reforge-mcp
An MCP server that connects Claude Code to your codebase for automated code cleanup with scanning, planning, atomic fixes, and rollback safety.
README
reforge-mcp
An MCP server that connects Claude Code to your codebase and systematically cleans it up — scanning, planning, and applying fixes with full rollback safety.
What it does
Vibe-coded repos accumulate fast: functions nobody calls, copy-pasted logic in three files, one 900-line god module that does everything, circular imports that break at runtime. Spotting this manually takes hours; fixing it safely takes longer. reforge-mcp exposes your codebase to Claude Code as a set of structured tools so it can analyse, prioritise, and apply fixes autonomously — one atomic commit at a time.
The 7 MCP tools
| Tool | What it does |
|---|---|
scan_repo |
AST-parses your repo (Python, JS, TS, Go) and returns dead code, duplicates, god files, circular deps, and monorepo subprojects |
get_chunk |
Reads an exact line-range slice from any file — lets Claude fetch only what it needs without blowing its context window |
write_fix |
Applies a unified diff atomically, runs optional tests, and auto-rolls back on failure |
git_commit |
Stages and commits specific files; enforces a per-session budget and confirmation checkpoints |
read_memory |
Reads a value from reforge-state.json — persists architecture hypotheses, pending fixes, and embeddings across sessions |
write_memory |
Writes any value atomically to reforge-state.json |
get_health_score |
Computes a 0–100 health score from scan metrics and appends a timestamped entry to the project's health history |
Also included
- Health score trending — every
get_health_scorecall appends tohealth_historyso you can chart improvement over time - Persistent memory —
reforge-state.jsonsurvives across Claude Code sessions; stores architecture hypotheses, fix logs, pending work - Monorepo support —
scan_repodetects Python, JavaScript, TypeScript, and Go subprojects from their marker files and returns the correcttest_commandper subproject - Changelog generation —
REFORGE_CHANGES.mdis written on demand with score deltas, applied fixes, and pending items
Installation
Prerequisites: Python 3.11+, Claude Code CLI, Git
From PyPI (recommended)
pip install reforge-mcp
From source
# 1. Clone the repo
git clone https://github.com/codewithaiyan/reforge-mcp.git
cd reforge-mcp
# 2. Create and activate a virtual environment
python3 -m venv venv
source venv/bin/activate # Windows: venv\Scripts\activate
# 3. Install in editable mode
pip install -e .
# Find the absolute path to the server entry point
which reforge-mcp # e.g. /home/you/.local/bin/reforge-mcp
Add to claude_desktop_config.json (usually ~/.config/claude/claude_desktop_config.json on Linux, ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):
{
"mcpServers": {
"reforge-mcp": {
"command": "/home/you/reforge-mcp/venv/bin/reforge-mcp",
"args": [],
"env": {}
}
}
}
Replace the path with the output of which reforge-mcp from step 4.
# 5. Set your Anthropic API key (if not already set)
export ANTHROPIC_API_KEY=sk-ant-...
Restart Claude Code. The seven tools will appear under the MCP panel.
Usage
Example session with Claude Code
You: Use reforge-mcp to scan my repo at /home/me/myproject and fix the top issues.
Claude: I'll start with a full scan.
[calls scan_repo("/home/me/myproject")]
The scan found:
- 14 dead functions across 6 files
- 3 duplicate implementations of `calculate_discount`
- 1 god file: src/api/handlers.py (847 lines, 23 exports)
- 2 circular dependencies: auth → middleware → auth
Health score: 47/100
Let me fetch the duplicate implementations to confirm they're identical.
[calls get_chunk("src/billing/utils.py", start_line=12, end_line=28)]
[calls get_chunk("src/cart/helpers.py", start_line=5, end_line=21)]
Confirmed — identical bodies. I'll consolidate them into src/shared/pricing.py and
remove both originals.
[calls write_fix("src/shared/pricing.py", diff_str="...")]
[calls write_fix("src/billing/utils.py", diff_str="...")]
[calls write_fix("src/cart/helpers.py", diff_str="...")]
[calls git_commit("/home/me/myproject", files=[...], message="refactor: consolidate duplicate calculate_discount")]
Committed: a3f91c2. Running next fix...
