coder-config

🥈Silver

Manage configurations for AI coding tools like Claude Code, Gemini CLI, Codex CLI, and Antigravity. Set up MCPs, rules, permissions, memory, and workstreams. Ideal for operations teams automating AI developer workflows.

4320Updated 2mo ago

Intermediate30min to implementautomation

Saves ~60 min per use

Quick InstallView Source

git clone https://github.com/regression-io/coder-config.git

Works with:

Claude

Overview

About This Skill

Coder Config simplifies management of AI coding tool configurations across multiple projects and repositories. It provides a unified MCP registry, hierarchical rules system, persistent memory, and workstream grouping—eliminating the need to reconfigure tools or re-explain project context with each session. The tool generates outputs compatible with Claude Code, Gemini CLI, Codex CLI, and Antigravity from a single configuration. A web UI (localhost:3333) offers visual management of MCPs, rules, memory, and workstreams, while CLI commands enable project initialization, MCP toggling, and multi-repo coordination. Ideal for operations teams, development teams managing microservices or monorepos, and organizations standardizing AI developer workflows.

How to Use

[{"step":"Identify the AI coding tool you're using (e.g., Claude Code, Gemini CLI) and the project name.","tip":"Use the tool’s documentation to confirm supported configuration formats. For example, Claude Code uses `.claude` files, while Gemini CLI uses `gemini-config.yaml`."},{"step":"List the configurations you need to apply. Include memory limits, permission rules, MCP integrations, and workstreams.","tip":"Start with a small subset of configurations (e.g., memory and permissions) and expand incrementally. Use the tool’s `--dry-run` flag if available to preview changes."},{"step":"Run the configuration command or update the relevant config file in your project directory.","tip":"Back up your existing configuration files before making changes. For CLI tools, use `tool-name --config-path` to specify a custom config file location."},{"step":"Validate the configuration using the tool’s built-in validation command (e.g., `claude code --validate-config`).","tip":"Check the tool’s logs for errors or warnings. Common issues include permission conflicts or MCP server connection failures."},{"step":"Test the configuration by performing a task (e.g., generating code, running a script) and verify that the settings work as expected.","tip":"Use a staging environment or a small test project to validate configurations before applying them to production projects."}]

Use Cases

Managing MCPs and configurations across multiple microservices in a monorepo

Maintaining consistent coding rules and conventions globally while allowing project-specific overrides

Persisting coding preferences and corrections that Claude remembers across sessions

Grouping related repositories into workstreams so AI knows project relationships

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/regression-io/coder-config

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Configure [TOOL_NAME] for [PROJECT_NAME] with the following specifications: [LIST_OF_CONFIGURATIONS]. Ensure the setup includes [SPECIFIC_FEATURES] like memory management, permission rules, and MCP integrations. Validate the configuration against [CHECKLIST_ITEMS] and provide a summary of changes made. Tool: [TOOL_NAME], Project: [PROJECT_NAME], Configurations: [LIST_OF_CONFIGURATIONS], Features: [SPECIFIC_FEATURES], Checklist: [CHECKLIST_ITEMS]

Example Output

### Configuration Summary for Project 'Acme-AI-Dev' with Claude Code

**Changes Applied:**
1. **Memory Management:** Enabled persistent memory for the project with a 500MB limit and auto-prune settings for files older than 30 days. The system now retains context for the last 500 interactions to improve code generation accuracy.

2. **Permission Rules:** Configured role-based access with the following permissions:
- **Developers:** Read/write access to `/src` and `/tests`, execute permissions for `npm run dev`.
- **QA Team:** Read-only access to `/src` and `/tests`, execute permissions for `npm run test`.
- **Admins:** Full access to all directories, including `/config` and `/scripts`.

3. **MCP Integrations:** Added the following Model Context Protocol (MCP) servers:
- **GitHub MCP:** Configured with a PAT token for repository access. Enabled auto-pull for PRs and issue tracking.
- **Jira MCP:** Integrated with the team's Jira board. Configured to auto-sync tickets with the `dev` label.
- **Docker MCP:** Set up to manage local Docker containers for the project’s microservices.

4. **Workstreams:** Defined three workstreams for the team:
- **Feature Development:** Automatically assigns new tickets to the `dev` branch with a pre-commit hook for linting.
- **Bug Fixes:** Routes critical bugs to the `hotfix` branch with a mandatory review by the QA team.
- **Documentation:** Syncs updates to the `/docs` directory with Confluence via the MCP.

**Validation Results:**
- All permissions tested successfully. Developers can push to `/src` but cannot modify `/config`.
- Memory usage is stable at 420MB (within the 500MB limit).
- MCP servers are responsive: GitHub (1.2s response), Jira (0.8s response), Docker (0.5s response).

**Next Steps:**
Run `claude code --validate-config` to ensure no conflicts exist. Monitor memory usage over the next 24 hours to confirm auto-pruning works as expected.