safe

Create, manage, and execute Safe multisig transactions using the Safe{Core} SDK across EVM chains.

6|20|Updated Mar 14, 2026
One-click install
npx skills add https://github.com/andresdefi/cryptoskills --skill safe-andresdefi
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: safe
Source: https://github.com/andresdefi/cryptoskills/tree/main/skills/safe
Command: npx skills add https://github.com/andresdefi/cryptoskills --skill safe-andresdefi

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI agents frequently use deprecated Safe SDK packages, wrong contract addresses, and outdated API patterns when working with Safe multisig wallets, leading to failed transactions and integration errors. This Skill provides current Safe{Core} SDK patterns (protocol-kit, api-kit, relay-kit), verified contract addresses, and correct transaction lifecycle workflows. ## Core Features & Use Cases - Safe Deployment & Management: Deploy new Safes with deterministic addresses, configure owners and thresholds, and manage owner sets across 15+ EVM chains. - Full Transaction Lifecycle: Create, propose, confirm, and execute multisig transactions through the Safe Transaction Service, including batch transactions via MultiSend and rejection flows. - Modules, Guards & ERC-7579: Enable Safe Modules, set transaction guards, and integrate ERC-7579 modular smart account features like session keys and passkey validators. - Use Case: A developer building a treasury management tool can use this Skill to deploy a 3-of-5 Safe, propose an ERC-20 transfer, collect confirmations from owners, and execute it once the threshold is met. ## Quick Start Use the safe skill to deploy a 2-of-3 multisig Safe on Ethereum mainnet and propose a 0.5 ETH transfer transaction.

Frequently Asked Questions about safe

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create and deploy a Safe multisig wallet programmatically?

Use Safe.init from @safe-global/protocol-kit with a predictedSafe config containing owners, threshold, saltNonce, and safeVersion. Call getAddress to predict the address, then createSafeDeploymentTransaction to deploy it on-chain.

How to propose and execute a Safe multisig transaction?

Create the transaction with protocolKit.createTransaction, sign it with signTransaction, then propose it via apiKit.proposeTransaction. Other owners confirm with apiKit.confirmTransaction, and anyone executes once the threshold is met using executeTransaction.

What is the difference between Safe modules and guards?

Modules can execute transactions on behalf of a Safe without owner signatures, bypassing the threshold. Guards are hooks that run pre- and post-execution checks on every transaction for policy enforcement but cannot initiate transactions.

Why does my Safe transaction fail with GS013?

GS013 means execTransaction reverted, usually due to invalid signatures. Verify signatures are sorted by signer address ascending, the signer count meets the threshold, the nonce matches the on-chain nonce, and all signers are current owners.

Does the Safe SDK still use EthersAdapter?

No, EthersAdapter and the old @gnosis.pm packages are deprecated. Protocol Kit v4+ takes a provider (RPC URL or EIP-1193) and signer directly in Safe.init, with no adapter classes required.

Are Safe contract addresses the same on all chains?

Yes, Safe v1.4.1 uses the ERC-2470 singleton factory for deterministic deployment, so the singleton, factory, and MultiSend addresses are identical across Ethereum, Arbitrum, Base, Optimism, Polygon, and other supported networks.