Overview

This section explains the ideas behind Syntic Code — how Amara turns a request into working changes, how you can extend its capabilities, and how it manages the context that shapes every response.

What you will learn

Start with how Syntic Code works to understand the agent loop: reading your repository, planning, editing, running commands, and reviewing the result. From there, extend Syntic Code surveys the ways you can add capabilities, from skills and hooks to MCP servers, plugins, and subagents.

The remaining pages cover the machinery that keeps sessions grounded. The .syntic/ directory is where your settings, rules, skills, commands, and memory live. The context window page explains what fills Amara’s working memory and how compaction keeps long sessions on track. Finally, prompt caching shows how reused context lowers latency and cost on the Syntic model.

How to read this section

You can read these pages in order for a complete mental model, or jump to whatever you need. If you are new, the first two pages give you enough to be productive. If you are configuring a project or a team, the .syntic/ directory page is the natural anchor. If you are tuning cost and speed, focus on the context window and prompt caching pages.

Each page is short and self-contained, so treat this section as a reference you return to as your usage grows.