{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:C43N4527I7EBNA4U3P6FEMSGNT","short_pith_number":"pith:C43N4527","schema_version":"1.0","canonical_sha256":"1736de775f47c8168394dbfc5232466cc48c2b225a6cac2e82cdcf2700ae03c2","source":{"kind":"arxiv","id":"2605.11053","version":3},"attestation_state":"computed","paper":{"title":"Content-Aware Attack Detection in LLM Agent Tool-Call Traffic: An Empirical Study of Features, Architectures, and Evaluation Protocols","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"MCPShield shows that content embeddings of tool arguments and responses are essential for detecting attacks on LLM agent traffic, lifting AUROC from 0.64 to above 0.89.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Sultan Zavrak","submitted_at":"2026-05-11T14:55:48Z","abstract_excerpt":"The Model Context Protocol (MCP) has become a widely adopted interface for LLM agents to invoke external tools, yet learned monitoring of MCP tool-call traffic remains underexplored. In this article, the proposed detector is presented as an attack detection framework for MCP tool-call traffic that encodes each agent session as a graph (tool calls as nodes, sequential and data-flow links as edges), enriches nodes with sentence-embedding features over arguments and responses, and classifies sessions as benign or attacked. Three GNN architectures (GAT, GCN, GraphSAGE), a no-graph MLP, and classic"},"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":true},"canonical_record":{"source":{"id":"2605.11053","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CR","submitted_at":"2026-05-11T14:55:48Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"56536732f9de1996a11d36b08864f8e9afceabd57fc31b3cd927a753d8126632","abstract_canon_sha256":"93766ad117e27044b395a3531afd639c203eda23c05c7c1bc349e4308001ddeb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:01:22.999856Z","signature_b64":"aPg0A+/zkLrGMJkwkcGUh5okbPEzR00NdZeFJknmjfgMQMsez8wWg61j4sjUALTptE+Da7J4NXX1yMOM5mrODQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1736de775f47c8168394dbfc5232466cc48c2b225a6cac2e82cdcf2700ae03c2","last_reissued_at":"2026-05-25T02:01:22.999408Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:01:22.999408Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Content-Aware Attack Detection in LLM Agent Tool-Call Traffic: An Empirical Study of Features, Architectures, and Evaluation Protocols","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"MCPShield shows that content embeddings of tool arguments and responses are essential for detecting attacks on LLM agent traffic, lifting AUROC from 0.64 to above 0.89.","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CR","authors_text":"Sultan Zavrak","submitted_at":"2026-05-11T14:55:48Z","abstract_excerpt":"The Model Context Protocol (MCP) has become a widely adopted interface for LLM agents to invoke external tools, yet learned monitoring of MCP tool-call traffic remains underexplored. In this article, the proposed detector is presented as an attack detection framework for MCP tool-call traffic that encodes each agent session as a graph (tool calls as nodes, sequential and data-flow links as edges), enriches nodes with sentence-embedding features over arguments and responses, and classifies sessions as benign or attacked. Three GNN architectures (GAT, GCN, GraphSAGE), a no-graph MLP, and classic"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Content-level features are essential: metadata-only detection plateaus around an AUROC of 0.64 regardless of architecture, while content embeddings push the AUROC above 0.89.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The attack examples in RAS-Eval and ATBench are representative of real-world threats and that SBERT embeddings capture generalizable attack signals rather than dataset-specific artifacts.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MCPShield detects attacks on LLM agent tool-call traffic by encoding sessions as graphs enriched with SBERT content embeddings, achieving AUROC above 0.89 with content features versus 0.64 for metadata alone.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MCPShield shows that content embeddings of tool arguments and responses are essential for detecting attacks on LLM agent traffic, lifting AUROC from 0.64 to above 0.89.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"2be7d0554a6898f025b0c21336f8fcb86f458ff6dc2839868e23af16eb035329"},"source":{"id":"2605.11053","kind":"arxiv","version":3},"verdict":{"id":"18baac78-0ee7-4c11-9d03-a76e77dfd00c","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T21:13:06.877944Z","strongest_claim":"Content-level features are essential: metadata-only detection plateaus around an AUROC of 0.64 regardless of architecture, while content embeddings push the AUROC above 0.89.","one_line_summary":"MCPShield detects attacks on LLM agent tool-call traffic by encoding sessions as graphs enriched with SBERT content embeddings, achieving AUROC above 0.89 with content features versus 0.64 for metadata alone.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The attack examples in RAS-Eval and ATBench are representative of real-world threats and that SBERT embeddings capture generalizable attack signals rather than dataset-specific artifacts.","pith_extraction_headline":"MCPShield shows that content embeddings of tool arguments and responses are essential for detecting attacks on LLM agent traffic, lifting AUROC from 0.64 to above 0.89."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.11053/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-20T05:22:00.477103Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T14:38:23.932674Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T11:01:17.183645Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T09:04:11.242458Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"4029f4fd7ed3d4315e94a5eedda203f9c01f678d2ad697d8e7bca38106607601"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":1,"snapshot_sha256":"935fcd76a398a5376ac7bcde7de16c8ebef2edfa366fa86a3a77fd3f313afb2a"},"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":"2605.11053","created_at":"2026-05-25T02:01:22.999466+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.11053v3","created_at":"2026-05-25T02:01:22.999466+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.11053","created_at":"2026-05-25T02:01:22.999466+00:00"},{"alias_kind":"pith_short_12","alias_value":"C43N4527I7EB","created_at":"2026-05-25T02:01:22.999466+00:00"},{"alias_kind":"pith_short_16","alias_value":"C43N4527I7EBNA4U","created_at":"2026-05-25T02:01:22.999466+00:00"},{"alias_kind":"pith_short_8","alias_value":"C43N4527","created_at":"2026-05-25T02:01:22.999466+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":1,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/C43N4527I7EBNA4U3P6FEMSGNT","json":"https://pith.science/pith/C43N4527I7EBNA4U3P6FEMSGNT.json","graph_json":"https://pith.science/api/pith-number/C43N4527I7EBNA4U3P6FEMSGNT/graph.json","events_json":"https://pith.science/api/pith-number/C43N4527I7EBNA4U3P6FEMSGNT/events.json","paper":"https://pith.science/paper/C43N4527"},"agent_actions":{"view_html":"https://pith.science/pith/C43N4527I7EBNA4U3P6FEMSGNT","download_json":"https://pith.science/pith/C43N4527I7EBNA4U3P6FEMSGNT.json","view_paper":"https://pith.science/paper/C43N4527","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.11053&json=true","fetch_graph":"https://pith.science/api/pith-number/C43N4527I7EBNA4U3P6FEMSGNT/graph.json","fetch_events":"https://pith.science/api/pith-number/C43N4527I7EBNA4U3P6FEMSGNT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/C43N4527I7EBNA4U3P6FEMSGNT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/C43N4527I7EBNA4U3P6FEMSGNT/action/storage_attestation","attest_author":"https://pith.science/pith/C43N4527I7EBNA4U3P6FEMSGNT/action/author_attestation","sign_citation":"https://pith.science/pith/C43N4527I7EBNA4U3P6FEMSGNT/action/citation_signature","submit_replication":"https://pith.science/pith/C43N4527I7EBNA4U3P6FEMSGNT/action/replication_record"}},"created_at":"2026-05-25T02:01:22.999466+00:00","updated_at":"2026-05-25T02:01:22.999466+00:00"}