agent-sync 1.2.0

Bidirectional synchronization tool for custom agents and settings between Claude Code and GitHub Copilot.

Agent Sync

A universal synchronization tool for custom agents, permissions, and slash commands (saved prompts) between Claude Code and GitHub Copilot.

Support for Codex and Gemini CLI coming soon!

Manage your configuration in your preferred tool's native settings files, and automatically sync and convert those changes to all your other supported AI coding agents.

Features

• Bidirectional Sync: Automatically syncs changes in both directions between Claude and Copilot configurations.
• Permission Management: Translates and syncs permission configurations (Claude settings.json ↔ Copilot .perm.json).
• Slash Commands: Syncs slash command definitions and prompt files.
• Smart Conflict Resolution: Detects conflicts and offers interactive or automatic resolution strategies.
• State Tracking: Intelligently tracks file modifications and deletions to keep directories clean.
• Format Conversion: Seamlessly converts between format-specific schemas (e.g., Markdown structure, YAML frontmatter).
• Dry-Run Mode: Preview changes safely before applying them to your file system.

Installation

The recommended way to install Agent Sync is via pip:

pip install agent-sync

Building from Source

If you prefer to build from source:

Requirements: Python 3.8+

# Clone the repository
git clone https://github.com/ZacheryGlass/agent-sync.git
cd agent-sync

# Install dependencies
pip install -r requirements.txt

# Run via module
python -m cli.main

Usage

Agent Sync is a Command Line Interface (CLI) tool.

CLI Mode

The general syntax for the CLI is:

agent-sync [options]

CLI Options

The tool offers various flags to customize the synchronization process.

Core Configuration

• --source-dir: Specifies the directory containing your source configuration files.
• --target-dir: Specifies the directory where files should be synced to.
• --source-format: Defines the format of the source files (claude or copilot).
• --target-format: Defines the format for the destination files (claude or copilot).
• --config-type: Determines what type of data is being synced. Options are:
  • agent: For AI agent definitions.
  • permission: For tool use permissions and settings.
  • slash-command: For prompt and command definitions.
• --direction: Controls the synchronization flow.
  • both: Bidirectional sync (default).
  • source-to-target: One-way sync from source to target.
  • target-to-source: One-way sync from target to source.

Operation Control

• --dry-run: Simulates the operation and prints what would happen without modifying any files.
• --force: Automatically resolves conflicts by choosing the newest file, bypassing interactive prompts.
• --state-file: Path to a custom state file (defaults to ~/.agent_sync_state.json). This file tracks sync history.
• --verbose, -v: Enables detailed logging output for debugging.

Single File Operations

• --convert-file: Path to a single file to convert. Mutually exclusive with directory options.
• --output: Destination path for the single converted file.
• --sync-file & --target-file: Used for in-place merging of two specific files.

Format-Specific Flags

• --add-argument-hint: Adds an argument-hint field (useful for Copilot) based on the description when converting from Claude.
• --add-handoffs: Adds a handoffs placeholder field when converting to Copilot format.

Configuration Details

File Matching Strategy

Files are matched between formats based on their base names:

• Agents: planner.md (Claude) ↔ planner.agent.md (Copilot)
• Permissions: settings.json (Claude) ↔ settings.perm.json (Copilot)
• Slash Commands: command.md (Claude) ↔ command.prompt.md (Copilot)

Field Mapping

Claude → Copilot

Claude Field	Copilot Field	Notes
name	name	Direct mapping
description	description	Direct mapping
description	argument-hint	Optional (requires --add-argument-hint)
tools	tools	Converts comma-separated string to array
model	model	Maps model names (e.g., sonnet → Claude Sonnet 4)
permissionMode	-	Dropped (handled via Permission Sync)

Copilot → Claude

Copilot Field	Claude Field	Notes
name	name	Direct mapping
description	description	Direct mapping
tools	tools	Converts array to comma-separated string
model	model	Maps model names (e.g., Claude Sonnet 4 → sonnet)
argument-hint	-	Dropped

Permission Conversion

The tool handles complex logic to translate between Claude's permission system and VS Code's (Copilot) permission structure.

VS Code (Copilot) uses specific boolean flags for commands and URLs (e.g., "chat.tools.terminal.autoApprove"). Claude Code uses categories (allow, ask, deny) with specific patterns.

• Auto-Approve: Maps between VS Code true and Claude allow.
• Require Approval: Maps between VS Code false and Claude ask.
• Regex Patterns: Preserved as-is during conversion.
• Lossy Conversions: Since VS Code does not support a hard "deny" (block) state, Claude deny rules are converted to "require approval" in VS Code, and a warning is logged.

Contributing

Contributions are welcome! Please submit a Pull Request or open an issue to discuss proposed changes.

1. Fork the repository.
2. Create your feature branch (git checkout -b feature/amazing-feature).
3. Commit your changes (git commit -m 'Add some amazing feature').
4. Push to the branch (git push origin feature/amazing-feature).
5. Open a Pull Request.

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.