ToolPrivacyBench is a new benchmark that evaluates purpose-bound privacy over-disclosure in multi-tool LLM agent trajectories by auditing tool arguments against policy knowledge bases across 2,150 cases.
Title resolution pending
14 Pith papers cite this work, alongside 607 external citations. Polarity classification is still indexing.
14
Pith papers citing it
607
external citations · Crossref
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
ProcRoute enforces process-scoped authorization for split-tunnel routes on Linux using cgroups and eBPF, blocking unauthorized internal access while matching baseline VPN performance.
NLAC architecture translates natural language requests to access policies via LLMs, with embedding-based subgraph selection enabling up to 98.7% accuracy on large networks per NLACBench evaluations.
Identifies five attack classes specific to agentic cyber-physical systems and proposes ZTPM with 25 typed primitives across five domains plus Physical Impact Tiers, motivated by 60-trace evidence of model-dependent non-deterministic actuation.
The paper reframes manufacturing ransomware recovery as an interdependency problem, identifies nine evidence-backed failure modes from a multivocal review, and defines Minimum Viable Factory Recovery as an analytical objective for resuming minimal safe operations.
A threat-oriented digital twinning methodology and open-source modular twin is introduced for security evaluation of autonomous platforms, translating threat analysis into controllable tests for spoofing, replay, and adversarial ML attacks.
A first multivocal literature review finds academia modifies IoT to meet Zero Trust rules while industry integrates IoT into existing NIST-guided Zero Trust frameworks, exposing gaps in socio-technical and cost-benefit analysis.
HBHC protocol binds hierarchical credentials to heartbeat proofs for deterministic bounded-time revocation in AI agent swarms without network round-trips.
A rubric is proposed for zero-trust assurance of O-RAN RIC app supply chains, using threat modeling, WG11-aligned mappings, and evidence-based onboarding levels.
The paper proposes an operator-first 6G model with a three-layer Control Compact and five-tier Guarantee Economy, claiming feasibility from Rakuten Mobile's profitable Open RAN deployment.
AI Trust OS is a proposed always-on operating layer that discovers undocumented AI systems via telemetry and produces continuous zero-trust compliance artifacts for regulations including ISO 42001, EU AI Act, SOC 2, GDPR, and HIPAA.
Network segmentedness is 1 minus edge density; 97 random node pairs yield a 95% CI of ±0.1 via the standard binomial sample-size formula.
A literature review of industry publications on Zero Trust for IoT that identifies trends and challenges in real-world integration.
Proposes Triple-Identity Authentication by combining credentials with IMEI/IMSI and using a custom hash algorithm for decentralized verification at login and server points.
citing papers explorer
-
ToolPrivacyBench: Benchmarking Purpose-Bound Privacy in Tool-Using LLM Agents
ToolPrivacyBench is a new benchmark that evaluates purpose-bound privacy over-disclosure in multi-tool LLM agent trajectories by auditing tool arguments against policy knowledge bases across 2,150 cases.
-
ProcRoute: Process-Scoped Authorization of Split-Tunnel Routes
ProcRoute enforces process-scoped authorization for split-tunnel routes on Linux using cgroups and eBPF, blocking unauthorized internal access while matching baseline VPN performance.
-
Natural Language Access Control (NLAC): From Help Desk Requests to Structured Policies
NLAC architecture translates natural language requests to access policies via LLMs, with embedding-based subgraph selection enabling up to 98.7% accuracy on large networks per NLACBench evaluations.
-
When Agents Control Robots: A Zero Trust Policy Model for Agentic Cyber-Physical Systems
Identifies five attack classes specific to agentic cyber-physical systems and proposes ZTPM with 25 typed primitives across five domains plus Physical Impact Tiers, motivated by 60-trace evidence of model-dependent non-deterministic actuation.
-
From Backup Restoration to Minimum Viable Factory Recovery: A Systematization of Ransomware Recovery in Manufacturing Systems
The paper reframes manufacturing ransomware recovery as an interdependency problem, identifies nine evidence-backed failure modes from a multivocal review, and defines Minimum Viable Factory Recovery as an analytical objective for resuming minimal safe operations.
-
Threat-Oriented Digital Twinning for Security Evaluation of Autonomous Platforms
A threat-oriented digital twinning methodology and open-source modular twin is introduced for security evaluation of autonomous platforms, translating threat analysis into controllable tests for spoofing, replay, and adversarial ML attacks.
-
Converging Zero Trust and IoT Security: A Multivocal Literature Review
A first multivocal literature review finds academia modifies IoT to meet Zero Trust rules while industry integrates IoT into existing NIST-guided Zero Trust frameworks, exposing gaps in socio-technical and cost-benefit analysis.
-
Heartbeat-Bound Hierarchical Credentials: Cryptographic Revocation for AI Agent Swarms
HBHC protocol binds hierarchical credentials to heartbeat proofs for deterministic bounded-time revocation in AI agent swarms without network round-trips.
-
Towards a Zero-Trust Supply-Chain Assurance Rubric for ORAN RIC Applications
A rubric is proposed for zero-trust assurance of O-RAN RIC app supply chains, using threat modeling, WG11-aligned mappings, and evidence-based onboarding levels.
-
Operator-Controlled 6G: From Connectivity Infrastructure to Guaranteed Digital Services
The paper proposes an operator-first 6G model with a three-layer Control Compact and five-tier Guarantee Economy, claiming feasibility from Rakuten Mobile's profitable Open RAN deployment.
-
AI Trust OS -- A Continuous Governance Framework for Autonomous AI Observability and Zero-Trust Compliance in Enterprise Environments
AI Trust OS is a proposed always-on operating layer that discovers undocumented AI systems via telemetry and produces continuous zero-trust compliance artifacts for regulations including ISO 42001, EU AI Act, SOC 2, GDPR, and HIPAA.
-
How segmented is my network?
Network segmentedness is 1 minus edge density; 97 random node pairs yield a 95% CI of ±0.1 via the standard binomial sample-size formula.
-
Zero Trust in the Context of IoT: Industrial Literature Review, Trends, and Challenges
A literature review of industry publications on Zero Trust for IoT that identifies trends and challenges in real-world integration.
-
Triple-Identity Authentication: The Future of Secure Access
Proposes Triple-Identity Authentication by combining credentials with IMEI/IMSI and using a custom hash algorithm for decentralized verification at login and server points.