{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:DQDNX7CUP23JS6RDPP3VTDZ5N3","short_pith_number":"pith:DQDNX7CU","canonical_record":{"source":{"id":"2604.00392","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2026-04-01T02:21:55Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b5a2dfdfead89c6b2b27e23b612cab7c543552d4cee65cb6e4572e5db9fd056d","abstract_canon_sha256":"53003308bd30e697d90923ecf347728e2e39b9fd50678075f1ab551c9ba81eba"},"schema_version":"1.0"},"canonical_sha256":"1c06dbfc547eb6997a237bf7598f3d6ed2b438cf77cbfd0b67e91f127f246979","source":{"kind":"arxiv","id":"2604.00392","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.00392","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"arxiv_version","alias_value":"2604.00392v2","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.00392","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"pith_short_12","alias_value":"DQDNX7CUP23J","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"pith_short_16","alias_value":"DQDNX7CUP23JS6RD","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"pith_short_8","alias_value":"DQDNX7CU","created_at":"2026-07-07T02:18:39Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:DQDNX7CUP23JS6RDPP3VTDZ5N3","target":"record","payload":{"canonical_record":{"source":{"id":"2604.00392","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2026-04-01T02:21:55Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b5a2dfdfead89c6b2b27e23b612cab7c543552d4cee65cb6e4572e5db9fd056d","abstract_canon_sha256":"53003308bd30e697d90923ecf347728e2e39b9fd50678075f1ab551c9ba81eba"},"schema_version":"1.0"},"canonical_sha256":"1c06dbfc547eb6997a237bf7598f3d6ed2b438cf77cbfd0b67e91f127f246979","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-07T02:18:39.439841Z","signature_b64":"+R/fRV6oticP02F6FGmYn2ylvmEAKp52f3ws2vUWCgFCr3LpthcteLvTwQN2llgzEoOhdeyDtEVYxdGft9llBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1c06dbfc547eb6997a237bf7598f3d6ed2b438cf77cbfd0b67e91f127f246979","last_reissued_at":"2026-07-07T02:18:39.438998Z","signature_status":"signed_v1","first_computed_at":"2026-07-07T02:18:39.438998Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.00392","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-07T02:18:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iX7PoEQiSMOHentw6W44PdDh7mXwxaBbdOdlGLvCWUAahoEjNkdrpN1YeE+Q9++5F8CCMOs6pLZAbLRexKiHAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-02T10:31:28.378430Z"},"content_sha256":"96d9e92797baf5cd85e3b27d91c2902d3d13aeb06a348a106ec7332487e48102","schema_version":"1.0","event_id":"sha256:96d9e92797baf5cd85e3b27d91c2902d3d13aeb06a348a106ec7332487e48102"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:DQDNX7CUP23JS6RDPP3VTDZ5N3","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Beyond Task Completion: A Verification-vs.-Conformance Gap in Tool-Evolving Agents","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.SE","authors_text":"Alibek Kaliyev, Artem Maryanskyy","submitted_at":"2026-04-01T02:21:55Z","abstract_excerpt":"Agents that synthesize their own tools ship a second artifact alongside each answer: a software library that future tasks reuse, compose, and depend on. Task completion (TC) certifies the answer; it does not certify the library. On a Claude Haiku 4.5 pilot, we patch the harness to preserve per-tool source and replay every synthesised tool against a held-out conformance suite. Across 222 preserved tools and three protocols, 96.8% record per-tool correctness C=0.00: two protocols silent-rot at 100%, one at 91.7%. Hand-written reference implementations score C=1.00 on all 16 capability suites. Th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.00392","kind":"arxiv","version":2},"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/2604.00392/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-07T02:18:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vx5sJazeFu4ToYzjk9OrP3H//ycfQew7n0kgELantNONLkUjWQmttD1A55i4YX6+sQEZhjC9MOuwFgBhdjrxCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-02T10:31:28.378968Z"},"content_sha256":"808337cc30633e88dc5ee532b238d5f8ae19b31d84c5aa6bc828fdb352f8d17f","schema_version":"1.0","event_id":"sha256:808337cc30633e88dc5ee532b238d5f8ae19b31d84c5aa6bc828fdb352f8d17f"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/DQDNX7CUP23JS6RDPP3VTDZ5N3/bundle.json","state