AI Agent
An AI agent is like a digital assistant you give permission to use your computer's tools, like opening files or browsing the web. If a hacker tricks this assistant, they can use it as a remote-control puppet to run commands on your machine without you knowing.
An AI agent is an autonomous system integrated into a host environment with tool-use capabilities, such as shell execution and network access. Adversaries can compromise these agents via prompt injection or configuration tampering to establish a command-and-control (C2) channel, leveraging the agent's legitimate permissions to execute malicious operations covertly.
An AI agent is a software entity utilizing an LLM as a reasoning engine to orchestrate tool invocation within a host environment. Adversaries exploit this architecture by manipulating the agent's instruction set or context—via techniques such as prompt injection, context poisoning, or configuration modification—to repurpose the agent as a C2 conduit. This enables the execution of arbitrary shell commands, exfiltration of data, and lateral movement, effectively weaponizing the agent's authorized access to system resources and external services.