Feature Request: Expose Eclipse IDE as a Native MCP (Model Context Protocol) Server #2921

Feature Overview
Instead of building a proprietary AI chat interface into the Eclipse IDE, the Eclipse Foundation should provide official, built-in support for exposing Eclipse IDE capabilities as a Model Context Protocol (MCP) server.

This allows standalone terminal-based AI assistants and AI CLI tools (e.g., GitHub Copilot CLI, Claude Code, Codex, Antigravity CLI) running inside or alongside Eclipse to directly query, control, and inspect the IDE workspace.

Rationale & Key Use Case: AI-Driven Workflows & Method-Level Review
Developer workflows are shifting toward AI-assisted "vibe coding," but text-based AI diffs lack the structural context required for large codebase safety. Eclipse possesses best-in-class structural review tools that are far superior for verifying AI modifications.

Eclipse Compare view structural diffing. מקור: Stack Overflow

By exposing Eclipse via MCP, AI CLI tools can invoke native IDE actions as structured MCP Skills, creating a seamless loop between AI code generation and visual developer verification:

Advanced AI Skill Workflows
Exposing Eclipse via MCP enables powerful automated skills, such as:

Auto-Generate Commit & Stage Changes:

An AI CLI tool executes code edits across multiple files.

The AI auto-generates a structured commit message based on the AST diff.

An MCP tool call triggers Eclipse's Git Staging View, bringing focus directly to the staged changes for immediate visual inspection.

Two-Phase Smart Staging Strategy for Complex Projects:

Phase 1 (Low-Risk New Files): The AI separates newly created files from modified files, writes an appropriate commit message, and places new files into Git Staging. Since new files don't alter existing logic, committing them in complex codebases carries minimal risk.

Phase 2 (High-Risk Modified Files & Method Compare): The AI auto-generates a separate commit message for existing modified files and routes them directly to the Eclipse Git Staging / Compare View.

Visual Method Review: Instead of reviewing crude line-by-line diffs, the developer uses Eclipse’s Synchronize View / Compare Editor, which maps and highlights changes method-by-method. The developer can focus on specific AST nodes and methods affected by the AI before confirming the commit.

Proposed MCP Tool Capabilities
The native Eclipse MCP Server module should expose workspace capabilities through standard transport protocols (stdio / WebSockets):

1. Workspace & AST Inspection
   Expose project structures, active editor context, and Java/C++ AST nodes.

Stream real-time compiler errors, warnings, and marker logs directly to the AI CLI tool.

2. Git & Staging Automation (MCP Skills)
   eclipse/git/stage: Move specific files or AST modifications into the Git Staging area.

eclipse/git/openStagingView: Focus the native Git Staging Perspective in Eclipse.

eclipse/compare/openMethodCompare: Launch the Compare Editor targeting specific modified methods or structural diffs.

3. Execution & Testing
   Trigger Eclipse incremental builds and execute JUnit/TestNG test suites.

4. Configuration Options
   Provide a preference panel (Preferences -> AI / MCP Server) to manage connected CLI clients, toggle server status, and configure permission levels (e.g., read-only inspection vs. execution/staging privileges).

Expected Outcome
Out-of-the-box availability in future Eclipse IDE packages.

Positions Eclipse as the leading IDE for terminal-first AI development by uniting modern AI generation with unmatched AST-aware, method-level code review and staging tools.