rulesync 3.11.4 → 3.12.1

package/README.md

@@ -60,7 +60,7 @@ Rulesync supports both **generation** and **import** for All of the major AI cod
 | Tool                  | rules | ignore | mcp   | commands | subagents |
 |------------------------|:-----:|:------:|:-----:|:--------:|:---------:|
 | AGENTS.md            |  ✅   |      |       |     🎮     |      🎮     |
-| Claude Code            |  ✅ 🌏   |  ✅   |  ✅ 🌏   |    ✅ 🌏     |    ✅ 🌏     |
+| Claude Code            |  ✅ 🌏   |  ✅   |  ✅ 🌏 📦   |    ✅ 🌏     |    ✅ 🌏     |
 | Codex CLI              |  ✅ 🌏   |      |   🌏   |     🌏    |    🎮      |
 | Gemini CLI             |  ✅ 🌏  |   ✅   |  ✅ 🌏  |     ✅ 🌏  |      🎮     |
 | GitHub Copilot         |  ✅    |       |  ✅    |     ✅     |    🎮      |
@@ -79,6 +79,7 @@ Rulesync supports both **generation** and **import** for All of the major AI cod
 * ✅: Supports project mode
 * 🌏: Supports global mode
 * 🎮: Supports simulated commands/subagents (Project mode only)
+* 📦: Supports modular MCP (Experimental)
 
 ## Why Rulesync?
 
@@ -148,7 +149,7 @@ Example:
   // Basically, you can specify a `["."]` only.
   // However, for example, if your project is a monorepo and you have to launch the AI agent at each package directory, you can specify multiple base directories.
   "baseDirs": ["."],
-  
+
   // Delete existing files before generating
   "delete": true,
 
@@ -158,7 +159,8 @@ Example:
   // Advanced options
   "global": false,  // Generate for global(user scope) configuration files
   "simulatedCommands": false,  // Generate simulated commands
-  "simulatedSubagents": false  // Generate simulated subagents
+  "simulatedSubagents": false,  // Generate simulated subagents
+  "modularMcp": false  // Enable modular-mcp for context compression (experimental, Claude Code only)
 
   // Deprecated experimental options (for backward compatibility)
   // "experimentalGlobal": false,
@@ -248,6 +250,7 @@ Example:
 {
   "mcpServers": {
     "serena": {
+      "description": "Code analysis and semantic search MCP server",
       "type": "stdio",
       "command": "uvx",
       "args": [
@@ -265,6 +268,7 @@ Example:
       "env": {}
     },
     "context7": {
+      "description": "Library documentation search server",
       "type": "stdio",
       "command": "npx",
       "args": [
@@ -277,6 +281,37 @@ Example:
 }
 ```
 
+#### Modular MCP Support (Experimental)
+
+Rulesync supports [modular-mcp](https://github.com/d-kimuson/modular-mcp) for context compression. When enabled with `--modular-mcp`, it generates both `.mcp.json` and `modular-mcp.json` files for Claude Code.
+
+**Benefits:**
+- Reduces context usage by loading tool schemas on-demand
+- Better scalability when using many MCP servers
+- Improved performance with large tool collections
+
+**Requirements:**
+- Each MCP server must have a `description` field (required when `modularMcp` is enabled)
+- Currently only supported for Claude Code
+
+**Usage:**
+
+```bash
+# Enable modular-mcp via CLI
+npx rulesync generate --targets claudecode --features mcp --modular-mcp
+
+# Or via configuration file
+{
+  "modularMcp": true
+}
+```
+
+**Generated files:**
+- `.mcp.json` - Contains modular-mcp proxy configuration
+- `modular-mcp.json` - Contains actual MCP server configurations
+
+For more information, see the [modular-mcp documentation](https://github.com/d-kimuson/modular-mcp).
+
 ### `.rulesyncignore`
 
 Example: