Ajar: Measuring Open Privilege in Agent Defenses

arXiv:2609.26900 · cs.CR, cs.AI, cs.SE · Submitted 2026-09-22 · Read on arXiv

cs.CR, cs.AI, cs.SE

Submitted: 2026-09-22

Updated: 2026-09-22

Code: https://github.com/reSHARMA/Ajar

Terminology

• Benchmark — 81× in this paper · explained in 2 episodes
• Benchmarking — 5× in this paper · explained in 4 episodes
• Symposium on Operating Systems Principles — 3× in this paper
• AI Agents — 2× in this paper · explained in 2 episodes
• Core — 2× in this paper · explained in 2 episodes
• Provenance — 2× in this paper · explained in 4 episodes
• ACM Conference on Computer and Communications Security — 1× in this paper
• Ablation — 1× in this paper · explained in 1 episode
• Abstraction — 1× in this paper · explained in 1 episode
• Agent Skills — 1× in this paper · explained in 1 episode
• Formal Methods — 1× in this paper · explained in 3 episodes
• Formal Verification — 1× in this paper · explained in 2 episodes
• Fourteenth International Conference on Learning Representations — 1× in this paper
• In Formal Methods in Computer Aided Design — 1× in this paper
• Trace — 1× in this paper · explained in 2 episodes
• Verification — 1× in this paper · explained in 2 episodes

Sources

• Concrete Problems in AI Safety
• Defeating Prompt Injections by Design
• Defending Against Indirect Prompt Injection Attacks With Spotlighting
• Penalizing side effects using stepwise relative reachability
• FORTIS: Benchmarking Over-Privilege in Agent Skills
• Ignore Previous Prompt: Attack Techniques For Language Models
• AC4A: Access Control for Agents
• SPML: A DSL for Defending Language Models Against Prompt Attacks
• Progent: Securing AI Agents with Privilege Control
• Extending the Formalism and Theoretical Foundations of Cryptography to AI
• The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions
• When Lower Privileges Suffice: Investigating Over-Privileged Tool Selection in LLM Agents
• $\tau$-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains

Related papers

• SoK: AI-Augmented Binary Reversing
• Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
• Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
• Efficient Fuzzy PSI under One-Sided Assumptions
• Sealing the Audit-Runtime Gap for LLM Skills
• Token Composition: A Graph Based on EVM Logs