{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ZSZB5PARFDOFWWLE54TLHT4432","short_pith_number":"pith:ZSZB5PAR","schema_version":"1.0","canonical_sha256":"ccb21ebc1128dc5b5964ef26b3cf9cdead83f7312955a901fe1c18764c50bf97","source":{"kind":"arxiv","id":"2607.22368","version":1},"attestation_state":"computed","paper":{"title":"Do Agent Benchmarks Measure Capability? Protocol Validity in the Age of Agentic AI","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bing Luo, Hanck Chen, Jiaqi Shao, Maxm Pan, Wei Zhang","submitted_at":"2026-07-24T14:55:19Z","abstract_excerpt":"Agent benchmarks increasingly evaluate repository editing, web research, terminal use, and long-horizon interaction. Their scores support capability claims only when the evaluation protocol keeps the intended capability necessary for success. Recent reward-hacking benchmarks and system reports show that agents can instead recover public solutions, read evaluation artifacts, infer generator structure, manipulate feedback, or benefit from invalid scoring paths; existing responses do not provide a common procedure for attributing these shortcuts and quantifying their effect across benchmarks. We "},"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":"2607.22368","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-07-24T14:55:19Z","cross_cats_sorted":[],"title_canon_sha256":"744ca0099228b3a94fe680cc78658da5542d00d2c3b22c86cf7869e32010f917","abstract_canon_sha256":"192b91dec1d46a7aa7df6c5b508523b56284a6091cbfe5c86ce24507657aeaed"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-27T01:21:16.918660Z","signature_b64":"ipFF61sHT3IuvFHC+LyUyAojy+oMc/gJd7k2wJaUFktml7+8w86Efakrb5MTTUZRoQyRqtVbIY+W7WeOIfgdAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ccb21ebc1128dc5b5964ef26b3cf9cdead83f7312955a901fe1c18764c50bf97","last_reissued_at":"2026-07-27T01:21:16.917794Z","signature_status":"signed_v1","first_computed_at":"2026-07-27T01:21:16.917794Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do Agent Benchmarks Measure Capability? Protocol Validity in the Age of Agentic AI","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bing Luo, Hanck Chen, Jiaqi Shao, Maxm Pan, Wei Zhang","submitted_at":"2026-07-24T14:55:19Z","abstract_excerpt":"Agent benchmarks increasingly evaluate repository editing, web research, terminal use, and long-horizon interaction. Their scores support capability claims only when the evaluation protocol keeps the intended capability necessary for success. Recent reward-hacking benchmarks and system reports show that agents can instead recover public solutions, read evaluation artifacts, infer generator structure, manipulate feedback, or benefit from invalid scoring paths; existing responses do not provide a common procedure for attributing these shortcuts and quantifying their effect across benchmarks. We "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.22368","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/2607.22368/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":"2607.22368","created_at":"2026-07-27T01:21:16.918240+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.22368v1","created_at":"2026-07-27T01:21:16.918240+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.22368","created_at":"2026-07-27T01:21:16.918240+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZSZB5PARFDOF","created_at":"2026-07-27T01:21:16.918240+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZSZB5PARFDOFWWLE","created_at":"2026-07-27T01:21:16.918240+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZSZB5PAR","created_at":"2026-07-27T01:21:16.918240+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/ZSZB5PARFDOFWWLE54TLHT4432","json":"https://pith.science/pith/ZSZB5PARFDOFWWLE54TLHT4432.json","graph_json":"https://pith.science/api/pith-number/ZSZB5PARFDOFWWLE54TLHT4432/graph.json","events_json":"https://pith.science/api/pith-number/ZSZB5PARFDOFWWLE54TLHT4432/events.json","paper":"https://pith.science/paper/ZSZB5PAR"},"agent_actions":{"view_html":"https://pith.science/pith/ZSZB5PARFDOFWWLE54TLHT4432","download_json":"https://pith.science/pith/ZSZB5PARFDOFWWLE54TLHT4432.json","view_paper":"https://pith.science/paper/ZSZB5PAR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.22368&json=true","fetch_graph":"https://pith.science/api/pith-number/ZSZB5PARFDOFWWLE54TLHT4432/graph.json","fetch_events":"https://pith.science/api/pith-number/ZSZB5PARFDOFWWLE54TLHT4432/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZSZB5PARFDOFWWLE54TLHT4432/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZSZB5PARFDOFWWLE54TLHT4432/action/storage_attestation","attest_author":"https://pith.science/pith/ZSZB5PARFDOFWWLE54TLHT4432/action/author_attestation","sign_citation":"https://pith.science/pith/ZSZB5PARFDOFWWLE54TLHT4432/action/citation_signature","submit_replication":"https://pith.science/pith/ZSZB5PARFDOFWWLE54TLHT4432/action/replication_record"}},"created_at":"2026-07-27T01:21:16.918240+00:00","updated_at":"2026-07-27T01:21:16.918240+00:00"}