One Context For All Your AI Coding Tools

Every AI coding tool wants its own configuration and instructions. Over time they drift apart and assistants start contradicting each other. AgentCTX lets you maintain project context once and generate aligned configs for every editor your team uses.

Install AgentCTX View quickstart

Generates configs for the assistants your team already uses

End config sprawl. Keep every AI tool on the same page

AgentCTX is for teams that rely on more than one AI coding assistant. You define how the project should be described, built, and protected in one maintained context. AgentCTX generates what each tool expects so standards stick without copy and paste between editors.

A single source of truth that syncs everywhere

Your context lives in one place you maintain. No more duplicating project rules in every AI editor. Just update your context, run AgentCTX, and each supported assistant stays aligned.

One context. Every assistant.

Sources you maintain

Rules, commands, agents, and metadata in .agenstra/

Generated everywhere

Automatic generation of Claude, Cursor, Copilot, and other tool configs

Same rules everywhere

Consistent instructions and standards across every editor and assistant

Simple workflow for any project

Most repositories already repeat the same guidance in several tool-specific places. AgentCTX replaces that loop with a short workflow you can use in every project, from install to fresh configs in minutes.

Install AgentCTX

Install AgentCTX with a single command.

curl -fsSL https://downloads.agenstra.com/agentctx/install.sh | bash

Add project context

Add a .agenstra/ directory with rules, commands, agents, and metadata for architecture, standards, workflows, and security.

Run AgentCTX

Run AgentCTX to generate tool specific config files for the AI coding tools you use.

agentctx

Commit or keep as artifacts

Commit the generated files if needed or let them stay as build artifacts.

Less drift, more alignment

AgentCTX helps teams treat project context as a first class asset. You capture architecture, standards, security, and documentation as linked knowledge assistants can rely on. When it stays current, every tool follows the same rules, people onboard faster, and you can add or swap editors without starting over.

Consistency

Every assistant follows the same rules and understands the same project story.

Architecture & domain

High level architecture and domain overview

Faster onboarding

New team members get clear guidance from .agenstra/ and consistent behavior from their tools.

Coding style

Coding styles, naming conventions, and review expectations

AgentCTX

One context, every assistant

Security

Security considerations, secrets handling, and what must never be changed automatically

Documentation

Links to documentation, ADRs, and design docs

Version control friendly

.agenstra/ lives in your repository with pull requests, reviews, and history.

Tool flexibility

You can add or swap AI tools without redefining everything from scratch.

Built for teams that code with AI every day

AgentCTX is a natural fit for everybody who:

Uses multiple AI coding tools across different editors
Works in monorepos or large codebases with many contributors
Cares about consistent standards and secure defaults
Wants AI to feel like part of the engineering system, not a personal side tool

Platform engineers

Standardize tool config across repos

Dev leads

Keep teams aligned on standards

Security-conscious teams

Govern what AI can see and do

Open source maintainers

One context for all contributors

Add AgentCTX to your next sprint

Start with one repository. Define enough context that assistants understand your architecture, standards, and how you ship. Generate configs for the tools your team already uses, gather feedback, and expand when the approach sticks.

Once your team feels the difference, you can roll out AgentCTX across repositories to standardize how AI coding tools understand your systems.

View installation guide Explore example context files