{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:XZ5RJ2FZF4EZSMLSKBPAOWA4G6","short_pith_number":"pith:XZ5RJ2FZ","schema_version":"1.0","canonical_sha256":"be7b14e8b92f09993172505e07581c37a1a080657d62ef41b4b72ebe98dbe6b3","source":{"kind":"arxiv","id":"2505.03574","version":1},"attestation_state":"computed","paper":{"title":"LlamaFirewall: An open source guardrail system for building secure AI agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Abraham Montilla, Alekhya Gampa, Beto De Paola, Cyrus Nikolaidis, Daniel Song, David Molnar, Dominik Gabi, James Crnkovich, Jean-Christophe Testud, Joshua Saxe, Kat He, Lauren Deason, Nicholas Doucette, Rashnil Chaturvedi, Sahana Chennabasappa, Shengye Wan, Spencer Whitman, Stephanie Ding, Wu Zhou","submitted_at":"2025-05-06T14:34:21Z","abstract_excerpt":"Large language models (LLMs) have evolved from simple chatbots into autonomous agents capable of performing complex tasks such as editing production code, orchestrating workflows, and taking higher-stakes actions based on untrusted inputs like webpages and emails. These capabilities introduce new security risks that existing security measures, such as model fine-tuning or chatbot-focused guardrails, do not fully address. Given the higher stakes and the absence of deterministic solutions to mitigate these risks, there is a critical need for a real-time guardrail monitor to serve as a final laye"},"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":"2505.03574","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2025-05-06T14:34:21Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"4c9282d87b25cd6552675216de2c26115d4defd27961976b8820f18466c110ab","abstract_canon_sha256":"52a51777470df2c9178b32e10ae8d352e86254898376ca2c7da4b55c38a3cd7b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:59:18.621497Z","signature_b64":"ZrstmJTxJ8qvdU12vg8nwveUYQAID7TKqMZ9V0MnyOOR68HyQFT+Ci4H01EFGJnT3IN7VAQoF/+oJ+ywbMP5DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"be7b14e8b92f09993172505e07581c37a1a080657d62ef41b4b72ebe98dbe6b3","last_reissued_at":"2026-07-05T10:59:18.620854Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:59:18.620854Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LlamaFirewall: An open source guardrail system for building secure AI agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CR","authors_text":"Abraham Montilla, Alekhya Gampa, Beto De Paola, Cyrus Nikolaidis, Daniel Song, David Molnar, Dominik Gabi, James Crnkovich, Jean-Christophe Testud, Joshua Saxe, Kat He, Lauren Deason, Nicholas Doucette, Rashnil Chaturvedi, Sahana Chennabasappa, Shengye Wan, Spencer Whitman, Stephanie Ding, Wu Zhou","submitted_at":"2025-05-06T14:34:21Z","abstract_excerpt":"Large language models (LLMs) have evolved from simple chatbots into autonomous agents capable of performing complex tasks such as editing production code, orchestrating workflows, and taking higher-stakes actions based on untrusted inputs like webpages and emails. These capabilities introduce new security risks that existing security measures, such as model fine-tuning or chatbot-focused guardrails, do not fully address. Given the higher stakes and the absence of deterministic solutions to mitigate these risks, there is a critical need for a real-time guardrail monitor to serve as a final laye"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.03574","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/2505.03574/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":"2505.03574","created_at":"2026-07-05T10:59:18.620934+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.03574v1","created_at":"2026-07-05T10:59:18.620934+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.03574","created_at":"2026-07-05T10:59:18.620934+00:00"},{"alias_kind":"pith_short_12","alias_value":"XZ5RJ2FZF4EZ","created_at":"2026-07-05T10:59:18.620934+00:00"},{"alias_kind":"pith_short_16","alias_value":"XZ5RJ2FZF4EZSMLS","created_at":"2026-07-05T10:59:18.620934+00:00"},{"alias_kind":"pith_short_8","alias_value":"XZ5RJ2FZ","created_at":"2026-07-05T10:59:18.620934+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":32,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08395","citing_title":"Token-Flow Firewall: Semantic Runtime Auditing for Persistent AI Agents","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2606.25189","citing_title":"ActPlane: Programmable OS-Level Policy Enforcement for Agent Harnesses","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26627","citing_title":"Agents That Know Too Much: A Data-Centric Survey of Privacy in LLM Agents","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2606.22504","citing_title":"Lingering Authority: Revocable Resource-and-Effect Capabilities for Coding