{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:EPDCRWPQTHVH4LXCCV3ZDBGXNC","short_pith_number":"pith:EPDCRWPQ","canonical_record":{"source":{"id":"2604.00594","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-01T07:59:59Z","cross_cats_sorted":[],"title_canon_sha256":"3a0ff1fcd1a0b893f86c67e8801785ee4e77626319bdb5f379c351c321b2325e","abstract_canon_sha256":"0b9fdc285353d5b68c3079dd81bd1460a2e677f9e25a2107901f8a5803b0528b"},"schema_version":"1.0"},"canonical_sha256":"23c628d9f099ea7e2ee215779184d76881abd37290a72eda5c3bd77ec7cdf30a","source":{"kind":"arxiv","id":"2604.00594","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.00594","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"arxiv_version","alias_value":"2604.00594v2","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.00594","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"pith_short_12","alias_value":"EPDCRWPQTHVH","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"pith_short_16","alias_value":"EPDCRWPQTHVH4LXC","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"pith_short_8","alias_value":"EPDCRWPQ","created_at":"2026-07-21T02:21:34Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:EPDCRWPQTHVH4LXCCV3ZDBGXNC","target":"record","payload":{"canonical_record":{"source":{"id":"2604.00594","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-01T07:59:59Z","cross_cats_sorted":[],"title_canon_sha256":"3a0ff1fcd1a0b893f86c67e8801785ee4e77626319bdb5f379c351c321b2325e","abstract_canon_sha256":"0b9fdc285353d5b68c3079dd81bd1460a2e677f9e25a2107901f8a5803b0528b"},"schema_version":"1.0"},"canonical_sha256":"23c628d9f099ea7e2ee215779184d76881abd37290a72eda5c3bd77ec7cdf30a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-21T02:21:34.774834Z","signature_b64":"Gv3e93XQejMUnsuuA7Ag3du/sbAUKXQohFczxRS20B5YofGxg2iOq472EYEcscrwOTiwTxMiBNm3u6qlkhaVDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"23c628d9f099ea7e2ee215779184d76881abd37290a72eda5c3bd77ec7cdf30a","last_reissued_at":"2026-07-21T02:21:34.773921Z","signature_status":"signed_v1","first_computed_at":"2026-07-21T02:21:34.773921Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.00594","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":"2026-07-21T02:21:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"XPzhQD1gw8u5N8gpHRw3KKPVMWQ3xNBrM6zk1CWwRdVeCDBiKbNXzEXr8Su/qrXjOvlFRCAIrR/olkBTPNSeAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T05:07:26.653309Z"},"content_sha256":"7538f414229f321483a3024bac9caf9ba45dd8e470de19ee0ef0bc34a7e2a3cd","schema_version":"1.0","event_id":"sha256:7538f414229f321483a3024bac9caf9ba45dd8e470de19ee0ef0bc34a7e2a3cd"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:EPDCRWPQTHVH4LXCCV3ZDBGXNC","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Agent psychometrics: Task-level performance prediction in agentic coding benchmarks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Chris Ge, Daniel Fried, Daria Kryvosheieva, Kaivalya Hariharan, Uzay Girit","submitted_at":"2026-04-01T07:59:59Z","abstract_excerpt":"As the focus in LLM-based coding shifts from static single-step code generation to multi-step agentic interaction with tools and environments, understanding which tasks will challenge agents and why becomes increasingly difficult. This is compounded by current practice: agent performance is typically measured by aggregate pass rates on benchmarks, but single-number metrics obscure the diversity of tasks within a benchmark. We present a framework for predicting success or failure on individual tasks tailored to the agentic coding regime. Our approach augments Item Response Theory (IRT) with ric"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.00594","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/2604.00594/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":"2026-07-21T02:21:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CplaISJYis7+gM3iVCcMDyohxYE0KOQn0ejT/kCzeuRKTarRiJlv2NLJlQrcvNE+QxEzCH0hhKEzpEg5W4jwBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T05:07:26.653926Z"},"content_sha256":"c6d1f033c167042be153f1d3b8e0b897f089dbf75d5f7a8138f018184852d030","schema_version":"1.0","event_id":"sha256:c6d1f033c167042be153f1d3b8e0b897f089dbf75d5f7a8138f018184852d030"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/EPDCRWPQTHVH4LXCCV3ZDBGXNC/bundle.json","state_url":"https://pith.science