Connect your AI to Claude Code routines

A Claude Code routine is a saved prompt that runs as its own Claude session on a schedule or when its API trigger is fired. Firing it is one POST with a per-routine bearer token, and the body's text member reaches the session as an untrusted payload the prompt has to opt in to. That makes a routine a destination an event can wake: the routine reads what woke it, does its work with the connectors it has, and stops. Only Claude Code routines are supported here; the routine and its prompt are made by hand on claude.ai, and the trigger endpoint sits behind an experimental beta header that can change.

The credential never enters the model context: it lives in the FlurryPORT secret store, deliveries are signed server-side, and every send returns a receipt your agent can quote.

Auth model

static_header

Suggested secret

CLAUDE_ROUTINE_FIRE_TOKEN

Scopes

The routine's API trigger token, shown once when the trigger is created. One token per routine.

Provider docs

https://claude.ai/code

Recipes for Claude Code routines

More in Collaboration

• Team collaboration

• Collaboration recipes
• All recipes