Prompt Injection

Your coding agent trusts the tools it reads. That trust is the vulnerability. When an attacker poisons a Sentry error report, a Datadog monitor alert, a PagerDuty incident, or a Jira ticket description with hidden prompt injection payloads, your agent doesn’t know the difference between a legitimate instruction and a hijack attempt. I’ve spent the last few months digging into this attack surface across the four most common integrations teams wire up to Claude Code, Cursor, and Codex. Here’s what I found and exactly how to fix it. ...

On June 25, 2026, the Mozilla 0DIN team demonstrated an attack that should change how every team deploys AI coding agents. They published a normal-looking Python repository on GitHub. A developer cloned it and pointed Claude Code at it. The agent read the README, installed the requirements, hit a routine initialization error, and — trying to be helpful — ran the suggested fix. That fix queried a DNS TXT record, decoded the value, and executed it as a shell command, opening a reverse shell on the developer’s machine. ...

Autonomous AI agents are executing real actions — writing code, querying databases, sending emails, and calling third-party APIs — and the security industry is finally treating them as the high-value attack surface they represent. The AI security market is projected to reach $12.8B by 2026 at a 28% CAGR, driven almost entirely by enterprise urgency around agent deployments. Unlike traditional software vulnerabilities, AI agent attacks are often semantic rather than syntactic: a well-crafted prompt in a retrieved document can silently redirect an agent’s entire task chain without triggering a single firewall rule. Security teams that treat agents like ordinary microservices will discover this difference the hard way. ...

MCP (Model Context Protocol) is now the de facto standard for connecting AI agents to external tools — but 43% of analyzed MCP servers are vulnerable to command injection, and over 2,000 internet-exposed servers were found leaking API keys in early 2026. This guide covers every major attack vector, real CVEs, and the exact controls you need before shipping to production. What Is MCP and Why Security Is Now a Developer Responsibility MCP (Model Context Protocol) is an open standard developed by Anthropic that gives AI agents a structured way to interact with external tools, APIs, filesystems, and databases through a uniform interface. Unlike a traditional REST API where a human decides which endpoint to call, MCP delegates tool selection and invocation to the AI agent itself — creating a radically different trust model that most existing security tooling was never designed to handle. As of mid-April 2026, over 9,400 public MCP servers exist with projections reaching 18,000 by year-end, and the MCP SDK has surpassed 97 million monthly downloads — a 970× increase in 18 months. 67% of CTOs surveyed in Q1 2026 say MCP is or will be their default agent-integration standard within 12 months. That velocity is exactly why security has become every developer’s problem: the attack surface is exploding faster than defenses are being built. In a traditional API integration, a developer writes code that calls a specific endpoint with known parameters. With MCP, a language model reads tool descriptions at runtime, decides which tools to call, interprets their outputs, and may chain multiple tools together — all without a human in the loop. Compromising any link in that chain can cascade silently across an entire session. ...

The threat surface for large language models has expanded beyond what most security teams anticipated three years ago. What began as a concern about chatbot misuse has evolved into a full-spectrum attack discipline targeting autonomous AI agents that browse the web, execute code, manage files, and call external APIs on behalf of users. This guide consolidates the current state of LLM red teaming as of 2026, covering the attack categories, specialized tooling, and operational processes that security teams need to protect AI-powered systems in production. ...