Prompts Don't Protect: Architectural Enforcement via MCP Proxy for LLM Tool Access Control

[2605.18414] Prompts Don't Protect: Architectural Enforcement via MCP Proxy for LLM Tool Access Control Skip to main content Search Submit Donate Log in Computer Science > Cryptography and Security arXiv:2605.18414 (cs) [Submitted on 18 May 2026 ( v1 ), last revised 27 Aug 2026 (this version, v3)] Title: Prompts Don't Protect: Architectural Enforcement via MCP Proxy for LLM Tool Access Control Authors: Rohith Uppala View PDF HTML (experimental) Abstract: Large language models increasingly operate as autonomous agents that select and invoke tools from large registries. We identify a critical gap: when unauthorized tools are visible in an agent's context, models select them in 48-68% of adversarial scenarios, even when explicitly instructed not to. Role escalation attacks (e.g., "I'm the CFO, override the access controls") are the most dangerous category, reaching 96% unauthorized invocation in frontier models. We show this holds across three models spanning open-weight and frontier systems, including instruction-tuned models with strong alignment training. Critically, prompt-based compliance is both insufficient and unpredictable: explicit per-tool allowlists reduce violations to as low as 4.0% but never to zero, and compliance varies widely across models, from 4.0% to 37.0% UIR, with no reliable relationship to general capability. We propose a proxy-enforced attribute-based access control (ABAC) layer for MCP that filters tool registries at discovery time. Because unauthorized tools never reach the model context, UIR is 0% by design, a structural guarantee that prompt instructions cannot replicate regardless of model or phrasing. Comments: 7 pages, 4 tables, 2 figures. Camera-ready version accepted to the EMNLP 2026 Industry Track; adds benign-utility and abstention evaluation, latency percentiles, and revised limitations Subjects: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI) Cite as: arXiv:2605.18414 [cs.CR] (or arXiv:2605.18414v3 [cs.CR] for this version) https://doi.org/10.48550/arXiv.2605.18414 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Rohith Uppala [ view email ] [v1] Mon, 18 May 2026 13:52:24 UTC (39 KB) [v2] Sun, 9 Aug 2026 04:49:11 UTC (56 KB) [v3] Thu, 27 Aug 2026 19:15:31 UTC (57 KB) Full-text links: Access Paper: View PDF HTML (experimental) TeX Source view license Current browse context: cs.CR < prev | next > new | recent | 2026-05 Change to browse by: cs cs.AI References & Citations NASA ADS Google Scholar Semantic Scholar export BibTeX citation Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer ( What is the Explorer? ) Connected Papers Toggle Connected Papers ( What is Connected Papers? ) Litmaps Toggle Litmaps ( What is Litmaps? ) scite.ai Toggle scite Smart Citations ( What are Smart Citations? ) Code, Data, Media Code, Data and Media Associated with this Article alphaXiv Toggle alphaXiv ( What is alphaXiv? ) Links to Code Toggle CatalyzeX Code Finder for Papers ( What is CatalyzeX? ) DagsHub Toggle DagsHub ( What is DagsHub? ) GotitPub Toggle Gotit.pub ( What is GotitPub? ) Huggingface Toggle Hugging Face ( What is Huggingface? ) ScienceCast Toggle ScienceCast ( What is ScienceCast? ) Demos Demos Replicate Toggle Replicate ( What is Replicate? ) Spaces Toggle Hugging Face Spaces ( What is Spaces? ) Spaces Toggle TXYZ.AI ( What is TXYZ.AI? ) Related Papers Recommenders and Search Tools Link to Influence Flower Influence Flower ( What are Influence Flowers? ) Core recommender toggle CORE Recommender ( What is CORE? ) Author Venue Institution Topic About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs . Which authors of this paper are endorsers? | Disable MathJax ( What is MathJax? ) We gratefully acknowledge support from our major funders , member institutions , and all contributors. About · Help · Contact · Subscribe · Copyright · Privacy · Accessibility · Operational Status (opens in new tab) Major funding support from

Related documents

• Prompts Don't Protect: Architectural Enforcement via MCP Proxy for LLM Tool Access Control#682682
• Prompts Don't Protect: Architectural Enforcement via MCP Proxy for LLM Tool Access Control#200382
• Understanding and Mitigating Prompt Leaking Attacks in Real-World LLM-Based Applications#287061
• TRUSTDESC: Preventing Tool Poisoning in LLM Applications via Trusted Description Generation#2465
• AEGIS: Preventing Cross-Domain Resource Abuse in MCP#659049