Server data from the Official MCP Registry
Local-first MCP runtime for coding agents with safe workspace access, AST/LSP navigation, and Git.
About
Local-first MCP runtime for coding agents with safe workspace access, AST/LSP navigation, and Git.
Security Report
Valid MCP server (3 strong, 2 medium validity signals). No known CVEs in dependencies. ⚠️ Package registry links to a different repository than scanned source. Imported from the Official MCP Registry. Trust signals: 4 highly-trusted packages. 1 finding(s) downgraded by scanner intelligence.
15 files analyzed · 1 issue found
Security scores are indicators to help you make informed decisions, not guarantees. Always review permissions before connecting any MCP server.
Permissions Required
This plugin requests these system permissions. Most are normal for its category.
How to Install
Add this to your MCP configuration file:
{
"mcpServers": {
"io-github-boxboxmari-my-pi": {
"args": [
"-y",
"@koonwang03/my-pi"
],
"command": "npx"
}
}
}Documentation
View on GitHubFrom the project's GitHub README.
my-pi
What is my-pi?
my-pi is a deterministic local coding capability runtime exposed through the official Model Context Protocol (MCP). It gives MCP-capable coding agents controlled access to a real workspace through explicit workspace authority, bounded reads, guarded writes, structural search, language-server intelligence, and Git operations.
It is aimed at developers who want agentic coding tools to understand and modify code without handing the agent a general-purpose shell or silently granting the current working directory.
Why use it?
- Local-first: source code and detailed workspace state remain on the host.
- Read-only by default: writes and language-server process startup require explicit
trustedelevation. - Safer mutation: writes can require SHA-256 content preconditions instead of blind overwrite semantics.
- Code-aware search: Tree-Sitter structural search across TypeScript, JavaScript, Python, Rust, and Go.
- Semantic navigation: LSP symbols, definitions, references, hover, and diagnostics for supported host language servers.
- Git-aware context: bounded status and diff operations with secret filtering and large-diff spillover.
- No secondary paid LLM dependency: the core runtime executes locally and does not select models or spawn agents.
Release channel: Alpha. The npm badge above is authoritative for the currently published package version. This repository may contain a newer release candidate before publication completes. Suitable for evaluation and controlled local development; review the security model before enabling the trusted profile.
Install in under a minute
Requires Node.js >=22.6.0.
npm install -g @koonwang03/my-pi
my-pi-mcp --workspace /path/to/your/project
The server starts read-only. For a workspace you explicitly trust:
my-pi-mcp --workspace /path/to/your/project --security-profile trusted
Or inspect a host configuration without a global install:
npx --yes --package @koonwang03/my-pi my-pi-mcp host-config cursor-local
Generate host-specific configuration snippets:
my-pi-mcp host-config claude-code-local
my-pi-mcp host-config cursor-local
my-pi-mcp host-config opencode-current-local
Starting without --workspace or MY_PI_WORKSPACE_ROOT fails closed. Use --allow-cwd only when granting the current directory is intentional.
13-tool MCP surface
| Area | Tools | Purpose |
|---|---|---|
| Filesystem | fs_read, fs_write, fs_patch, fs_stat | Bounded reads, guarded writes/patches, metadata |
| Search & workspace | search, workspace_info | Repository exploration and authoritative workspace state |
| AST & LSP | ast_search, lsp_status, lsp_symbols, lsp_navigate, lsp_diagnostics | Structural and semantic code intelligence |
| Git | vcs_status, vcs_diff | Repository status and bounded/filtered diffs |
Key guarantees
| Capability | Behavior |
|---|---|
| Content-preconditioned mutation | File updates verify raw SHA-256 fingerprints and reject stale guarded overwrites |
| Pre-read sensitive-path policy | Sensitive paths such as .env*, .aws/, .ssh/, and *.key are denied before content is allocated to model context |
| Explicit security profiles | Default is read-only; mutation and LSP process startup require explicit elevation |
| Encoding/mode fidelity | File replacement preserves relevant encoding, line endings, BOM, and POSIX executable mode behavior |
| Cancellation | Long-running Git/search/LSP subprocess work supports cancellation and cleanup |
Architecture
MCP-capable coding host
│
│ stdio
▼
┌─────────────────────┐
│ my-pi MCP edge │
└──────────┬──────────┘
▼
┌─────────────────────┐
│ capability contracts│
└──────────┬──────────┘
▼
┌─────────────────────────────────────────────────────────┐
│ workspace │ policy │ filesystem │ search │ AST │ LSP │ Git │
└─────────────────────────────────────────────────────────┘
│
└── local host workspace
The stable public claim is the 13-tool MCP capability surface. The repository also contains an opt-in Production Next coordination/evaluation candidate; those coordination, code-state, change-receipt, evaluation, and feedback features remain experimental and subject to their promotion gates.
