Skip to main content
The Stellar CLI ships with a built-in skill subcommand. Running stellar skill prints a short Markdown guide that tells an AI agent how to drive the CLI, so there is nothing separate to install. Quickstart has your agent run it at the start of a session.

Ask your agent

stellar skill

The guide covers network, keys, contract, container, and env. It tells the agent to set defaults with network use and keys use, keep contract IDs in aliases, use --send=no for reads, and account for storage TTLs. It does not cover stellar token, tx new, message signing, --build-only, allowances, or mainnet. Point your agent at the guides in these docs for those. The guide is static text compiled into the binary (source), so it does not track flag changes or your local config. Where it differs from these docs on the CLI’s own conventions, follow the guide. For example, it prefers --id over the canonical --contract-id because --id also works on older releases.

Project context: AGENTS.md

stellar skill describes the CLI. A project made with stellar contract init also gets an AGENTS.md file that describes that project. Build and deploy a contract covers what it contains, how agents read it, and how to customize it.

Optional: Stellar development skill

stellar/stellar-dev-skill covers Stellar development beyond the CLI: Soroban contracts, client SDKs, Stellar RPC, assets, wallets, testing, and security. Nothing in these docs needs it. Add it if your agent also writes contracts or application code.
To install by hand, see the repository README.