AGENTS.md is a project instruction file that gives AI coding agents task context, workflow rules, and constraints so agent behavior aligns with team standards.

Approval Modes are policy settings that control which operations an agent can run automatically and which require explicit human approval.

A System Prompt is a top-level instruction layer that sets goals, boundaries, and response style before user messages are processed.

Prompt Injection is an adversarial technique that embeds manipulative instructions in user or retrieved content to subvert model behavior.

An agent instruction is a rule or request that guides how an AI agent should plan, edit, format, or avoid certain actions.

Agent instructions shape the behavior of an AI assistant. They can describe tone, linking rules, CMS conventions, design constraints, or publishing boundaries so repeated work stays consistent across a project.

An agent trigger is an event or command that starts an automated AI action, such as updating content or checking a project.

Agent triggers can be manual, scheduled, or connected to outside systems. In a website workflow, a trigger might start a content import, run a QA pass, or ask an agent to update CMS entries after source data changes.

Conversational editing is the process of changing a website through natural-language feedback instead of direct manual edits alone.

Conversational editing lets a user describe what should change and have an agent update the project. It works best when each request names the target clearly and the agent keeps the result editable for follow-up refinements.