9db2b3ff2c6935893b22ff762378c381dacc","first_computed_at":"2026-05-17T23:39:16.456876Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:16.456876Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"R3TUlFeydp9u6o/AQZmlTY/3Tfj4j4hNCjDJdPHiLw+AqlOsrYJwd469kBLQapp+D6Tmc8C2VNRDnOSINahMDg==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:16.457464Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.01664","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3ae8e5aa2aff419099c67441bd6fca52705ccd0f9a0fded4151155ceab6f3ee2","sha256:425159f78a02881f02620eabdc9bb9e603b43db0059b0f5d95ccaeabbec3d633"],"state_sha256":"b158a3c15c813a98b8479ba62eae04906e2a565dead73a305a098496390a08a9"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3tedaH2c85GAk/XOIhI92bh1hlDFtww9RR2h6NhW4c4Q9kLKjGUfoaNuwOo4S4UQlwI2uAUq/svryxp9ttZVCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-05T17:50:29.824733Z","bundle_sha256":"1c80b7c94e04afe3b1688aa9cddf941d3552073f9bd1beea52a18aeb213ae643"}}