{"schema_version":"papers/paper-detail-v1","title":"How Reliable Are AI Attackers Against a Fixed Vulnerable Target? A 400-Run Empirical Study of LLM Penetration Testing Consistency","surface":"papers","opportunity_kernel":{"paper_id":"63fade7e-45d8-45ff-a3fe-7f4f1a08dc07","title":"How Reliable Are AI Attackers Against a Fixed Vulnerable Target? A 400-Run Empirical Study of LLM Penetration Testing Consistency","authors":["Galip Tolga Erdem"],"arxiv_id":"2605.30096v1","doi":null,"published_at":"2026-05-28T15:39:43.000Z","score_object":{"overall":{"value":5,"scale":"0-10","confidence":0.85,"confidence_reason":"Backfilled from persisted papers.viability_score.","model_version":"phase0-web-fallback-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-29T20:15:12.052Z","fresh_until":"2026-06-28T20:15:12.052Z","is_stale":false,"source_count":1,"missingness":[]},"technical":{"value":0,"scale":"0-10","confidence":0.15,"confidence_reason":"No canonical technical score row was available yet; marked low confidence.","model_version":"phase0-web-fallback-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-29T20:15:12.052Z","fresh_until":"2026-06-12T20:15:12.052Z","is_stale":false,"source_count":0,"missingness":["paper_score_objects.technical"]},"commercial":{"value":4,"scale":"0-10","confidence":0.7,"confidence_reason":"Backfilled from persisted commercial_flags and repo availability.","model_version":"phase0-web-fallback-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-29T20:15:12.052Z","fresh_until":"2026-06-28T20:15:12.052Z","is_stale":false,"source_count":1,"missingness":[]},"market":{"value":0,"scale":"0-10","confidence":0.15,"confidence_reason":"No canonical market score row was available yet; marked low confidence.","model_version":"phase0-web-fallback-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-29T20:15:12.052Z","fresh_until":"2026-06-12T20:15:12.052Z","is_stale":false,"source_count":0,"missingness":["paper_score_objects.market"]},"team":{"value":0,"scale":"0-10","confidence":0.15,"confidence_reason":"No canonical team score row was available yet; marked low confidence.","model_version":"phase0-web-fallback-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-29T20:15:12.052Z","fresh_until":"2026-06-12T20:15:12.052Z","is_stale":false,"source_count":0,"missingness":["paper_score_objects.team"]},"methodology":{"value":0,"scale":"0-10","confidence":0.15,"confidence_reason":"No canonical methodology score row was available yet; marked low confidence.","model_version":"phase0-web-fallback-v1","pipeline_version":"phase0-kernel-v1","computed_at":"2026-05-29T20:15:12.052Z","fresh_until":"2026-06-28T20:15:12.052Z","is_stale":false,"source_count":0,"missingness":["paper_score_objects.methodology"]}},"evidence_receipt":{"freshness":"fresh","proof_status":"unverified","repo_status":"missing","references_count":0,"source_count":0,"coverage":0,"missingness":["paper_evidence_receipts.references_count","paper_evidence_receipts.coverage"],"unresolved_unknowns":["Canonical evidence receipt has not been materialized yet."],"last_verification_at":"2026-05-29T20:15:12.052Z"},"lineage_hash":"c8582e4aed3fa3112b7abc088c43afa2fbaa4614e682ad654b3d14e75056ebe0"},"distribution":null,"replication_evidence":[],"author_dna":[]}