MCP Java Dev Tools
Bridges agentic coding tools and live Java runtime behavior through a lightweight sidecar agent.
mcp-java-dev-tools
MCP Java Dev Tools bridges agentic coding tools and live Java runtime behavior through a lightweight sidecar agent.
Static analysis only gets you so far. By attaching directly to a running JVM, this tool surfaces bytecode-level runtime signals that static analysis alone can't see — enabling probe-verified inspection, targeted regression checks, runtime-path validation, and deterministic debugging workflows.
The runtime agent is built with ByteBuddy and works alongside JDWP rather than replacing it. On top of the probe layer, the system adds framework-aware data synthesis and strict, fail-closed tool contracts — so agent orchestrators can make decisions grounded in actual runtime proof, not inference.
The current focus is HTTP entrypoints. Non-HTTP protocol support is on the horizon but not yet implemented — it will need concrete models and validation targets before the core contracts can be generalized.
For operator workflows and end-to-end execution flows, see docs/how-it-works/README.md.
Requirements
| Requirement | Version |
|---|---|
| Node.js | v24.13.0 (tested) |
| npm | 11.6.2 (tested) |
| JDK | 17+ |
| Maven | any recent |
Build
npm.cmd install
npm.cmd run build
mvn -f java-agent\pom.xml package
This produces two artifacts:
- MCP server →
dist/server.js - Java agent bundle →
java-agent/core/core-probe/target/mcp-java-dev-tools-agent-0.1.0-all.jar
Installation
Installer
The installer currently supports Codex and Kiro.
./scripts/install-integrations.sh
This installs the default skill set:
mcp-java-dev-tools-line-probe-runmcp-java-dev-tools-regression-suitemcp-java-dev-tools-issue-report
Manual Setup
Java Agent Setup
The target JVM must run on Java 17 or newer. If you're on Java 21, see Java 21 compatibility mode before continuing.
Add the following as a JVM argument when launching your application, replacing {desktopName} and the include package with your own:
-javaagent:C:\Users\{desktopName}\repository\mcp-java-dev-tools\java-agent\core\core-probe\target\mcp-java-dev-tools-agent-0.1.0.jar=host=0.0.0.0;port=9191;include=com.{package_to_capture}.**;exclude=com.nimbly.mcpjavadevtools.agent.**,**.config.**,**Test
Tip: The
includefilter is optional. By default, the agent instruments all classes within the module it was run from. Settingincludeto your root package (e.g.com.acme.**) keeps instrumentation focused and avoids capturing classes you don't care about.
To confirm the agent is instrumenting your classes, check the startup logs for lines like:
[mcp-probe]: com.yourpackagename.yourclassname
If you don't see your classes listed, check your include filter.
IntelliJ IDEA — Step by Step
- Open Run > Edit Configurations... from the top menu
- Select the run configuration for your target application (or create one if it doesn't exist)
- Expand the Modify options dropdown and enable Add VM options if it isn't already visible
- In the VM options field, paste the full
-javaagent:...argument from above - Click Apply, then OK
- Run your application normally — the agent attaches on startup
Finding the JAR path: If you're unsure of the absolute path, right-click the agent JAR in the Project panel and choose Copy Path > Absolute Path.
On Windows, use backslashes in the path (
C:\Users\...). On macOS/Linux, use forward slashes (/home/...or/Users/...).
Eclipse — Step by Step
- Go to Run > Run Configurations... (or Debug Configurations... if you're debugging)
- Select your application under Java Application, or create a new one
- Open the Arguments tab
- In the VM arguments field, paste the full
-javaagent:...argument from above - Click Apply, then Run (or Debug)
Finding the JAR path: Navigate to the JAR in your file system, right-click it, and copy the full path. Paste it into the agent argument, replacing the placeholder path.
On Windows, Eclipse accepts both forward and backslashes in paths, but backslashes are safer. Wrap the path in quotes if it contains spaces:
-javaagent:"C:\path with spaces\agent.jar"=...
Runtime Configuration
Java Agent Options
Capture History Buffer Size
Controls how many method captures the agent retains per probe point.
| Method | Value |
|---|---|
| Agent arg | captureMethodBufferSize=<1..32> |
| JVM property | -Dmcp.probe.capture.method.buffer.size=<1..32> |
| Environment variable | MCP_PROBE_CAPTURE_METHOD_BUFFER_SIZE=<1..32> |
Default is 3. Increase this if you need deeper capture history for a single probe point.
