{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:MH7AEZCLHSBBIBWDPIDHF47P2J","short_pith_number":"pith:MH7AEZCL","schema_version":"1.0","canonical_sha256":"61fe02644b3c821406c37a0672f3efd2769ad655f4008c3b6eb620d0dd5f0d37","source":{"kind":"arxiv","id":"2605.28122","version":1},"attestation_state":"computed","paper":{"title":"SNARE: Adaptive Scenario Synthesis for Eliciting Overeager Behavior in Coding Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Gelei Deng, Leo Yu Zhang, Yanjun Zhang, Yi Liu, Ying Zhang, Yubin Qu, Yuekang Li","submitted_at":"2026-05-27T08:14:07Z","abstract_excerpt":"A coding agent executes a benign task as a sequence of shell, file, and network actions, any of which can quietly exceed the authorized scope while the task still completes. We call this overeager behavior: the prompt is not adversarial and the run succeeds, yet an out-of-scope step can leak credentials or delete files. Existing benchmarks miss it: task-completion suites credit any finished run, jailbreak suites probe adversarial prompts, and the one prior overeager benchmark applies a single fixed prompt set to every agent-model pair, leaving its easiest and most resistant pairs under-measure"},"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":"2605.28122","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-05-27T08:14:07Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"17e40b7e1651d32a436d4284ea5d931eb4b906fa3a79c4bcc1bae32095c32a18","abstract_canon_sha256":"60c79517e3e5150dc8efec47b4fd1f34deac87e7d194a0948d30c1d51de23f6a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:59.679250Z","signature_b64":"VFUsOh45S7OyCAXsbEpK48z5WzOmuqZIgCXmCau/odY7H0SSBguqRSRemYjzqsVGgZ8WmlcPD4+5wM19ajKxCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"61fe02644b3c821406c37a0672f3efd2769ad655f4008c3b6eb620d0dd5f0d37","last_reissued_at":"2026-05-28T01:04:59.678649Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:59.678649Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SNARE: Adaptive Scenario Synthesis for Eliciting Overeager Behavior in Coding Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CR","authors_text":"Gelei Deng, Leo Yu Zhang, Yanjun Zhang, Yi Liu, Ying Zhang, Yubin Qu, Yuekang Li","submitted_at":"2026-05-27T08:14:07Z","abstract_excerpt":"A coding agent executes a benign task as a sequence of shell, file, and network actions, any of which can quietly exceed the authorized scope while the task still completes. We call this overeager behavior: the prompt is not adversarial and the run succeeds, yet an out-of-scope step can leak credentials or delete files. Existing benchmarks miss it: task-completion suites credit any finished run, jailbreak suites probe adversarial prompts, and the one prior overeager benchmark applies a single fixed prompt set to every agent-model pair, leaving its easiest and most resistant pairs under-measure"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28122","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/2605.28122/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":"2605.28122","created_at":"2026-05-28T01:04:59.678738+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28122v1","created_at":"2026-05-28T01:04:59.678738+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28122","created_at":"2026-05-28T01:04:59.678738+00:00"},{"alias_kind":"pith_short_12","alias_value":"MH7AEZCLHSBB","created_at":"2026-05-28T01:04:59.678738+00:00"},{"alias_kind":"pith_short_16","alias_value":"MH7AEZCLHSBBIBWD","created_at":"2026-05-28T01:04:59.678738+00:00"},{"alias_kind":"pith_short_8","alias_value":"MH7AEZCL","created_at":"2026-05-28T01:04:59.678738+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/MH7AEZCLHSBBIBWDPIDHF47P2J","json":"https://pith.science/pith/MH7AEZCLHSBBIBWDPIDHF47P2J.json","graph_json":"https://pith.science/api/pith-number/MH7AEZCLHSBBIBWDPIDHF47P2J/graph.json","events_json":"https://pith.science/api/pith-number/MH7AEZCLHSBBIBWDPIDHF47P2J/events.json","paper":"https://pith.science/paper/MH7AEZCL"},"agent_actions":{"view_html":"https://pith.science/pith/MH7AEZCLHSBBIBWDPIDHF47P2J","download_json":"https://pith.science/pith/MH7AEZCLHSBBIBWDPIDHF47P2J.json","view_paper":"https://pith.science/paper/MH7AEZCL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28122&json=true","fetch_graph":"https://pith.science/api/pith-number/MH7AEZCLHSBBIBWDPIDHF47P2J/graph.json","fetch_events":"https://pith.science/api/pith-number/MH7AEZCLHSBBIBWDPIDHF47P2J/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MH7AEZCLHSBBIBWDPIDHF47P2J/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MH7AEZCLHSBBIBWDPIDHF47P2J/action/storage_attestation","attest_author":"https://pith.science/pith/MH7AEZCLHSBBIBWDPIDHF47P2J/action/author_attestation","sign_citation":"https://pith.science/pith/MH7AEZCLHSBBIBWDPIDHF47P2J/action/citation_signature","submit_replication":"https://pith.science/pith/MH7AEZCLHSBBIBWDPIDHF47P2J/action/replication_record"}},"created_at":"2026-05-28T01:04:59.678738+00:00","updated_at":"2026-05-28T01:04:59.678738+00:00"}