SUDP: Secret-Use Delegation Protocol for Agentic Systems
cs.CR, cs.AI
Submitted: 2026-04-27
Updated: 2026-10-21
Code: https://github.com/xhyumiracle/sudp
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- StruQ: Defending Against Prompt Injection with Structured Queries
- SecAlign: Defending Against Prompt Injection with Preference Optimization
- How Your Credentials Are Leaked by LLM Agent Skills: An Empirical Study
- LlamaFirewall: An open source guardrail system for building secure AI agents
- HDP: A Lightweight Cryptographic Protocol for Human Delegation Provenance in Agentic AI Systems
- Defeating Prompt Injections by Design
- Taming OpenClaw: Security Analysis and Mitigation of Autonomous LLM Agent Threats
- AgentLeak: A Benchmark for Internal-Channel Privacy Leakage in Multi-Agent LLM Systems
- Agentic JWT: A Secure Delegation Protocol for Autonomous AI Agents
- Defending Against Indirect Prompt Injection Attacks With Spotlighting
- The Attack and Defense Landscape of Agentic AI: A Comprehensive Survey
- Security Considerations for Artificial Intelligence Agents
- VIGIL: Defending LLM Agents Against Tool Stream Injection via Verify-Before-Commit
- ClawKeeper: Comprehensive Safety Protection for OpenClaw Agents Through Skills, Plugins, and Watchers
- OpenID Connect for Agents (OIDC-A) 1.0: A Standard Extension for LLM-Based Agent Identity and Authorization
- The Attacker Moves Second: Stronger Adaptive Attacks Bypass Defenses Against Llm Jailbreaks and Prompt Injections
- GPT-4 Technical Report
- Agent Tools Orchestration Leaks More: Dataset, Benchmark, and Mitigation
- Identifying the Risks of LM Agents with an LM-Emulated Sandbox
- Progent: Securing AI Agents with Privilege Control
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
- Estimating the Social Cost of Corporate Data Breaches
- Efficient Fuzzy PSI under One-Sided Assumptions
- Sealing the Audit-Runtime Gap for LLM Skills