# NoFly > NoFly finds every AI agent, MCP server, and tool running in your org and shows which ones are risky before they cause a problem. NoFly is an AI agent security company. Its first product, the Discovery Scan, finds every AI agent, MCP server, and tool running in an organization and scores each one for risk. The NoFly list is a free, public risk score for MCP servers and agent tools. ## Pages - [Home](https://www.nofly.ai/): Discovery Scan overview - [The NoFly List](https://www.nofly.ai/nofly-list): check whether an MCP server is safe - [Glossary](https://www.nofly.ai/glossary): AI agent security terms - [Trust](https://www.nofly.ai/trust): platform security, data handling, compliance status - [Get a demo](https://www.nofly.ai/demo) ## Questions and answers ### What does a shadow AI discovery tool find? A shadow AI discovery tool finds the AI agents and tools people use at work without security sign-off. Many watch browser and SaaS traffic, which catches chat apps. NoFly also looks inside developer tooling, where coding agents like Claude Code and Cursor load MCP servers that can reach files, secrets, and the terminal. ### What is shadow AI? Shadow AI is any AI agent or tool people use at work without the security team knowing about it. A NoFly Discovery Scan finds these agents and the MCP servers they connect to, so you can decide what stays. ### What is an MCP server, and why is it a risk? An MCP server gives an AI agent new abilities, like reading files, running commands, or calling an API. A malicious or careless server can make an agent leak secrets or take actions nobody approved, and most security tools never see it because it happens inside the agent. ### How do I know if an MCP server is safe? Look at what it can reach (files, secrets, the terminal, the network), who publishes it, and whether it sends data anywhere it does not declare. The NoFly list runs these checks for you and gives each server a public risk score. ### Does the scan change anything on our machines? No. The Discovery Scan is read-only. It reports what it finds and leaves your agents, tools, and settings as they are. ### Which agents can NoFly find? Coding assistants such as Claude Code, Cursor, GitHub Copilot, Windsurf, and OpenAI Codex, plus internal agents and bots that connect to MCP servers. ### Is the NoFly list free? Yes. Anyone can look up an MCP server on the NoFly list and see its risk score, with no signup and no install. ### What is the NoFly list? A public reputation list for MCP servers and agent tools. Each entry has a risk score and a plain explanation of what the server can reach and why it is or is not safe to load. ### How is a risk score worked out? NoFly looks at what a server can access, what it actually does when connected, who publishes it, and whether it sends data to hosts it does not declare. Untrusted servers are only ever connected inside a disposable, locked-down sandbox. ### Is this MCP server safe if it is popular on GitHub? Not necessarily. Stars and downloads measure popularity, not behavior, and a server can change between versions. The NoFly list scores what a server can reach and what it does when connected, so check the current version before your agents load it. ### Can I get my MCP server reviewed? Yes. Send us the server name or repository and we will score it. If your server is listed and you think the score is wrong, tell us and we will look again. ## Glossary - **AI agent**: Software that uses a language model to plan and take actions on its own, such as editing code, running commands, or calling other services. Coding assistants like Claude Code and Cursor are AI agents. - **Model Context Protocol (MCP)**: An open standard that lets AI agents connect to outside tools and data through a common interface. It is how most agents gain new abilities today. - **MCP server**: A program that exposes tools to an AI agent over the Model Context Protocol. Examples include servers that read files, query a database, or post to a chat app. Each one widens what the agent can reach. - **Shadow AI**: AI agents and tools used inside a company without the security team knowing or approving them. Shadow AI is the agent version of shadow IT. - **Agent discovery**: Finding every AI agent and MCP server running across an organization, along with who runs it and what it can access. Discovery is the first step in securing agents. - **Prompt injection**: An attack that hides instructions inside content an agent reads, such as a web page, file, or tool response, so the agent follows the attacker instead of the user. - **Tool poisoning**: A prompt injection placed in an MCP tool description or response. The agent trusts the tool, so the hidden instructions can make it leak data or run commands. - **Data exfiltration**: Moving data out of an organization without permission. For agents this often means a tool reading secrets like an .env file and sending them to an outside host. - **AI bill of materials (AI-BOM)**: An inventory of the AI models, agents, and tools a system depends on. It does for AI what a software bill of materials does for code libraries. - **AI security posture management (AI-SPM)**: The practice of continuously finding, scoring, and fixing the security risks in an organization’s AI agents, models, and tools. - **NoFly list**: NoFly’s public reputation list of MCP servers and agent tools. Each entry has a risk score explaining whether it is safe for an agent to load.