{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:LJWA5EL3HA3KQY4QTWOABPVCNN","short_pith_number":"pith:LJWA5EL3","canonical_record":{"source":{"id":"2604.11950","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-04-13T18:44:02Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"title_canon_sha256":"0a2dd23cd449d39253f2ddc01d0ae18fe53aafe02015ab395d4f8dcaf961f91b","abstract_canon_sha256":"0c08f5ea0276acc59364e73694ea554bf554bd31db0d5d2bd6300d81581c465d"},"schema_version":"1.0"},"canonical_sha256":"5a6c0e917b3836a863909d9c00bea26b5e7d5fdc4cdd098c24f1adbdebf022be","source":{"kind":"arxiv","id":"2604.11950","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.11950","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"arxiv_version","alias_value":"2604.11950v2","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.11950","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"pith_short_12","alias_value":"LJWA5EL3HA3K","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"pith_short_16","alias_value":"LJWA5EL3HA3KQY4Q","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"pith_short_8","alias_value":"LJWA5EL3","created_at":"2026-07-09T01:20:35Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:LJWA5EL3HA3KQY4QTWOABPVCNN","target":"record","payload":{"canonical_record":{"source":{"id":"2604.11950","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-04-13T18:44:02Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"title_canon_sha256":"0a2dd23cd449d39253f2ddc01d0ae18fe53aafe02015ab395d4f8dcaf961f91b","abstract_canon_sha256":"0c08f5ea0276acc59364e73694ea554bf554bd31db0d5d2bd6300d81581c465d"},"schema_version":"1.0"},"canonical_sha256":"5a6c0e917b3836a863909d9c00bea26b5e7d5fdc4cdd098c24f1adbdebf022be","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-09T01:20:35.844744Z","signature_b64":"KC7Lh1dvpfODpYyhb/j8zI7WtO20wcQ9ySM5hsl5vhnFINz0LS8EBbcwTsjPLcdtkC+BTPReY8ippfm/TLTgAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5a6c0e917b3836a863909d9c00bea26b5e7d5fdc4cdd098c24f1adbdebf022be","last_reissued_at":"2026-07-09T01:20:35.844135Z","signature_status":"signed_v1","first_computed_at":"2026-07-09T01:20:35.844135Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.11950","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-09T01:20:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jfg7iSSoBsb3RdIFhook8rcSUEs5nNLcQtFs6yEj3GT4vIbZb95OJd2vm/Twu+NsE2rBcflcyVwo+4XQDttgAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T09:47:05.620441Z"},"content_sha256":"42b5861466120c356985ffc0e93a43305d854f09a1246a317d9088fe8f5f8637","schema_version":"1.0","event_id":"sha256:42b5861466120c356985ffc0e93a43305d854f09a1246a317d9088fe8f5f8637"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:LJWA5EL3HA3KQY4QTWOABPVCNN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AnyPoC: Universal Proof-of-Concept Test Generation for Scalable LLM-Based Bug Detection","license":"http://creativecommons.org/licenses/by/4.0/","headline":"AnyPoC generates executable proof-of-concept tests to automatically validate and confirm bugs reported by LLM agents.","cross_cats":["cs.AI","cs.CL","cs.CR"],"primary_cat":"cs.SE","authors_text":"Chenyuan Yang, Lingming Zhang, Weidong Wang, Yihan Yang, Zijie Zhao, Ziqi Zhang","submitted_at":"2026-04-13T18:44:02Z","abstract_excerpt":"While recent LLM-based agents can identify many candidate bugs in source code, their reports remain static hypotheses that require manual validation, limiting the practicality of automated bug detection. We frame this challenge as a test generation task: given a candidate report, synthesizing an executable proof-of-concept (PoC) - such as a script, command sequence, or crafted input - to trigger the suspected defect. Automated PoC generation can act as a scalable validation oracle, enabling end-to-end autonomous bug detection by providing concrete execution evidence. However, naive LLM agents "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"AnyPoC produces 1.3x more valid PoCs for true-positive bug reports and rejects 9.8x more false-positive bug reports. To date, AnyPoC has discovered 122 new bugs (105 confirmed, 86 already fixed), with 45 generated PoCs adopted as official regression tests.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the combination of fact-checking, iterative PoC synthesis with execution traces, and independent re-execution reliably prevents LLM hallucination and reward-hacking across heterogeneous languages and domains without introducing new failure modes.