{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ERAQLQKVHDHJKU74JIG22HCIV6","short_pith_number":"pith:ERAQLQKV","schema_version":"1.0","canonical_sha256":"244105c15538ce9553fc4a0dad1c48afb8028155429e549016d1a81e172a9a1d","source":{"kind":"arxiv","id":"2605.29178","version":1},"attestation_state":"computed","paper":{"title":"The Best-Laid SCHEMEs: Coordinated Sabotage and Monitoring in Multi-Agent Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.MA"],"primary_cat":"cs.CR","authors_text":"Benjamin Arnav, Lennart Haas, Nikolay Radev, Pablo Bernabeu-P\\'erez","submitted_at":"2026-05-27T23:30:21Z","abstract_excerpt":"As agentic coding systems decompose work across multiple model instances, a critical safety question is whether those instances can coordinate to achieve a hidden malicious objective while remaining aligned with user intent. We introduce SCHEME, a benchmark of 17 task instances across 7 settings and 8 real open-source libraries, each pairing a legitimate software-engineering task with a covert side task. Every setting is designed so that no proper subset of agents can succeed alone: agents must decompose a shared sabotage plan, relay partial requirements under different communication topologie"},"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":"2605.29178","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-05-27T23:30:21Z","cross_cats_sorted":["cs.MA"],"title_canon_sha256":"bdf72a5e63a4e6aeae9d78b6cb6d079ad6de60abd3e941e03942648c2f6ec0c2","abstract_canon_sha256":"e339969a67d03666f463734b75b2a7395a9bdb2d4d49ef815a7ee56d2e0624f0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T01:05:22.956512Z","signature_b64":"qo9hnIRYSZLSb666fcf1BMVYGbfL94FbjvJFXafUoELZsx/rCYRDqfaLsuw/3+OwGM+SwKkTiCezCxslzayKDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"244105c15538ce9553fc4a0dad1c48afb8028155429e549016d1a81e172a9a1d","last_reissued_at":"2026-05-29T01:05:22.955914Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T01:05:22.955914Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"The Best-Laid SCHEMEs: Coordinated Sabotage and Monitoring in Multi-Agent Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.MA"],"primary_cat":"cs.CR","authors_text":"Benjamin Arnav, Lennart Haas, Nikolay Radev, Pablo Bernabeu-P\\'erez","submitted_at":"2026-05-27T23:30:21Z","abstract_excerpt":"As agentic coding systems decompose work across multiple model instances, a critical safety question is whether those instances can coordinate to achieve a hidden malicious objective while remaining aligned with user intent. We introduce SCHEME, a benchmark of 17 task instances across 7 settings and 8 real open-source libraries, each pairing a legitimate software-engineering task with a covert side task. Every setting is designed so that no proper subset of agents can succeed alone: agents must decompose a shared sabotage plan, relay partial requirements under different communication topologie"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.29178","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/2605.29178/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":"2605.29178","created_at":"2026-05-29T01:05:22.956012+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.29178v1","created_at":"2026-05-29T01:05:22.956012+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.29178","created_at":"2026-05-29T01:05:22.956012+00:00"},{"alias_kind":"pith_short_12","alias_value":"ERAQLQKVHDHJ","created_at":"2026-05-29T01:05:22.956012+00:00"},{"alias_kind":"pith_short_16","alias_value":"ERAQLQKVHDHJKU74","created_at":"2026-05-29T01:05:22.956012+00:00"},{"alias_kind":"pith_short_8","alias_value":"ERAQLQKV","created_at":"2026-05-29T01:05:22.956012+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/ERAQLQKVHDHJKU74JIG22HCIV6","json":"https://pith.science/pith/ERAQLQKVHDHJKU74JIG22HCIV6.json","graph_json":"https://pith.science/api/pith-number/ERAQLQKVHDHJKU74JIG22HCIV6/graph.json","events_json":"https://pith.science/api/pith-number/ERAQLQKVHDHJKU74JIG22HCIV6/events.json","paper":"https://pith.science/paper/ERAQLQKV"},"agent_actions":{"view_html":"https://pith.science/pith/ERAQLQKVHDHJKU74JIG22HCIV6","download_json":"https://pith.science/pith/ERAQLQKVHDHJKU74JIG22HCIV6.json","view_paper":"https://pith.science/paper/ERAQLQKV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.29178&json=true","fetch_graph":"https://pith.science/api/pith-number/ERAQLQKVHDHJKU74JIG22HCIV6/graph.json","fetch_events":"https://pith.science/api/pith-number/ERAQLQKVHDHJKU74JIG22HCIV6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ERAQLQKVHDHJKU74JIG22HCIV6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ERAQLQKVHDHJKU74JIG22HCIV6/action/storage_attestation","attest_author":"https://pith.science/pith/ERAQLQKVHDHJKU74JIG22HCIV6/action/author_attestation","sign_citation":"https://pith.science/pith/ERAQLQKVHDHJKU74JIG22HCIV6/action/citation_signature","submit_replication":"https://pith.science/pith/ERAQLQKVHDHJKU74JIG22HCIV6/action/replication_record"}},"created_at":"2026-05-29T01:05:22.956012+00:00","updated_at":"2026-05-29T01:05:22.956012+00:00"}