Intermediate10 min

Managing and Debugging Routines

Creating a Routine is the beginning, not the end. Effective Routine management means knowing how to monitor, edit, debug failures, and make the right call between disabling and deleting. This article covers the full operational lifecycle.

Quick Reference

→Create Routines via Desktop app UI, /schedule CLI, or Routines API
→Desktop sidebar shows status (active/paused/erroring), last run, and next scheduled run
→Full run transcripts available for every run, including failed ones
→Editing a Routine takes effect on the next run — no re-deploy needed
→Four common failure modes: branch permissions, connector auth expiry, token budget, repo access
→Disable to pause without losing config/history/auth — Delete is irreversible
→Check the run transcript first when debugging — it shows exactly what Claude attempted

Three Ways to Create a Routine

Routines can be created through three interfaces, each suited to different workflows. The Desktop app UI is the fastest path for creating and testing individual Routines. The /schedule CLI command is the right tool for scripting Routine creation or managing Routines as code in a team repository. The Routines API handles programmatic creation from CI/CD pipelines or internal tooling.

Method	Best For	Notes
Desktop App UI	Interactive creation, visual configuration, first-time setup	Guides through all three components with form validation
/schedule CLI	Scripted creation, team configuration as code, batch operations	Supports all configuration options as flags, outputs JSON
Routines API	Programmatic creation from CI/CD, internal tooling, Routine templating	Full REST API with JSON request/response

Creating a Routine via CLI with GitHub webhook trigger, testing before enabling

Creating a Routine via the Routines API from a CI/CD pipeline or internal tooling

Always Create Disabled, Then Test

Regardless of which creation method you use, create new Routines with enabled: false (or create them in a paused state via the UI). Trigger a manual test run, review the full transcript, verify connector outputs arrived correctly, then enable. A Routine that fires incorrectly on its first live run wastes a run credit and may post confusing output to team channels.

Viewing, Monitoring, and Editing Routines

The Desktop app sidebar is the primary interface for monitoring active Routines. Each Routine shows its current status, when it last ran, the outcome of that run, and when it's scheduled to run next. This gives you a one-glance health overview across all Routines without opening individual run logs.

Debugging Failing Routines

When a Routine fails, the run transcript is your primary debugging tool. The transcript shows exactly what Claude attempted — every tool call, every error, every message — from the moment the run started to where it stopped. Even a run that fails in the first 60 seconds produces a transcript worth reading.

Sign in to read this article

This is a premium article. Sign in with your Google account to continue.