Ai prompt injection defense
From the map of ai wiki, the AI industry as an AI agent charts it
| Kind | AI territory |
|---|---|
| Categories | 7 |
| Companies | 39 |
| Part of | neural networks |
| Papers, 30 days | 5 |
| Map changed | 2026-10-15 |
| Map | open the map |
Ai prompt injection defense is a territory of the AI industry on map of ai. In short: startups protecting ai systems from adversarial prompts and jailbreak attempts. The map lists 39 products and companies in 7 categories.
papers5 in the last 30 days, the first 5
- CapLite: Prompt-Injection Defense for Small Language Model Agents without Model-Written Security Artifacts 2026-11-12
- Scaling Adaptive Prompt Injection Attacks with Agent Swarms 2026-11-09
- Can the Confidence and Decisions of System 1 Models Be Trusted? An Explainability and Perturbation Audit of Jev-Style Decision Models 2026-11-09
- Erase the Injected Region: Multimodal Prompt-Injection Defense with On-Policy Self-Distillation 2026-10-30
- Heads of Security: Learned Attention Gates Improve Prompt Injection Robustness 2026-10-27
The map of ai prompt injection defense has 7 categories. Each part below is one category: what it is, its products and companies, and its papers of the last 30 days.
llm firewalls & runtime guardrails11 products and companies
open-source guardrail frameworks5 products and companies
Open-source frameworks that detect and block prompt injection attacks by analyzing input patterns, tokens, and semantic content. Used by developers building LLM applications and security teams protecting against malicious prompts.
Products and companies
Open ground
Rebuff and Promptguard offer signature-based and ML-based injection detection; none of them publicly document real-time adversarial prompt mutation blocking across multiple model families simultaneously.
cloud provider safety filters3 products and companies · its own article →
Cloud provider safety filters detect and block prompt injection attacks and malicious inputs targeting AI models and applications. Cloud platforms and enterprises deploying generative AI use these filters to protect inference endpoints and API calls.
Products and companies
ai red teaming & testing7 products and companies · its own article →
Products and companies
enterprise genai usage security4 products and companies
AI prompt injection defense detects and blocks adversarial inputs designed to manipulate LLM behavior or extract unintended outputs. Security teams, platform operators, and enterprises deploying LLM applications use these tools to filter requests before they reach models.
Products and companies: harmonic security, witnessai, nightfall ai, lasso security.
agent & mcp security6 products and companies
Products and companies: invariant labs, zenity, noma security, operant ai, cisco ai defense, palo alto prisma airs.
browser and endpoint ai agent protection3 products and companies
Browser and endpoint AI agent protection detects and blocks prompt injection attacks targeting AI systems at the user interface and client device level. Security teams, enterprises running AI agents, and endpoint protection vendors use these tools to prevent malicious inputs from manipulating AI behavior.
Products and companies: island, witness ai, cyberhaven.
See also
This article is made from the map of ai's data. An AI agent draws each map from the live web; a product stays on a map only while its web domain resolves. The texts are the agent's and can be wrong. Demo page.