A Claude Code plugin that lets Claude discover and respond to code review comments left by a human in the Remata IntelliJ plugin.
No server to run. No accounts. Install the plugin, and Claude can see your review comments in any project.
- Python 3.10+
- Claude Code CLI
pip install mcp
# Install the Python dependency
pip install mcp
# Install the plugin into Claude Code
claude plugin add /path/to/remata-review-plugin/remata:review
Claude will list all open review comments, read the referenced code, address each issue, and resolve the comments with explanations.
| Tool | Description |
|---|---|
list_reviews |
List comments — filter by status (open/resolved/wontfix/all) or file |
show_review |
Show a comment with full thread and code context |
reply_to_review |
Append a reply without changing status |
resolve_review |
Mark resolved (or wontfix) with an explanation |
create_review |
Create a new comment on a file and line |
Claude discovers these tools via Tool Search and can use them directly in any conversation.
Human in IntelliJ Claude Code
┌──────────────┐ ┌──────────────┐
│ Right-click │ │ /remata:review│
│ Add Review │──► .review/comments/ ◄──│ list_reviews │
│ Comment │ ├── a3f1c2d4.json │ show_review │
│ │ └── b7e902f1.json │ resolve_review│
└──────────────┘ └──────────────┘
Both sides read and write the same JSON files. The IntelliJ plugin watches the directory for changes in real time. Claude Code uses the MCP tools to interact with the files structurally.
Each comment is a JSON file in .review/comments/:
{
"id": "a3f1c2d4",
"schema_version": 1,
"author": "human",
"created": "2026-03-25T10:14:00Z",
"status": "open",
"anchor": {
"file": "src/Main.java",
"commit": "f3a9b1c",
"line_hint": 142,
"hunk": {
"context_before": ["previous line 1", "previous line 2"],
"target": ["the commented line"],
"context_after": ["next line 1"]
}
},
"body": "This should handle the null case.",
"thread": []
}See the full file format spec for details.
# Run the MCP server directly (for testing)
cd remata-review-plugin
python3 server/review-server.py
# The server uses stdio transport — it reads JSON-RPC from stdin
# and writes responses to stdout.