Trust No AI: Prompt Injection Along The CIA Security Triad
On the map: LLM safety and security → Prompt injection and agent security
Abstract (arXiv)
The CIA security triad - Confidentiality, Integrity, and Availability - is a cornerstone of data and cybersecurity. With the emergence of large language model (LLM) applications, a new class of threat, known as prompt injection, was first identified in 2022. Since then, numerous real-world vulnerabilities and exploits have been documented in production LLM systems, including those from leading vendors like OpenAI, Microsoft, Anthropic and Google. This paper compiles real-world exploits and proof-of concept examples, based on the research conducted and publicly documented by the author, demonstrating how prompt injection undermines the CIA triad and poses ongoing risks to cybersecurity and AI systems at large.
Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Placed in Prompt injection and agent security, one of 12 directions in LLM safety and security. First posted on arXiv 8 Dec 2024.
Builds on
0 of 0 references.
No references here yet.
Built on by
Citing papers per year: 2025: 4; 2026: 4.
8 papers.
- 2025Advances and Challenges in Foundation Agents: From Brain-Inspired Intelligence to Evolutionary, Collaborative, and Safe Systemscited by 126
- 2026The Promptware Kill Chain: How Prompt Injections Gradually Evolved Into a Multistep Malware Delivery Mechanismcited by 9
- 2025Agentic LLM Framework for Adaptive Decision Discoursecited by 6
- 2025Logic layer Prompt Control Injection (LPCI): A Novel Security Vulnerability Class in Agentic Systemscited by 6
- 2025Enterprise AI Must Enforce Participant-Aware Access Controlcited by 3
- 2026Training a General Purpose Automated Red Teaming Modelcited by 0
- 2026Beware of Agentic Botnets: Scalable Untargeted Promptware Attacks via Universal and Transferable Adversarial HalluSquattingcited by 0
- 2026Workspace Topology as an Attack Vector in Agentic Coding Assistantscited by 0
- 2026Workspace Topology as an Attack Vector in Agentic Coding Assistants14 Aug 2026
- 2026Beware of Agentic Botnets: Scalable Untargeted Promptware Attacks via Universal and Transferable Adversarial HalluSquatting8 Jul 2026
- 2026Training a General Purpose Automated Red Teaming Model24 Apr 2026
- 2026The Promptware Kill Chain: How Prompt Injections Gradually Evolved Into a Multistep Malware Delivery Mechanism14 Jan 2026
- 2025Enterprise AI Must Enforce Participant-Aware Access Control18 Sep 2025
- 2025Logic layer Prompt Control Injection (LPCI): A Novel Security Vulnerability Class in Agentic Systems14 Jul 2025
- 2025Advances and Challenges in Foundation Agents: From Brain-Inspired Intelligence to Evolutionary, Collaborative, and Safe Systems31 Mar 2025
- 2025Agentic LLM Framework for Adaptive Decision Discourse16 Feb 2025
Similar, not cited
Close in wording, no citation either way
- 2025Prompt Injection 2.0: Hybrid AI Threats
- 2023Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection
- 2024Prompt Infection: LLM-to-LLM Prompt Injection within Multi-Agent Systems
- 2025Meta SecAlign: A Secure Foundation LLM Against Prompt Injection Attacks
- 2024Exploring Vulnerabilities and Protections in Large Language Models: A Survey
- 2025Cybersecurity AI: Hacking the AI Hackers via Prompt Injection
- 2024Systematically Analyzing Prompt Injection Vulnerabilities in Diverse LLM Architectures
- 2023Tensor Trust: Interpretable Prompt Injection Attacks from an Online Game