Build from source
git clone https://github.com/BoxBoxmari/my-pi.git
cd my-pi
pnpm install --frozen-lockfile
pnpm build
Prerequisites for repository development:
- Node.js
v22.6.0+orv24 LTS - pnpm
v11.2.2+ - Rust stable is optional and currently relevant only to the deferred native-backend scaffold
Verification
# Local code, architecture, boundary, build, tests, gates and smoke verification
pnpm verify
# Unit/integration suite
pnpm test
# SBOM validation
pnpm verify:sbom
# Release admission checks
pnpm bind:evidence
pnpm verify:release
The configured CI matrix covers Ubuntu, Windows, and macOS lanes. See the live workflow badges above for current status rather than relying on static claims in this document.
Benchmarks
The repository contains deterministic synthetic benchmarks for MCP stdio overhead, search/traversal throughput, memory sampling, runtime boundaries, coordination behavior, impact routing, evaluation feedback, and local reliability. Benchmark outputs are candidate evidence; performance claims should be interpreted alongside their qualification criteria and runner variance.
Production Next (experimental)
Start the local coordination candidate for a logical project:
my-pi-daemon --workspace /path/to/your/project
my-pi-mcp --workspace /path/to/your/project --coordination
Add --evaluation only when the evaluation plane is required. The candidate keeps source and detailed code state local, does not select models or spawn agents, and does not require a hosted control plane.
Relevant qualification commands include:
pnpm bench:impact-arms
pnpm bench:evaluation-feedback-arms
pnpm dogfood:self-host
pnpm bench:local-reliability
pnpm verify:production-next
Package topology
packages/
├── contracts/ # Core interfaces, error codes, fingerprinting
├── workspace-runtime/ # Workspace/path normalization and mutation coordination
├── policy/ # Sensitive-path protection
├── artifact-store/ # Disk-backed spillover artifacts
├── observability/ # Tracing and metrics contracts
├── fs/ # Hardened filesystem capabilities
├── search/ # Grep/glob traversal
├── hashline/ # Hashline-anchored patch engine
├── ast/ # Tree-Sitter structural search
├── lsp/ # Multi-language LSP lifecycle/client
├── vcs/ # Git-backed status and diff
├── mcp-adapter/ # MCP stdio server adapter
├── host-profiles/ # Host configuration renderers
├── change-runtime/ # Content preconditions and change receipts
├── code-state/ # Filesystem/AST/LSP/VCS code state
├── coordination-client/ # Local daemon client
├── coordination-runtime/ # Work graph, claims, intents, sync
├── coordination-store/ # SQLite event/projection store
├── evaluation-runtime/ # Evaluation and feedback flow
└── impact-engine/ # Bounded impact/routing decisions
Search-ignore behavior is documented in docs/SEARCH_IGNORE.md. It is a traversal optimization, not a substitute for sensitive-path policy.
Security
Before using trusted mode, read docs/SECURITY_MODEL.md. Security findings are welcome through the repository's documented reporting process.
Contributing
Issues, reproducible bug reports, benchmark counterexamples, integration feedback, and focused pull requests are welcome. If you are evaluating my-pi in a real coding host, include the host, OS, Node version, security profile, and a minimal reproduction where possible.
License
MIT — see LICENSE.
Reviews
No reviews yet
Be the first to review this server!
More Developer Tools MCP Servers
Paperclip
Freeby Paperclipai · Developer Tools
Trending hip-hop artist momentum scores across four cultural dimensions.
Git
Freeby Modelcontextprotocol · Developer Tools
Read, search, and manipulate Git repositories programmatically
Toleno
Freeby Toleno · Developer Tools
Toleno Network MCP Server — Manage your Toleno mining account with Claude AI using natural language.
mcp-creator-python
Freeby mcp-marketplace · Developer Tools
Create, build, and publish Python MCP servers to PyPI — conversationally.
MCP Marketplace
Freeby mcp-marketplace · Developer Tools
Search and install MCP servers from inside your AI client.
MarkItDown
Freeby Microsoft · Content & Media
Convert files (PDF, Word, Excel, images, audio) to Markdown for LLM consumption