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"AnyPoC introduces a multi-agent system for generating and validating PoC tests from LLM bug reports, producing 1.3x more valid PoCs, rejecting 9.8x more false positives, and discovering 122 new bugs across 12 major projects.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"AnyPoC generates executable proof-of-concept tests to automatically validate and confirm bugs reported by LLM agents.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"17190d3e002e09ce4fd782ea4cdd1b83133531b86e99662be0bca14dfe53e477"},"source":{"id":"2604.11950","kind":"arxiv","version":2},"verdict":{"id":"f1e7da3a-3b49-49ae-8d0a-33802c9dbee7","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T15:55:35.410089Z","strongest_claim":"AnyPoC produces 1.3x more valid PoCs for true-positive bug reports and rejects 9.8x more false-positive bug reports. To date, AnyPoC has discovered 122 new bugs (105 confirmed, 86 already fixed), with 45 generated PoCs adopted as official regression tests.","one_line_summary":"AnyPoC introduces a multi-agent system for generating and validating PoC tests from LLM bug reports, producing 1.3x more valid PoCs, rejecting 9.8x more false positives, and discovering 122 new bugs across 12 major projects.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the combination of fact-checking, iterative PoC synthesis with execution traces, and independent re-execution reliably prevents LLM hallucination and reward-hacking across heterogeneous languages and domains without introducing new failure modes.","pith_extraction_headline":"AnyPoC generates executable proof-of-concept tests to automatically validate and confirm bugs reported by LLM agents."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.11950/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"},"verdict_id":"f1e7da3a-3b49-49ae-8d0a-33802c9dbee7"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-09T01:20:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dgPnPBpjY05bkL0+uT9HBAa4ZHFZ5Mlzw44kx3Q7SDFjUUulWEQnC3SU0/W2gB3KNYfuzxNGNXliwJcS0CUEDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-13T09:47:05.621184Z"},"content_sha256":"f255343d075f6460b8fc274af8b30fdf03309af6c8b83fdbb18ac284739eac55","schema_version":"1.0","event_id":"sha256:f255343d075f6460b8fc274af8b30fdf03309af6c8b83fdbb18ac284739eac55"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LJWA5EL3HA3KQY4QTWOABPVCNN/bundle.json","state_url":"https://pith.science/pith/LJWA5EL3HA3KQY4QTWOABPVCNN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LJWA5EL3HA3KQY4QTWOABPVCNN/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-13T09:47:05Z","links":{"resolver":"https://pith.science/pith/LJWA5EL3HA3KQY4QTWOABPVCNN","bundle":"https://pith.science/pith/LJWA5EL3HA3KQY4QTWOABPVCNN/bundle.json","state":"https://pith.science/pith/LJWA5EL3HA3KQY4QTWOABPVCNN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LJWA5EL3HA3KQY4QTWOABPVCNN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:LJWA5EL3HA3KQY4QTWOABPVCNN","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"0c08f5ea0276acc59364e73694ea554bf554bd31db0d5d2bd6300d81581c465d","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-04-13T18:44:02Z","title_canon_sha256":"0a2dd23cd449d39253f2ddc01d0ae18fe53aafe02015ab395d4f8dcaf961f91b"},"schema_version":"1.0","source":{"id":"2604.11950","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.11950","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"arxiv_version","alias_value":"2604.11950v2","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.11950","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"pith_short_12","alias_value":"LJWA5EL3HA3K","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"pith_short_16","alias_value":"LJWA5EL3HA3KQY4Q","created_at":"2026-07-09T01:20:35Z"},{"alias_kind":"pith_short_8","alias_value":"LJWA5EL3","created_at":"2026-07-09T01:20:35Z"}],"graph_snapshots":[{"event_id":"sha256:f255343d075f6460b8fc274af8b30fdf03309af6c8b83fdbb18ac284739eac55","target":"graph","created_at":"2026-07-09T01:20:35Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"AnyPoC produces 1.3x more valid PoCs for true-positive bug reports and rejects 9.8x more false-positive bug reports. To date, AnyPoC has discovered 122 new bugs (105 confirmed, 86 already fixed), with 45 generated PoCs adopted as official regression tests."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the combination of fact-checking, iterative PoC synthesis with execution traces, and independent re-execution reliably prevents LLM hallucination and reward-hacking across heterogeneous languages and domains without introducing new failure modes."