What is Prompt Injection? Posted in Security Janet Wagner October 29, 2026 Prompt injection is currently the top risk when it comes to building applications driven by large language models (LLMs), like AI agents. That’s because prompts are the main control interface for these models, and they can follow text instructions embedded in user input or any external content the model reads as part of a given prompt. Without guardrails in place, an LLM may follow those instructions, whether authorized or unauthorized. Recent research has uncovered novel techniques for deploying prompt injections. Cybercriminals are actively exploring new types of prompt injection as well. This article explains what prompt injection is, its potential consequences, and how APIs expand the attack surface. It also includes a few tips on defending against it. What is Prompt Injection? Prompt injection is where an LLM (or an AI agent powered by one) is manipulated to produce unintended output or behavior. This occurs by prompting the model one or multiple times or the model reading externally sourced content that contains an injection — sometimes both. For…
What is Prompt Injection?
In short
The threat encompasses direct inputs from users as well as indirect commands hidden inside external media such as images, error logs, and websites. Malicious attacks can force models to leak confidential data or execute unwanted chains of programmatic actions. Concrete examples include proof-of-concept exploits that compromise AI-powered browsers or steal repository secrets via GitHub pull requests. Google researchers analyzing public web data tracked a rapid rise in these hidden web-based threats in recent months.
Key points
- Prompt injection currently stands as the leading security vulnerability affecting systems…: Prompt injection currently stands as the leading security vulnerability affecting systems and applications driven by large language models.
- Exploits can occur unintentionally when regular text mimics instructions, or intentionall…: Exploits can occur unintentionally when regular text mimics instructions, or intentionally through targeted tactics engineered by cybercriminals.
- Integrating external tools and APIs with language models widens vulnerability risks unles…: Integrating external tools and APIs with language models widens vulnerability risks unless developers implement appropriate security guardrails.