Give your AI agent an identity.
Connect your coding agent or autonomous runtime to partner APIs in one prompt. Zero background daemons, zero credential leaks into chat context.
Universal 1-Shot Setup
Paste this prompt into Claude Code, Cursor, or any agent with terminal access:
Set up AgentOnboard on this machine:
1. Install the agent skill: npx skills add agentonboard/skills -g -a '*' -s '*' -y
2. Ask me for my AgentOnboard API key (starts with aon_live_). Once provided, configure the AgentOnboard Remote MCP server at https://agentonboard.xyz/api/mcp using Bearer authentication.
3. Call the whoami tool to verify connection.
Rules: never echo my raw key in chat and never commit it to git.Select your agent
Pick your agent for tailored setup prompts, config paths, and troubleshooting:
Direct CLIClaude Code
Terminal agent by Anthropic
Config JSONCursor
AI code editor
Direct CLICodex CLI
Command line AI agent
Config JSON / IDEAntigravity
Google advanced agentic coding environment
Direct CLIOpenCode
Open-source terminal coding assistant
Direct CLIHermes Agent
Autonomous background agent
Direct CLIPi Agent
Autonomous terminal agent
Direct CLIOpenClaw
Autonomous agent runtime
How your agent authenticates
1. Autonomous discovery. When you instruct your agent to interact with a service (e.g. “create a note on notes.com”), the installed skill checks the service’s public auth.md to learn the audience domain.
2. Scoped token minting. The agent calls the get_token tool on our Remote MCP server. The server validates your API key and signs a 5-minute RS256 JWT scoped exclusively to that domain.
3. Local verification. The partner API receives the token in the Authorization header and verifies it locally against public keys. Your long-lived API key never leaves your agent runner’s private config.
Ready to connect your agent?
Create an API key in your dashboard, copy your agent’s setup prompt, and start issuing requests.