{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:R5IDVVFMYOFG2EJPHRINMP4VLJ","short_pith_number":"pith:R5IDVVFM","schema_version":"1.0","canonical_sha256":"8f503ad4acc38a6d112f3c50d63f955a56b6633c75a7102ba1697c236a374d81","source":{"kind":"arxiv","id":"2606.22659","version":1},"attestation_state":"computed","paper":{"title":"Confidently Wrong: Severity-Aware Calibration of Prompt-Injection Detectors under Attack Shift","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Md Anas Biswas","submitted_at":"2026-06-21T20:30:12Z","abstract_excerpt":"Prompt-injection detectors are deployed as guards: a model scores an input and a downstream system trusts or blocks it on that score. I study the confidence of these scores, not only their accuracy, when the attack distribution shifts away from the clean benchmark on which the operating point was chosen. I evaluate three released detectors, ProtectAI-v2 and two Prompt-Guard-2 checkpoints, at a single source-calibrated threshold that I freeze and transport across five shifts. I report a severity metric S, how confident a detector is on the attacks it misses, alongside the false-negative rate an"},"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":"2606.22659","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-06-21T20:30:12Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"d939ba009c8f7c94bd9c80f5db6340f54da8121606bf48a7496e725244f8284c","abstract_canon_sha256":"9b88f921f7c37a50c94a0d9b2f9935acb3376fc4cbf61316f553fa4e771e4996"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-23T02:13:44.172174Z","signature_b64":"DTcArZTFMtn4e0VhtOwtIBib/apb0IhjsirPCAv7u3yAKpo7WUNrYDyDkqjIwkVpJXj6Ri9g8EkOWWdsZyayDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8f503ad4acc38a6d112f3c50d63f955a56b6633c75a7102ba1697c236a374d81","last_reissued_at":"2026-06-23T02:13:44.171746Z","signature_status":"signed_v1","first_computed_at":"2026-06-23T02:13:44.171746Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Confidently Wrong: Severity-Aware Calibration of Prompt-Injection Detectors under Attack Shift","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Md Anas Biswas","submitted_at":"2026-06-21T20:30:12Z","abstract_excerpt":"Prompt-injection detectors are deployed as guards: a model scores an input and a downstream system trusts or blocks it on that score. I study the confidence of these scores, not only their accuracy, when the attack distribution shifts away from the clean benchmark on which the operating point was chosen. I evaluate three released detectors, ProtectAI-v2 and two Prompt-Guard-2 checkpoints, at a single source-calibrated threshold that I freeze and transport across five shifts. I report a severity metric S, how confident a detector is on the attacks it misses, alongside the false-negative rate an"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.22659","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/2606.22659/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":"2606.22659","created_at":"2026-06-23T02:13:44.171819+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.22659v1","created_at":"2026-06-23T02:13:44.171819+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.22659","created_at":"2026-06-23T02:13:44.171819+00:00"},{"alias_kind":"pith_short_12","alias_value":"R5IDVVFMYOFG","created_at":"2026-06-23T02:13:44.171819+00:00"},{"alias_kind":"pith_short_16","alias_value":"R5IDVVFMYOFG2EJP","created_at":"2026-06-23T02:13:44.171819+00:00"},{"alias_kind":"pith_short_8","alias_value":"R5IDVVFM","created_at":"2026-06-23T02:13:44.171819+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/R5IDVVFMYOFG2EJPHRINMP4VLJ","json":"https://pith.science/pith/R5IDVVFMYOFG2EJPHRINMP4VLJ.json","graph_json":"https://pith.science/api/pith-number/R5IDVVFMYOFG2EJPHRINMP4VLJ/graph.json","events_json":"https://pith.science/api/pith-number/R5IDVVFMYOFG2EJPHRINMP4VLJ/events.json","paper":"https://pith.science/paper/R5IDVVFM"},"agent_actions":{"view_html":"https://pith.science/pith/R5IDVVFMYOFG2EJPHRINMP4VLJ","download_json":"https://pith.science/pith/R5IDVVFMYOFG2EJPHRINMP4VLJ.json","view_paper":"https://pith.science/paper/R5IDVVFM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.22659&json=true","fetch_graph":"https://pith.science/api/pith-number/R5IDVVFMYOFG2EJPHRINMP4VLJ/graph.json","fetch_events":"https://pith.science/api/pith-number/R5IDVVFMYOFG2EJPHRINMP4VLJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/R5IDVVFMYOFG2EJPHRINMP4VLJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/R5IDVVFMYOFG2EJPHRINMP4VLJ/action/storage_attestation","attest_author":"https://pith.science/pith/R5IDVVFMYOFG2EJPHRINMP4VLJ/action/author_attestation","sign_citation":"https://pith.science/pith/R5IDVVFMYOFG2EJPHRINMP4VLJ/action/citation_signature","submit_replication":"https://pith.science/pith/R5IDVVFMYOFG2EJPHRINMP4VLJ/action/replication_record"}},"created_at":"2026-06-23T02:13:44.171819+00:00","updated_at":"2026-06-23T02:13:44.171819+00:00"}