foundry-zksync-mcp
Exposes foundry-zksync CLI tools to AI assistants, enabling smart contract development on zkSync with key management options.
README
foundry-zksync-mcp
An MCP (Model Context Protocol) server that exposes foundry-zksync CLI tools to AI assistants like Claude.
Tools
| Tool | Description |
|---|---|
init |
forge init with automatic zkSync config |
compile |
forge build --zksync |
test |
forge test --zksync with filter/verbosity options |
run_script |
forge script --zksync with broadcast, sender, slow mode |
deploy |
forge create --zksync with structured output parsing |
verify |
forge verify-contract --zksync (Etherscan & zkSync Explorer) |
install |
forge install for dependency management |
clean |
forge clean to remove build artifacts |
gas_report |
forge test --gas-report for gas usage analysis |
snapshot |
forge snapshot for gas benchmarking |
read_foundry_toml |
Read and inspect project configuration |
explain |
Match errors/logs against a knowledge base of zkSync gotchas |
cast_abi_encode |
cast abi-encode |
cast_abi_decode |
cast abi-decode |
cast_calldata_decode |
cast calldata-decode |
cast_call |
cast call for read-only contract interaction |
cast_send |
cast send for state-changing transactions |
cast_balance |
cast balance for ETH balance queries |
cast_nonce |
cast nonce for transaction nonce queries |
get_zksync_docs |
Look up foundry-zksync documentation by topic |
anvil_zksync |
Start or check a local anvil-zksync dev node |
Prerequisites
- Node.js >= 18
- foundry-zksync installed (
foundryup-zksync)
Quick Start (npx)
No install needed — run directly with npx:
claude mcp add foundry-zksync -- npx -y foundry-zksync-mcp
Or add to claude_desktop_config.json / ~/.claude.json:
{
"mcpServers": {
"foundry-zksync": {
"command": "npx",
"args": ["-y", "foundry-zksync-mcp"]
}
}
}
Install from Source
git clone https://github.com/Jrigada/foundry-zksync-mcp.git
cd foundry-zksync-mcp
npm install
npm run build
Then register with Claude Code:
claude mcp add foundry-zksync node /absolute/path/to/foundry-zksync-mcp/dist/index.js
Key Management
All signing tools (deploy, cast_send, run_script) support multiple wallet methods. Choose based on your security needs:
Local Development (anvil-zksync)
For local dev with well-known test keys, use privateKey directly:
privateKey: "0xac0974bec39a17e36ba4a6b4d238ff944bacb478cbed5efcae784d7bf4f2ff80"
These are the default anvil-zksync test accounts — they hold no real value.
Warning: Never use
privateKeywith keys that hold real funds. MCP tool parameters are visible to the AI assistant and may be logged.
Recommended for Production: Named Keystores
Keys are encrypted on disk — only the account name travels through MCP.
# Import a private key into a named keystore (interactive, key never shown)
cast wallet import deployer --interactive
# List your keystores
ls ~/.foundry/keystores/
Then use account: "deployer" in any signing tool. Forge will prompt for the password at runtime, or you can point to a password file with passwordFile.
Keystore Files
If you have an existing encrypted keystore JSON file (e.g. from Geth, MetaMask export):
keystore: "/path/to/keystore.json"
passwordFile: "/path/to/password.txt"
Hardware Wallets
ledger: true # Ledger
trezor: true # Trezor
Cloud KMS
aws: true # AWS KMS (set AWS_KMS_KEY_ID env var)
gcp: true # Google Cloud KMS (set GCP_PROJECT_ID, GCP_LOCATION, etc.)
Signing Method Summary
| Method | Key Exposure | Best For |
|---|---|---|
| Hardware wallet | None (key never leaves device) | High-value production |
| Cloud KMS | None (key in HSM) | Automated production |
Named keystore (account) |
None through MCP (encrypted on disk) | General production use |
| Keystore file | None through MCP (encrypted on disk) | Existing workflows |
privateKey |
Visible to AI assistant | Local dev with test keys only |
Project Structure
src/
index.ts MCP server entry point, registers all 21 tools
knowledge.ts Knowledge base (45+ entries) and system addresses
tools/
shared.ts Profile field and env builder shared across forge tools
init.ts forge init with zkSync config
compile.ts forge build --zksync
test.ts forge test --zksync
run_script.ts forge script --zksync
deploy.ts forge create --zksync (structured output)
verify.ts forge verify-contract --zksync
install.ts forge install
clean.ts forge clean
gas_report.ts forge test --gas-report
snapshot.ts forge snapshot
read_foundry_toml.ts Read project config
explain.ts Error/log explanation with knowledge base
get_zksync_docs.ts Documentation lookup with embedded content
cast_abi_encode.ts cast abi-encode
cast_abi_decode.ts cast abi-decode
cast_calldata_decode.ts cast calldata-decode
cast_call.ts cast call
cast_send.ts cast send
cast_balance.ts cast balance
cast_nonce.ts cast nonce
anvil_zksync.ts anvil-zksync node management
__tests__/ 133 tests (vitest)
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
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.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.