{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:6IUX27RTTXN5DZIOGFOPUM5Q2R","short_pith_number":"pith:6IUX27RT","schema_version":"1.0","canonical_sha256":"f2297d7e339ddbd1e50e315cfa33b0d465e8285a0f3c74508afec8b2b0443961","source":{"kind":"arxiv","id":"2606.08755","version":1},"attestation_state":"computed","paper":{"title":"Co-Evolving Skill Generation and Policy Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fenglong Ma, Linlin Wu, Nikki Lijing Kuang, Songtao Liu, Xiaomin Li, Yudi Lin, Zhiwei Zhang","submitted_at":"2026-06-07T17:55:55Z","abstract_excerpt":"Skill-augmented reinforcement learning improves language agents by storing reusable procedural knowledge acquired from past experience. Existing methods typically use strong language models to analyze trajectories, generate skills, and update a retrievable skill bank during online training. However, they rarely assess whether a newly generated skill is useful before it is stored and reused. We find that this assumption is unreliable: even skills generated by proprietary frontier LLMs exhibit highly mixed utility, with many providing little benefit or even degrading performance. Once such skill"},"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":"2606.08755","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-07T17:55:55Z","cross_cats_sorted":[],"title_canon_sha256":"f7ecb9f0116c26872483ef8eee1c3fb39fe8e32147827db5ad579262c5c5a4b1","abstract_canon_sha256":"79d14ac2386c2d2bf8ebb1aa497ac13a0188571b956fdc3dd4f332b710aa7604"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-09T02:07:37.565781Z","signature_b64":"2OEA9OcEfAdpJRGJp65Ejh4dD9NHEm5nd98g9KUUS37N5iiwTYbUpIRGJl+At/7XoJWp/H8M5F4pI1GKCAMaBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f2297d7e339ddbd1e50e315cfa33b0d465e8285a0f3c74508afec8b2b0443961","last_reissued_at":"2026-06-09T02:07:37.564910Z","signature_status":"signed_v1","first_computed_at":"2026-06-09T02:07:37.564910Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Co-Evolving Skill Generation and Policy Optimization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Fenglong Ma, Linlin Wu, Nikki Lijing Kuang, Songtao Liu, Xiaomin Li, Yudi Lin, Zhiwei Zhang","submitted_at":"2026-06-07T17:55:55Z","abstract_excerpt":"Skill-augmented reinforcement learning improves language agents by storing reusable procedural knowledge acquired from past experience. Existing methods typically use strong language models to analyze trajectories, generate skills, and update a retrievable skill bank during online training. However, they rarely assess whether a newly generated skill is useful before it is stored and reused. We find that this assumption is unreliable: even skills generated by proprietary frontier LLMs exhibit highly mixed utility, with many providing little benefit or even degrading performance. Once such skill"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.08755","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/2606.08755/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":"2606.08755","created_at":"2026-06-09T02:07:37.565058+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.08755v1","created_at":"2026-06-09T02:07:37.565058+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.08755","created_at":"2026-06-09T02:07:37.565058+00:00"},{"alias_kind":"pith_short_12","alias_value":"6IUX27RTTXN5","created_at":"2026-06-09T02:07:37.565058+00:00"},{"alias_kind":"pith_short_16","alias_value":"6IUX27RTTXN5DZIO","created_at":"2026-06-09T02:07:37.565058+00:00"},{"alias_kind":"pith_short_8","alias_value":"6IUX27RT","created_at":"2026-06-09T02:07:37.565058+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/6IUX27RTTXN5DZIOGFOPUM5Q2R","json":"https://pith.science/pith/6IUX27RTTXN5DZIOGFOPUM5Q2R.json","graph_json":"https://pith.science/api/pith-number/6IUX27RTTXN5DZIOGFOPUM5Q2R/graph.json","events_json":"https://pith.science/api/pith-number/6IUX27RTTXN5DZIOGFOPUM5Q2R/events.json","paper":"https://pith.science/paper/6IUX27RT"},"agent_actions":{"view_html":"https://pith.science/pith/6IUX27RTTXN5DZIOGFOPUM5Q2R","download_json":"https://pith.science/pith/6IUX27RTTXN5DZIOGFOPUM5Q2R.json","view_paper":"https://pith.science/paper/6IUX27RT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.08755&json=true","fetch_graph":"https://pith.science/api/pith-number/6IUX27RTTXN5DZIOGFOPUM5Q2R/graph.json","fetch_events":"https://pith.science/api/pith-number/6IUX27RTTXN5DZIOGFOPUM5Q2R/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6IUX27RTTXN5DZIOGFOPUM5Q2R/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6IUX27RTTXN5DZIOGFOPUM5Q2R/action/storage_attestation","attest_author":"https://pith.science/pith/6IUX27RTTXN5DZIOGFOPUM5Q2R/action/author_attestation","sign_citation":"https://pith.science/pith/6IUX27RTTXN5DZIOGFOPUM5Q2R/action/citation_signature","submit_replication":"https://pith.science/pith/6IUX27RTTXN5DZIOGFOPUM5Q2R/action/replication_record"}},"created_at":"2026-06-09T02:07:37.565058+00:00","updated_at":"2026-06-09T02:07:37.565058+00:00"}