Skip to content

Agent Configurations

This project distributes coding standards to multiple AI coding assistants simultaneously. Each agent has its own configuration format and file location.

Location: Project root (.cursorrules)

Cursor reads this file automatically when opening a project. It contains instructions that guide Cursor’s AI responses, including references to all standards documents.

Format: Plain text with markdown. Cursor treats the entire file as system-level context for its AI features.

What’s included:

  • References to all architecture, language, and process standards
  • Instructions for Clean Architecture enforcement
  • Code review and PR generation guidelines

GitHub Copilot — .github/copilot-instructions.md

Section titled “GitHub Copilot — .github/copilot-instructions.md”

Location: .github/copilot-instructions.md

GitHub Copilot reads this file as custom instructions for code generation and chat. Available in VS Code, JetBrains, and Neovim with Copilot.

Format: Markdown. Copilot uses the content as additional context when generating suggestions.

What’s included:

  • Core standards summary (Clean Architecture, SOLID, naming conventions)
  • Language-specific guidelines for supported languages
  • Testing and error handling requirements

Location: Project root (.aiderrc)

Aider reads this YAML configuration file on startup. It controls model settings, conventions, and architectural rules.

Format: YAML with specific Aider configuration keys.

Key settings:

  • conventions — coding style and architecture rules
  • architect-mode — enables architecture-aware suggestions
  • read — references to standards files Aider should be aware of

Location: Project root (.codexrc)

OpenAI Codex CLI reads this configuration file for project-specific instructions.

Format: YAML configuration.

What’s included:

  • Model and approval mode settings
  • Instructions referencing core standards
  • Architecture and testing requirements

Location: .gemini/GEMINI.md and .gemini/settings.json

Gemini CLI and Google Antigravity read the .gemini/ directory for repository intelligence and configuration.

Files:

  • GEMINI.md — Repository context document. Contains project overview, architecture decisions, agent protocols, and standards references. Acts as the primary context for Gemini’s understanding of the codebase.
  • settings.json — CLI configuration including model selection, checkpointing, and MCP server settings.

To add support for a new AI coding assistant:

  1. Create the agent’s config file under standards/agents/<agent-name>/
  2. Add the config to the setup script (scripts/setup.sh) so it gets copied during installation
  3. Update the sync script (scripts/sync-standards.sh) to keep it updated
  4. Document the new agent in this file