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"AnyPoC introduces a multi-agent system for generating and validating PoC tests from LLM bug reports, producing 1.3x more valid PoCs, rejecting 9.8x more false positives, and discovering 122 new bugs across 12 major projects."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"AnyPoC generates executable proof-of-concept tests to automatically validate and confirm bugs reported by LLM agents."}],"snapshot_sha256":"17190d3e002e09ce4fd782ea4cdd1b83133531b86e99662be0bca14dfe53e477"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.11950/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"While recent LLM-based agents can identify many candidate bugs in source code, their reports remain static hypotheses that require manual validation, limiting the practicality of automated bug detection. We frame this challenge as a test generation task: given a candidate report, synthesizing an executable proof-of-concept (PoC) - such as a script, command sequence, or crafted input - to trigger the suspected defect. Automated PoC generation can act as a scalable validation oracle, enabling end-to-end autonomous bug detection by providing concrete execution evidence. However, naive LLM agents ","authors_text":"Chenyuan Yang, Lingming Zhang, Weidong Wang, Yihan Yang, Zijie Zhao, Ziqi Zhang","cross_cats":["cs.AI","cs.CL","cs.CR"],"headline":"AnyPoC generates executable proof-of-concept tests to automatically validate and confirm bugs reported by LLM agents.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-04-13T18:44:02Z","title":"AnyPoC: Universal Proof-of-Concept Test Generation for Scalable LLM-Based Bug Detection"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.11950","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-10T15:55:35.410089Z","id":"f1e7da3a-3b49-49ae-8d0a-33802c9dbee7","model_set":{"reader":"grok-4.3"},"one_line_summary":"AnyPoC introduces a multi-agent system for generating and validating PoC tests from LLM bug reports, producing 1.3x more valid PoCs, rejecting 9.8x more false positives, and discovering 122 new bugs across 12 major projects.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"AnyPoC generates executable proof-of-concept tests to automatically validate and confirm bugs reported by LLM agents.","strongest_claim":"AnyPoC produces 1.3x more valid PoCs for true-positive bug reports and rejects 9.8x more false-positive bug reports. To date, AnyPoC has discovered 122 new bugs (105 confirmed, 86 already fixed), with 45 generated PoCs adopted as official regression tests.","weakest_assumption":"That the combination of fact-checking, iterative PoC synthesis with execution traces, and independent re-execution reliably prevents LLM hallucination and reward-hacking across heterogeneous languages and domains without introducing new failure modes."}},"verdict_id":"f1e7da3a-3b49-49ae-8d0a-33802c9dbee7"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:42b5861466120c356985ffc0e93a43305d854f09a1246a317d9088fe8f5f8637","target":"record","created_at":"2026-07-09T01:20:35Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"0c08f5ea0276acc59364e73694ea554bf554bd31db0d5d2bd6300d81581c465d","cross_cats_sorted":["cs.AI","cs.CL","cs.CR"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2026-04-13T18:44:02Z","title_canon_sha256":"0a2dd23cd449d39253f2ddc01d0ae18fe53aafe02015ab395d4f8dcaf961f91b"},"schema_version":"1.0","source":{"id":"2604.11950","kind":"arxiv","version":2}},"canonical_sha256":"5a6c0e917b3836a863909d9c00bea26b5e7d5fdc4cdd098c24f1adbdebf022be","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5a6c0e917b3836a863909d9c00bea26b5e7d5fdc4cdd098c24f1adbdebf022be","first_computed_at":"2026-07-09T01:20:35.844135Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-09T01:20:35.844135Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KC7Lh1dvpfODpYyhb/j8zI7WtO20wcQ9ySM5hsl5vhnFINz0LS8EBbcwTsjPLcdtkC+BTPReY8ippfm/TLTgAA==","signature_status":"signed_v1","signed_at":"2026-07-09T01:20:35.844744Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.11950","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:42b5861466120c356985ffc0e93a43305d854f09a1246a317d9088fe8f5f8637","sha256:f255343d075f6460b8fc274af8b30fdf03309af6c8b83fdbb18ac284739eac55"],"state_sha256":"021f6323206c42dda9267ae6346b0e5c313f959875a08ed76cfe6cbc9400797d"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"V8wj4jO5Zb3JmFwQGTwPNo6HR/fC+4heQPxKnLhg38LKTbQbyRvgzXn2OuEnBhoIOALh63EeEPAlyD0ZQuVICA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-13T09:47:05.626763Z","bundle_sha256":"0d9ddd703b6a6074fc72dffead96d199566b223bdf124d06d0e75606d7e5352a"}}