/pith/EPDCRWPQTHVH4LXCCV3ZDBGXNC/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/EPDCRWPQTHVH4LXCCV3ZDBGXNC/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":"2026-08-07T05:07:26Z","links":{"resolver":"https://pith.science/pith/EPDCRWPQTHVH4LXCCV3ZDBGXNC","bundle":"https://pith.science/pith/EPDCRWPQTHVH4LXCCV3ZDBGXNC/bundle.json","state":"https://pith.science/pith/EPDCRWPQTHVH4LXCCV3ZDBGXNC/state.json","well_known_bundle":"https://pith.science/.well-known/pith/EPDCRWPQTHVH4LXCCV3ZDBGXNC/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:EPDCRWPQTHVH4LXCCV3ZDBGXNC","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":"0b9fdc285353d5b68c3079dd81bd1460a2e677f9e25a2107901f8a5803b0528b","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-01T07:59:59Z","title_canon_sha256":"3a0ff1fcd1a0b893f86c67e8801785ee4e77626319bdb5f379c351c321b2325e"},"schema_version":"1.0","source":{"id":"2604.00594","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.00594","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"arxiv_version","alias_value":"2604.00594v2","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.00594","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"pith_short_12","alias_value":"EPDCRWPQTHVH","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"pith_short_16","alias_value":"EPDCRWPQTHVH4LXC","created_at":"2026-07-21T02:21:34Z"},{"alias_kind":"pith_short_8","alias_value":"EPDCRWPQ","created_at":"2026-07-21T02:21:34Z"}],"graph_snapshots":[{"event_id":"sha256:c6d1f033c167042be153f1d3b8e0b897f089dbf75d5f7a8138f018184852d030","target":"graph","created_at":"2026-07-21T02:21:34Z","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/2604.00594/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As the focus in LLM-based coding shifts from static single-step code generation to multi-step agentic interaction with tools and environments, understanding which tasks will challenge agents and why becomes increasingly difficult. This is compounded by current practice: agent performance is typically measured by aggregate pass rates on benchmarks, but single-number metrics obscure the diversity of tasks within a benchmark. We present a framework for predicting success or failure on individual tasks tailored to the agentic coding regime. Our approach augments Item Response Theory (IRT) with ric","authors_text":"Chris Ge, Daniel Fried, Daria Kryvosheieva, Kaivalya Hariharan, Uzay Girit","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-01T07:59:59Z","title":"Agent psychometrics: Task-level performance prediction in agentic coding benchmarks"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.00594","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":"sha256:7538f414229f321483a3024bac9caf9ba45dd8e470de19ee0ef0bc34a7e2a3cd","target":"record","created_at":"2026-07-21T02:21:34Z","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":"0b9fdc285353d5b68c3079dd81bd1460a2e677f9e25a2107901f8a5803b0528b","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-04-01T07:59:59Z","title_canon_sha256":"3a0ff1fcd1a0b893f86c67e8801785ee4e77626319bdb5f379c351c321b2325e"},"schema_version":"1.0","source":{"id":"2604.00594","kind":"arxiv","version":2}},"canonical_sha256":"23c628d9f099ea7e2ee215779184d76881abd37290a72eda5c3bd77ec7cdf30a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"23c628d9f099ea7e2ee215779184d76881abd37290a72eda5c3bd77ec7cdf30a","first_computed_at":"2026-07-21T02:21:34.773921Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-21T02:21:34.773921Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Gv3e93XQejMUnsuuA7Ag3du/sbAUKXQohFczxRS20B5YofGxg2iOq472EYEcscrwOTiwTxMiBNm3u6qlkhaVDg==","signature_status":"signed_v1","signed_at":"2026-07-21T02:21:34.774834Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.00594","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7538f414229f321483a3024bac9caf9ba45dd8e470de19ee0ef0bc34a7e2a3cd","sha256:c6d1f033c167042be153f1d3b8e0b897f089dbf75d5f7a8138f018184852d030"],"state_sha256":"8b6a091dce1005c5f5b929e5e0aeff9316ce147c30c821fcec2fceacc17310b2"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7xYJKyzeFqiw7tw6P5cYjbbgtWwJZ/DSbVpMooRH/JydYyB2szpvXbUXX8KAvrsWffwAru6EK3UxJJOgVMW1DQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T05:07:26.658849Z","bundle_sha256":"8ae826ea861d65fd3cef349f1b5cfa92964fb0ac9d0b42abb8a59fc9f6632101"}}