Stop hardcoding API keys in your AI agents — how I built a governance layer in 3 weeks

Three weeks ago I got tired of pasting API keys into .env files every time I spun up a new AI agent. GitHub, Linear, Stripe, Notion, Slack, Vercel — each agent ended up with god-mode credentials to half my stack, with no approval flow, no audit trail, no revocation story. I went looking for a tool to fix this. Secrets managers like Vault and 1Password store secrets well but don't model agents , approvals , or agent-initiated tool requests . So I built one. It's called AgentKey , and it launched