AI Workshop
Episode 2 of 10Aired

Connecting and authorizing

Keys, OAuth, MCP servers, secrets and permissions — wiring an agent to services without leaking anything.

Wednesday 14 October at 19:00 Moscow time· 90 min

Aired

The recording is being prepared — it usually appears within two days. Sign up for the next streams so you don't miss them.

What we'll show

  1. Subscription sign-in vs. API key: the difference, where each is stored, how to revoke
  2. Connecting an MCP server (GitHub, a database, a browser) and OAuth sign-in from inside the agent
  3. Secrets: .env.local, a password manager, CI variables — and why a key never goes into a prompt
  4. Permissions: allowlists for commands and tools, permission modes, what to allow forever vs. once
  5. What to do when a key shows up on screen or in a commit: revoke and reissue in five minutes

What you take away

Connected MCP servers and a secrets setup you'd happily show a security engineer.

Safe connection checklist

  • Yes: Keys live in .env.local or a secrets manager, .env* is in .gitignore
  • Yes: One key per service, least privilege, a clear name
  • Yes: Only the MCP servers the project needs are connected; for each you know what it can reach
  • Yes: The agent settings have a command allowlist; dangerous ones (rm -rf, git push --force, deploy) need confirmation
  • Yes: The agent has no production access; production goes through CI with environment secrets
  • Yes: There's a leak plan: where to revoke, who reissues, what to check in the logs

Harness in this episode

Permissions and secrets are the harness perimeter: the agent can make mistakes, but can't step outside what it was given.

The full harness episode