Agents","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21732","citing_title":"Safe to Check, Unsafe to Use: Relinking at the Compression Boundary of LLM Agents","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20560","citing_title":"How Transparent is DiffusionGemma?","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01277","citing_title":"Cognitive Firewall: A Proactive, Zero-Trust, Multi-Gate Framework for LLM Safety","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2606.17257","citing_title":"Pulling The REINS: Training-Free Safety Alignment of Video Diffusion Models via Representation Steering","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2606.14517","citing_title":"From Shield to Target: Denial-of-Service Attacks on LLM-Based Agent Guardrails","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12329","citing_title":"PROJECTMEM: A Local-First, Event-Sourced Memory and Judgment Layer for AI Coding Agents","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2606.25189","citing_title":"ActPlane: Programmable OS-Level Policy Enforcement for Agent Harnesses","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10907","citing_title":"Engineering Robustness into Personal Agents with the AI Workflow Store","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26373","citing_title":"Hybrid privacy-aware semantic search: SVD-truncated document geometry and CKKS-encrypted query reranking under a restricted threat model","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.27489","citing_title":"HARP: Measuring Harm Amplification in Multi-Agent LLM Systems","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30883","citing_title":"TRACE: Task-Aware Adaptive Self-Evolving Agentic Jailbreaking","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00448","citing_title":"When Safe Skills Collide: Measuring Compositional Risk in Agent Skill Ecosystems","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10749","citing_title":"Toward Secure LLM Agents: Threat Surfaces, Attacks, Defenses, and Evaluation","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2505.13766","citing_title":"A Blueprint for AI-Driven Software Quality: Integrating LLMs with Established Standards","ref_index":176,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16265","citing_title":"AgentWall: A Runtime Safety Layer for Local AI Agents","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17380","citing_title":"ADR: An Agentic Detection System for Enterprise Agentic AI Security","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19149","citing_title":"Agent Meltdowns: The Road to Hell Is Paved with Helpful Agents","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2605.19240","citing_title":"CASPIAN: Online Detection and Attribution of Cascade Attacks in LLM Multi-Agent Systems via Cross-Channel Causal Monitoring","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17413","citing_title":"Ablating Safety: Mechanisms for Removing Alignment in Language Models for Security Applications","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10907","citing_title":"Engineering Robustness into Personal Agents with the AI Workflow Store","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10907","citing_title":"Engineering Robustness into Personal Agents with the AI Workflow Store","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XZ5RJ2FZF4EZSMLSKBPAOWA4G6","json":"https://pith.science/pith/XZ5RJ2FZF4EZSMLSKBPAOWA4G6.json","graph_json":"https://pith.science/api/pith-number/XZ5RJ2FZF4EZSMLSKBPAOWA4G6/graph.json","events_json":"https://pith.science/api/pith-number/XZ5RJ2FZF4EZSMLSKBPAOWA4G6/events.json","paper":"https://pith.science/paper/XZ5RJ2FZ"},"agent_actions":{"view_html":"https://pith.science/pith/XZ5RJ2FZF4EZSMLSKBPAOWA4G6","download_json":"https://pith.science/pith/XZ5RJ2FZF4EZSMLSKBPAOWA4G6.json","view_paper":"https://pith.science/paper/XZ5RJ2FZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.03574&json=true","fetch_graph":"https://pith.science/api/pith-number/XZ5RJ2FZF4EZSMLSKBPAOWA4G6/graph.json","fetch_events":"https://pith.science/api/pith-number/XZ5RJ2FZF4EZSMLSKBPAOWA4G6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XZ5RJ2FZF4EZSMLSKBPAOWA4G6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XZ5RJ2FZF4EZSMLSKBPAOWA4G6/action/storage_attestation","attest_author":"https://pith.science/pith/XZ5RJ2FZF4EZSMLSKBPAOWA4G6/action/author_attestation","sign_citation":"https://pith.science/pith/XZ5RJ2FZF4EZSMLSKBPAOWA4G6/action/citation_signature","submit_replication":"https://pith.science/pith/XZ5RJ2FZF4EZSMLSKBPAOWA4G6/action/replication_record"}},"created_at":"2026-07-05T10:59:18.620934+00:00","updated_at":"2026-07-05T10:59:18.620934+00:00"}