Build Anthropic agent workflows that execute through Swytchcode.
6
Setup steps
1
Provider path
Auditable
Tool calls
Safe
Execution
CLI
Pull the manifest, set your policy, execute. No SDK sprawl, no boilerplate. Just the tool you need, right now.
npm install @swytchcode/runtime @anthropic-ai/sdk dotenv
swytchcode login
swytchcode get githubHow it connects
Follow these steps to connect your agent, load a toolkit, and start executing tools through Swytchcode.
Install the CLI so Claude can discover and execute trusted tools from your project.
npm install -g swytchcodeSet up a Swytchcode project, install the GitHub integration, enable a tool, and connect your account via OAuth.
swytchcode init
swytchcode get github
swytchcode add method github.user.starred.update
swytchcode auth connect githubInstall the Runtime SDK together with the Anthropic SDK.
npm install @swytchcode/runtime @anthropic-ai/sdk dotenvStore your key in a .env file so the SDK can authenticate with Claude.
ANTHROPIC_API_KEY=sk-ant-...Initialize Swytchcode with the Anthropic provider, load your tools, and let Claude call them directly.
import { Swytchcode, TOOL_USE_INSTRUCTIONS } from '@swytchcode/runtime'
import { AnthropicProvider } from '@swytchcode/runtime/providers/anthropic'
const swx = new Swytchcode(new AnthropicProvider())
const tools = await swx.tools.get({ toolkits: ['github'] })Swytchcode validates the request, resolves credentials, evaluates policy, executes the call, and returns the result to Claude.
node main.jsWorks with
Install the CLI and pull your first manifest in under a minute. No account required.