sprawl-cli MCP server
Commit Nov 5, 2026Synchronizes AI agent rules and configurations across multiple IDEs using a GitOps-driven, local-first approach.
Also known as sprawl-software/sprawl-cli
- Stars
- 5
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- —
- Last commit
- Nov 5, 2026
- License
- —
What is sprawl-cli MCP?
sprawl-cli synchronizes AI agent rules and configurations across multiple IDEs using a GitOps-driven, local-first approach.
Summary generated from the project's README
About sprawl-cli MCP server
Sprawl CLI provides a centralized way to manage rules, skills, and boundaries for AI coding agents across different development environments. It uses a GitOps approach to keep configurations consistent, preventing the fragmentation that occurs when different editors like Cursor, VS Code, or Windsurf use separate rule files. By storing these instructions in a Git registry, teams can ensure that every developer follows the same standards without manual updates to every repository.
The tool operates entirely on your local machine, meaning it requires no cloud accounts and performs no telemetry. It works by compiling a single manifest into the native formats required by various IDEs and agents, including IntelliJ, Claude Code, and GitHub Copilot. Users can manage their setup through commands like sprawl sync to pull updates or sprawl bind to generate configuration files. It also includes a scoped MCP file server that restricts agent access to the specific workspace and explicitly mounted directories to maintain security.
What you can do with it
- 1Synchronize agent rules across teamsUpdate AI agent instructions across dozens of repositories simultaneously using a central Git registry.
- 2Detect configuration driftIdentify when a local workspace has deviated from the official organizational rules.
- 3Audit context token usagePerform a static audit of how many tokens are being used by different context loading strategies.
Written from the project's README. Every use case is checked against a quote from it.
README
View on GitHubDeveloper AI governance: GitOps rule management and multi-IDE consistency for AI coding agents.
Sprawl CLI
One source of truth for the rules, skills, and boundaries your AI coding agents follow.
Website · Docs · Quickstart
The Problem
AI coding agents (Cursor, Claude Code, Copilot, RooCode, Antigravity) each read their own configuration files, and teams end up with:
- Fragmented rules:
.cursorrules,.windsurfrules,.clinerules, andCLAUDE.mdscattered across every repository, each slightly different. - Configuration drift: a rule updated in one repo never reaches the other forty.
- Wasted context: agents load
node_modules, build caches, and irrelevant files into their context window. - No audit trail: security and platform teams cannot see which rules each team's agents actually run with.
The Solution
Sprawl CLI keeps agent rules in Git and compiles them into every editor's native format, locally, with no cloud service.
| Capability | How |
|---|---|
| GitOps Rule Sync | Rules, skills, and personas live in a Git registry. sprawl ws push updates every tracked repository. |
| Universal IDE Binding | One manifest compiles to native settings for Cursor, VS Code, IntelliJ, Claude, Gemini, and Copilot. |
| Drift Detection | sprawl diff shows where a workspace has moved away from the central registry. |
| Directory-Scoped MCP File Access | The bundled MCP file server only serves the workspace and explicitly mounted directories. |
| Local-Only | No accounts, no telemetry, no network calls except the git operations you run. |
Sprawl is a governance and configuration tool. It is not an operating system sandbox and does not replace OS-level isolation.
