Skip to content
Open access

Agent-access receipts for artificial intelligence-mediated content and application programming interface access: a reproducible synthetic security experiment

Sep 2026 · Frontiers of Computer Science · Vol 8 · 0 citations · 12 references

TL;DR

A reproducible controlled experiment is presented showing why AI-mediated access-control evidence should preserve bindings among identity, delegation, declared purpose, resource/API inventory, economic authorization, policy freshness, sensitivity, and validation survivability.

Abstract

Artificial intelligence (AI)-mediated web and application programming interface (API) access is often treated as a bot-management problem—identify an automated caller and then allow, block, delay, challenge, or meter the request. That framing is useful but incomplete. A disputed access decision can depend on the caller’s authenticated identity, the represented principal, delegated authority, declared purpose, and resource or API surface. It can also depend on the current policy snapshot, response sensitivity, payment or licensing state, and the survival of validation evidence after the request has completed. This article presents a reproducible synthetic experiment using an agent-access receipt—a compact decision record generated at a publisher or API edge and preserved for later inspection when corresponding validation material remains available. The study implements a deterministic Python laboratory with a synthetic policy and synthetic trust registry, evaluating 15 scenario cases, 15 mutation/control cases, and 3 post-issue receipt-integrity controls. It also emits generated receipts in JavaScript Object Notation (JSON), a JSON Schema, and result tables in comma-separated values (CSV) and JSON formats. In the deposited v0.1.6 package, all 15 scenario cases, all 15 mutation cases, and all 3 integrity controls match expected outcomes, and all 30 generated receipts pass the required-field/schema-shape guard. The artifact is not a production security protocol; it does not claim legal compliance, deployment readiness, or endorsement by any third party. The article contributes a reproducible controlled experiment showing why AI-mediated access-control evidence should preserve bindings among identity, delegation, declared purpose, resource/API inventory, economic authorization, policy freshness, sensitivity, and validation survivability. The deposited artifact provides data and code for rerunning and inspecting the experiment.

Read PDF

Similar papers

Review Sep 2026

Authorization Architectures for Tool-Using AI Agents

A principal hierarchy spanning human user, operator/deployer, orchestrator agent, sub-agent, and tool endpoint as an organizing framework is introduced, and seven structural requirements are proposed, derived from a four-layer reference architecture, and runtime enforcement and aggregation bounds are identified as the...

Rakesh Kumar Surapani, Pradeep Kumar Dolabehera Kakitapelli, Arun Morampudi et al. · 1 citation
#large language models Open access Sep 2026

ARES: Securing Agents for Computer Use Through Endpoint Resource Mediation and Behavioral Guardrails

Large language model (LLM)-based agents are evolving into agents for computer use (ACUs) that read files, invoke applications, communicate over networks, and operate graphical interfaces, moving the effective security boundary from model inputs and outputs to autonomous actions that alter endpoint state. Conventional i...

Changhee Kim, Seong-je Cho · 0 citations
Sep 2026

Runtime Authorization for Resources Acquired by AI Agents

A provenance-bounded runtime authorization architecture that quarantines acquired outputs, resolves their actual capabilities from authenticated provider evidence through a versioned resolver, and activates them only through a current activation transaction that checks the resolved manifest, provenance, epochs, and a d...

Gen-Liang Zhu, Chu Wang · 0 citations
Open access Sep 2026

Runtime Policy Firewall: A Zero-Trust Governance Layer for Enterprise Agentic AI

Enterprise adoption of generative AI is shifting from passive question answering to autonomous agentic execution. Modern agents can decompose goals, retrieve business context, call tools, update records, send messages, initiate transactions, and coordinate workflows across multiple systems. This creates productivity op...

S. Ray · 0 citations
Sep 2026

MetaPermit: Scalable and Auditable Access Control for AI Agents via LLM-Inferred Meta-Attributes

The results show that MetaPermit produces 31% more consistent authorization decisions than LLM-driven authorization and outperforms the state-of-the-art defenses CaMeL and IPIGuard in both task completion and robustness to IPI attacks, with no malicious tool calls executed.

Han-Zhang Ma, Alicia Y. Hariri, Tian-Xiang Shen et al. · 0 citations
Review Open access Aug 2026

A RAG and Agentic AI Architecture for Automated Fixed-Access Network Investigation

This review examines how retrieval-augmented generation (RAG) and agentic artificial intelligence can be combined to automate investigation without allowing a language model to substitute plausible narrative for verified engineering evidence.

Rizwan Farooq Khan · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.