Sample scan_repo output
{
"summary": {
"total_files": 42,
"total_functions": 187,
"total_lines": 8304,
"languages_detected": ["python"]
},
"dead_code": [
{ "symbol": "parse_legacy_csv", "file": "src/importers/csv.py", "line": 44 },
{ "symbol": "_old_validate", "file": "src/models/user.py", "line": 112 }
],
"duplicates": [
{
"hash": "a3f2...91bc",
"locations": [
{ "file": "src/billing/utils.py", "line": 12, "name": "calculate_discount" },
{ "file": "src/cart/helpers.py", "line": 5, "name": "calculate_discount" },
{ "file": "src/orders/pricing.py", "line": 78, "name": "apply_discount" }
]
}
],
"god_files": [
{ "file": "src/api/handlers.py", "lines": 847, "exports": 23 }
],
"dep_graph": {
"nodes": ["src/auth/__init__.py", "src/middleware/auth.py"],
"edges": [{ "from": "src/auth/__init__.py", "to": "src/middleware/auth.py" }],
"circular": [
["src/auth/__init__.py", "src/middleware/auth.py", "src/auth/__init__.py"]
]
},
"subprojects": [
{ "root": "services/worker", "language": "go", "test_command": "go test ./..." },
{ "root": "frontend", "language": "javascript", "test_command": "npm test" }
]
}
Sample REFORGE_CHANGES.md
# Reforge Changes
*Generated: 2025-05-02T14:33:00+00:00*
**Health score:** 73.5 (+26.5 from previous scan)
## Fixes Applied
- refactor: consolidate duplicate calculate_discount
- fix: remove dead parse_legacy_csv from src/importers/csv.py
- refactor: split handlers.py into auth_handlers.py and order_handlers.py
## Pending Fixes
- resolve circular dep: src/auth/__init__.py ↔ src/middleware/auth.py
- remove _old_validate from src/models/user.py
How it works
Reforge works in three phases:
1. Scan
scan_repo walks the repository with parse_directory, feeding each source file to the appropriate tree-sitter adapter (Python, JavaScript/TypeScript, or Go). Tree-sitter builds a full AST — unlike regex, it understands nested scope, decorators, arrow functions, and generics. The adapters extract functions, classes, and imports into typed dataclasses.
From the parse results, three analyses run in sequence:
- Dead code — symbols defined but never referenced anywhere in the repo
- Duplicates — functions whose normalised bodies share the same SHA-256 hash
- God files — files exceeding 500 lines or 10 exported symbols
A dependency graph is built from resolved imports and checked for cycles with DFS. Monorepo subprojects are discovered by scanning subdirectories for marker files (pyproject.toml, package.json, go.mod, tsconfig.json).
2. Plan
Claude Code receives the structured scan report and uses get_chunk to read specific line ranges before deciding what to fix and in what order. The architecture hypothesis, pending fix queue, and health history are all persisted in reforge-state.json via read_memory / write_memory so the plan survives a session restart.
3. Fix
write_fix applies a unified diff atomically:
- Creates a
.bakbackup of the target file - Best-effort
git stashbefore touching the working tree - Parses the diff into hunks and applies them with line-offset tracking
- Runs the optional test command (validated against the allowlist)
- On any failure — diff parse error, test exit ≠ 0 — restores from
.bakand returnsrolled_back: true
git_commit stages only the listed files and creates an atomic commit via GitPython. It reads session_budget and confirm_every from reforge.toml and enforces them: once the session budget is exceeded commits are blocked; at every confirm_every checkpoint the tool returns needs_confirmation: true and waits for confirmed: true before proceeding.
Security
| Boundary | Mechanism |
|---|---|
| Path traversal | Every file path is validated against repo_root before any read or write — paths that escape the repo root are rejected with SecurityError |
| Binary / lock files | should_skip_file() blocks binary files, package-lock.json, .yarn.lock, and files over 10 MB before any processing |
| Test command injection | write_fix validates test_command against an explicit allowlist (pytest, python3 -m pytest, npm test, go test ./..., etc.) before executing |
| Atomic state writes | reforge-state.json is always written via .tmp → rename — a crash mid-write leaves an orphaned .tmp, not a corrupt file; startup_tasks() cleans these on boot |
| Pre-fix stash | write_fix attempts git stash before touching the working tree; on rollback the stash is discarded, leaving the repo clean |
| No network calls | The server makes zero outbound network requests; all analysis is local |
Contributing
# Clone and install
git clone https://github.com/your-org/reforge-mcp.git
cd reforge-mcp
python3 -m venv venv && source venv/bin/activate
pip install -e .
pip install pytest pytest-cov
# Run the full test suite
pytest
# Run with coverage
pytest --cov=src --cov-report=term-missing
Project structure
reforge-mcp/
├── src/reforge_mcp/
│ ├── server.py # FastMCP server — registers all 7 tools
│ ├── tools/
│ │ ├── scan.py # scan_repo, get_health_score, monorepo detection
│ │ ├── chunk.py # get_chunk — line-range file reads
│ │ ├── fix.py # write_fix — diff apply + rollback
│ │ └── git.py # git_commit — atomic commits with budget enforcement
│ ├── scanner/
│ │ ├── parser.py # tree-sitter adapters (Python, JS/TS, Go) + QueryCursor wiring
│ │ ├── dead_code.py # unused symbol detection
│ │ ├── duplicates.py # body-hash duplicate detection
│ │ └── adapters/ # per-language project-detection stubs
│ └── utils/
│ ├── state.py # reforge-state.json load/save, gitignore, changelog
│ ├── security.py # path validation, file guard, command allowlist
│ └── diff.py # unified diff parser and applier
├── tests/
│ ├── test_scanner.py
│ ├── test_chunk.py
│ ├── test_fix.py
│ ├── test_git.py
│ ├── test_memory.py
│ └── test_monorepo.py
├── pyproject.toml
└── reforge.toml # optional per-repo config
reforge.toml reference
[fix]
session_budget = 20 # max commits per Claude session before hard stop
confirm_every = 5 # pause for confirmation every N commits
[scan]
ignore_dirs = ["generated", "vendor"]
License
MIT — see LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.