mlsanityData to 14 Oct 2026

PragLocker: Protecting Agent Intellectual Property in Untrusted Deployments via Non-Portable Prompts

arXiv 2605.05974 ↗First posted 7 May 2026cs.CR cs.AIcited by 0

On the map: LLM agents and tool use → Agent threats, guardrails and governance

Abstract (arXiv)

LLM agents rely on prompts to implement task-specific capabilities based on foundation LLMs, making agent prompts valuable intellectual property. However, in untrusted deployments, adversaries can copy and reuse these prompts with other proprietary LLMs, causing economic losses. To protect these prompts, we identify four key challenges: proactivity, runtime protection, usability, and non-portability that existing approaches fail to address. We present PragLocker, a prompt protection scheme that satisfies these requirements. PragLocker constructs function-preserving obfuscated prompts by anchoring semantics with code symbols and then using target-model feedback to inject noise, yielding prompts that only work on the target LLM. Experiments across multiple agent systems, datasets, and foundation LLMs show that PragLocker substantially reduces cross-LLM portability, maintains target performance, and remains robust against adaptive attackers.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Agent threats, guardrails and governance, one of 16 directions in LLM agents and tool use. Builds on 26 of its 53 references that are in the map. First posted on arXiv 7 May 2026.

Builds on

26 of 53 references.

References · 26

  1. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  2. 2021The Power of Scale for Parameter-Efficient Prompt Tuningcited by 3,795
  3. 2021Program Synthesis with Large Language Modelscited by 3,637
  4. 2023A Survey on Large Language Model based Autonomous Agentscited by 2,242
  5. 2025A-MEM: Agentic Memory for LLM Agentscited by 1,063
  6. 2024Evaluating Very Long-Term Conversational Memory of LLM Agentscited by 830
  7. 2017The NarrativeQA Reading Comprehension Challengecited by 597
  8. 2017Evolution Strategies as a Scalable Alternative to Reinforcement Learningcited by 574
  9. 2024Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attackscited by 486
  10. 2024A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applicationscited by 431
  11. 2022RLPrompt: Optimizing Discrete Text Prompts with Reinforcement Learningcited by 366
  12. 2021QuALITY: Question Answering with Long Input Texts, Yes!cited by 208
14 more
  1. 2024A Human-Inspired Reading Agent with Gist Memory of Very Long Contextscited by 146
  2. 2024PLeak: Prompt Leaking Attacks against Large Language Model Applicationscited by 127
  3. 2024Agent-Pro: Learning to Evolve via Policy-Level Reflection and Optimizationcited by 94
  4. 2024Prompt Compression for Large Language Models: A Surveycited by 67
  5. 2024Efficient Prompting Methods for Large Language Models: A Surveycited by 47
  6. 2025Language Models are Injective and Hence Invertiblecited by 35
  7. 2024Your Large Language Models Are Leaving Fingerprintscited by 33
  8. 2023PromptCARE: Prompt Copyright Protection by Watermark Injection and Verificationcited by 32
  9. 2024Raccoon: Prompt Extraction Benchmark of LLM-Integrated Applicationscited by 16
  10. 2024PRSA: Prompt Stealing Attacks against Real-World Prompt Servicescited by 10
  11. 2024Prompt Obfuscation for Large Language Modelscited by 10
  12. 2024EmojiPrompt: Generative Prompt Obfuscation for Privacy-Preserving Communication with Cloud-based LLMscited by 9
  13. 2025PromptCOS: Towards Content-only System Prompt Copyright Auditing for LLMscited by 3
  14. 2024The Reversing Machine: Reconstructing Memory Assumptionscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026AgentVisor: Defending LLM Agents Against Prompt Injection via Semantic Virtualization
  2. 2026Securing LLM Agents Need Intent-to-Execution Integrity
  3. 2026Black-Box Skill Stealing Attack from Proprietary LLM Agents: An Empirical Study
  4. 2026AgentSecBench: Measuring Prompt Injection, Privacy Leakage, and Tool-Use Integrity in LLM Agents
  5. 2025Prompt Flow Integrity to Prevent Privilege Escalation in LLM Agents
  6. 2026Can Agents Trust Their Skills? Uncovering Unsafe Chains of Trust in Skill-Based LLM Agents
  7. 2026Reachability-Based Capability Confinement for LLM Agents under Indirect Prompt Injection
  8. 2026AgentFlow: A Flow-Centric Policy Language and Framework for Securing LLM Agent Systems