_url":"https://pith.science/pith/DQDNX7CUP23JS6RDPP3VTDZ5N3/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/DQDNX7CUP23JS6RDPP3VTDZ5N3/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":"2026-08-02T10:31:28Z","links":{"resolver":"https://pith.science/pith/DQDNX7CUP23JS6RDPP3VTDZ5N3","bundle":"https://pith.science/pith/DQDNX7CUP23JS6RDPP3VTDZ5N3/bundle.json","state":"https://pith.science/pith/DQDNX7CUP23JS6RDPP3VTDZ5N3/state.json","well_known_bundle":"https://pith.science/.well-known/pith/DQDNX7CUP23JS6RDPP3VTDZ5N3/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DQDNX7CUP23JS6RDPP3VTDZ5N3","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":"53003308bd30e697d90923ecf347728e2e39b9fd50678075f1ab551c9ba81eba","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2026-04-01T02:21:55Z","title_canon_sha256":"b5a2dfdfead89c6b2b27e23b612cab7c543552d4cee65cb6e4572e5db9fd056d"},"schema_version":"1.0","source":{"id":"2604.00392","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.00392","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"arxiv_version","alias_value":"2604.00392v2","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.00392","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"pith_short_12","alias_value":"DQDNX7CUP23J","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"pith_short_16","alias_value":"DQDNX7CUP23JS6RD","created_at":"2026-07-07T02:18:39Z"},{"alias_kind":"pith_short_8","alias_value":"DQDNX7CU","created_at":"2026-07-07T02:18:39Z"}],"graph_snapshots":[{"event_id":"sha256:808337cc30633e88dc5ee532b238d5f8ae19b31d84c5aa6bc828fdb352f8d17f","target":"graph","created_at":"2026-07-07T02:18:39Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.00392/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Agents that synthesize their own tools ship a second artifact alongside each answer: a software library that future tasks reuse, compose, and depend on. Task completion (TC) certifies the answer; it does not certify the library. On a Claude Haiku 4.5 pilot, we patch the harness to preserve per-tool source and replay every synthesised tool against a held-out conformance suite. Across 222 preserved tools and three protocols, 96.8% record per-tool correctness C=0.00: two protocols silent-rot at 100%, one at 91.7%. Hand-written reference implementations score C=1.00 on all 16 capability suites. Th","authors_text":"Alibek Kaliyev, Artem Maryanskyy","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2026-04-01T02:21:55Z","title":"Beyond Task Completion: A Verification-vs.-Conformance Gap in Tool-Evolving Agents"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.00392","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:96d9e92797baf5cd85e3b27d91c2902d3d13aeb06a348a106ec7332487e48102","target":"record","created_at":"2026-07-07T02:18:39Z","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":"53003308bd30e697d90923ecf347728e2e39b9fd50678075f1ab551c9ba81eba","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2026-04-01T02:21:55Z","title_canon_sha256":"b5a2dfdfead89c6b2b27e23b612cab7c543552d4cee65cb6e4572e5db9fd056d"},"schema_version":"1.0","source":{"id":"2604.00392","kind":"arxiv","version":2}},"canonical_sha256":"1c06dbfc547eb6997a237bf7598f3d6ed2b438cf77cbfd0b67e91f127f246979","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1c06dbfc547eb6997a237bf7598f3d6ed2b438cf77cbfd0b67e91f127f246979","first_computed_at":"2026-07-07T02:18:39.438998Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-07T02:18:39.438998Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"+R/fRV6oticP02F6FGmYn2ylvmEAKp52f3ws2vUWCgFCr3LpthcteLvTwQN2llgzEoOhdeyDtEVYxdGft9llBQ==","signature_status":"signed_v1","signed_at":"2026-07-07T02:18:39.439841Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.00392","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:96d9e92797baf5cd85e3b27d91c2902d3d13aeb06a348a106ec7332487e48102","sha256:808337cc30633e88dc5ee532b238d5f8ae19b31d84c5aa6bc828fdb352f8d17f"],"state_sha256":"648209f4f6fff1a1eea76dcecd0a269ffa092450babb1af7afd7244c3ef4ef92"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"l/bnjFFLcXSVxu35Q1wni37YD9LEtJNEh62Sts/ZuPQ61+V4sOkuC9p4nE0DRa/J04+U93sl4lfZd1iHs9sjDg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-02T10:31:28.382582Z","bundle_sha256":"9c4b05ff3d40318df453eeed75d3222089f66dadf88f0c810dff05c54e07ae40"}}