Codex setup for FrankKi

Setup

[mcp_servers.frankki]
url = "https://mcp.frankki.app"
bearer_token_env_var = "FRANKKI_API_KEY"
default_tools_approval_mode = "writes"
enabled_tools = ["mcp_health", "address_validate", "letter_create_draft", "letter_preview", "shipping_quote", "order_send", "approval_submit", "order_status"]

1. 1.Create the narrowest suitable Test-Token when trial access is shown for your account; otherwise create a narrowly scoped production token and start with dryRun.
2. 2.Export FRANKKI_API_KEY in the Codex host environment, then add the table to ~/.codex/config.toml or trusted-project .codex/config.toml.
3. 3.Restart the host, run codex mcp list, and use /mcp to verify that FrankKi is enabled.
4. 4.Call mcp_health first. Keep write-tool approval at writes or stricter and follow the safe onboarding workflow before any live send.

Required FrankKi safety boundary

Validate the address, create a stable draft, inspect every preview page, obtain a fresh quote, set maxCostEuros, persist a stable clientOrderId, and obtain an explicit human decision before a live submission. Framework approval controls supplement FrankKi approval; model text never substitutes for the human decision.

Continue with canonical agent onboarding, the draft guide, send guide, and error recovery.