Java 21 Compatibility Mode
Required if your target JVM runs on Java 21. Enables ByteBuddy's experimental support for newer JVM internals.
| Method | Value |
|---|---|
| Agent arg | allowJava21=true (aliases: java21Compat=true, byteBuddyExperimental=true) |
| JVM property | -Dmcp.probe.bytebuddy.experimental=true (legacy alias: -Dmcp.probe.allow.java21=true) |
| Environment variable | MCP_PROBE_BYTEBUDDY_EXPERIMENTAL=true (legacy alias: MCP_PROBE_ALLOW_JAVA21=true) |
Default is false.
MCP Server Environment Variables
Required
| Variable | Purpose |
|---|---|
MCP_PROBE_BASE_URL | URL of the running probe agent |
Optional
| Variable | Default | Notes |
|---|---|---|
MCP_WORKSPACE_ROOT | — | Project root hint for path resolution |
MCP_JAVA_REQUEST_MAPPING_RESOLVER_JAR | — | |
MCP_JAVA_REQUEST_MAPPING_RESOLVER_CLASSPATH | — | |
MCP_JAVA_BIN | — | |
MCP_PROBE_LINE_SELECTION_MAX_SCAN_LINES | 120 | Range: 10–2000 |
MCP_PROBE_WAIT_MAX_RETRIES | 1 | Max: 10 |
MCP_PROBE_WAIT_UNREACHABLE_RETRY_ENABLED | false | |
MCP_PROBE_WAIT_UNREACHABLE_MAX_RETRIES | 3 | Max: 10 |
MCP_PROBE_INCLUDE_EXECUTION_PATHS | false | Set true to include executionPaths arrays in probe payloads |
Probe Endpoints
These paths are fixed and cannot be overridden.
| Endpoint | Path |
|---|---|
| Status | /__probe/status |
| Reset | /__probe/reset |
| Capture | /__probe/capture |
Skills
| Skill | Purpose |
|---|---|
mcp-java-dev-tools-line-probe-run | Line-level probe execution |
mcp-java-dev-tools-regression-suite | Regression check orchestration |
mcp-java-dev-tools-issue-report | Sanitized issue reporting from session, runtime, and probe evidence |
Contributing
Contribution guidance lives in CONTRIBUTING.md.
The guide distinguishes between:
- synthesizer and adapter contributions
- probe tools and recipe generation contributions
Start there before opening a large pull request or changing public tool contracts.
MCP Tools
| Tool | |
|---|---|
debug_check | |
project_context_validate | |
probe_check | |
probe_target_infer | |
probe_recipe_create | |
probe_get_status | |
probe_get_capture | |
probe_reset | |
probe_wait_for_hit | |
probe_enable |
Related Servers
Scout Monitoring MCP
sponsorPut performance and error data directly in the hands of your AI assistant.
Alpha Vantage MCP Server
sponsorAccess financial market data: realtime & historical stock, ETF, options, forex, crypto, commodities, fundamentals, technical indicators, & more
MCP Repo Search Server
MCP server that gives LLMs structural code intelligence across multiple repos
BlenderMCP
Connects Blender to Claude AI via the Model Context Protocol (MCP), enabling direct interaction and control for prompt-assisted 3D modeling, scene creation, and manipulation.
Hyperlend
Enables AI agents to interact with the Hyperlend protocol.
MCPJungle
A self-hosted MCP Server registry for private AI agents, supporting both PostgreSQL and SQLite databases.
Wopee MCP
AI testing agents for web apps — dispatch test runs, analysis crawls, and AI agent tests, fetch artifacts and project status
ServiceNow
A production-ready Model Context Protocol (MCP) server for ServiceNow platform integration. Built with TypeScript for Node.js 20+, this server enables LLMs and AI assistants to interact with ServiceNow instances through a standardized interface.
Developer MCP Server
A context management system designed for software development teams with customizable data storage.
Hex MCP
A server for listing, searching, running, and managing Hex projects.
Unified MCP Client Library
A TypeScript library for integrating MCP with tools like LangChain and Zod, providing helpers for schema conversion and event streaming.
Pipelock
Firewall for AI agents. MCP proxy that scans tool calls for credential leaks, prompt injection, and tool description poisoning.