Quickstart
Install
Linux & macOS:
# Automated bootstrap with pipx isolation
curl -sL https://sprawl.software/install.sh | bash
# Fallback via GitHub raw:
curl -sL https://raw.githubusercontent.com/sprawl-software/sprawl-cli/main/install.sh | bash
Windows (PowerShell 5.1+ / 7+):
# Native PowerShell bootstrap with pipx isolation
irm https://sprawl.software/install.ps1 | iex
# Fallback via GitHub raw:
irm https://raw.githubusercontent.com/sprawl-software/sprawl-cli/main/install.ps1 | iex
Cross-Platform (direct pipx):
pipx install git+https://github.com/sprawl-software/sprawl-cli.git
Initialize & Govern
# Import your team's global DNA registry
sprawl init https://github.com/your-org/your-dna-repo.git
# Graft onto an existing project (harvests legacy editor rules automatically)
cd my-project/
sprawl graft
# Sync DNA, provision venv, compile IDE bindings
sprawl sync
# Connect to your active editors
sprawl bind
Core Commands
| Command | Usage |
|---|---|
sprawl create |
Scaffold a new governance-compliant workspace |
sprawl graft |
Onboard an existing project without losing custom configs |
sprawl sync |
Pull DNA updates, provision venv, compile IDE bindings |
sprawl bind |
Generate native config files for all active editors |
sprawl ignore |
Compile stack-aware ignore shields ([detect|sync|check] [--all]) across up to 11 IDE targets |
sprawl add |
Inject rules, skills, or workflows from the registry |
sprawl mount |
Give agents scoped access to additional directories |
sprawl status |
Inspect workspace identity and DNA health |
sprawl token-size |
Statically audit Always-On (Turn 1) vs. Lazy-Loaded (On-Demand) context tokens ([--detailed] [--max-tokens <N>]) |
sprawl diff |
Detect configuration drift against the central registry |
sprawl doctor |
Run system-wide diagnostics |
sprawl ws |
Manage all tracked workspaces from a single registry |
sprawl demo |
Interactive demonstration |
sprawl man |
Offline terminal documentation |
sprawl wipe |
Clean uninstall with no configuration left behind |
Architecture
┌─────────────────────────────────────────────────────────────────┐
│ DEVELOPER WORKSTATION │
│ │
│ ┌──────────────┐ ┌─────────────┐ │
│ │ IDE / Agent │───────── HTTPS ────────▶│ Cloud LLM │ │
│ └──────────────┘ └─────────────┘ │
│ │ │
│ ┌──────┴───────────────────────────────────────────────────┐ │
│ │ SPRAWL (LOCAL) │ │
│ │ │ │
│ │ ┌───────────────────┐ ┌──────────────────────────┐ │ │
│ │ │ RULE REGISTRY │ │ SCOPED MCP FILE SERVER │ │ │
│ │ │ ~/.sprawl/ (Git) │ │ Path traversal guard │ │ │
│ │ │ Compiled per IDE │ │ Workspace + mounts only │ │ │
│ │ └───────────────────┘ └──────────────────────────┘ │ │
│ │ │ │
│ └──────────────────────────────────────────────────────────┘ │
└─────────────────────────────────────────────────────────────────┘
Supported Editors & Agents
- Cursor (
.cursorrules) - VS Code / GitHub Copilot (
.github/copilot-instructions.md,.vscode/settings.json) - Visual Studio
- Claude Code (
CLAUDE.md, MCP config) - Gemini / Google Antigravity (
gemini.json,.agent/) - IntelliJ / JetBrains
- Windsurf (
.windsurfrules) - RooCode / Cline (
.clinerules)
Design Principles
- 100% Local: Runs entirely on your machine. No cloud, no accounts, zero telemetry.
- Cross-Platform: Linux, macOS, and Windows 10/11.
- Agnostic Outputs: One manifest compiles to all editors. Switch tools without reconfiguring.
- Minimal Dependencies: Python standard library plus Rich. Sub-100ms boot.
- Auditable: Small, readable codebase.
Development & Testing
1. Setup Local Environment
Clone the repository and install Sprawl in editable mode with development dependencies:
git clone https://github.com/sprawl-software/sprawl-cli.git
cd sprawl-cli
python -m venv .venv
source .venv/bin/activate # On Windows: .venv\Scripts\Activate.ps1
pip install -e ".[dev]"
2. Lint & Code Formatting (Ruff)
ruff check src/ tests/
ruff format --check src/ tests/
3. Run Unit & Integration Tests
pytest
4. Run E2E Integration Suite (Smoke Test)
Run the automated 26-scenario end-to-end verification:
python tests/qa_run_verifier.py --ci
Contributing
Sprawl is in active private development. Contribution guidelines will be published at public launch. For early access or partnerships: hello@sprawl.software.
License
Business Source License 1.1. See LICENSE.
- Free for non-production use, and for production use by individuals and organisations with 5 people or fewer (see LICENSE for the exact thresholds).
- Other production use requires a commercial license.
- Agencies and consultants deploying Sprawl for clients require a Partner License.
Contact: hello@sprawl.software
Sprawl Software is a trading name of WebWizArt.be BV, Belgium.
sprawl-cli MCP questions
Is Sprawl an operating system sandbox?
No, it is a governance and configuration tool and should not be used as a replacement for OS-level isolation.
How can I detect if my workspace configuration has drifted?
You can use the sprawl diff command to see where a workspace has moved away from the central registry.