Sign inGet started
Skills

Skills troubleshooting

View as MarkdownOpen in ChatGPTOpen in Claude
Connect your agent
Add to CursorAdd to VS Code

Why scheduled skills say Paused, why a routine stopped running, why findings don't arrive and what to do when a finding is wrong, answered step by step.

Updated

Most problems here are one of two things: a schedule nobody allowed, or a week where nothing happened. Here's how to tell which.

Why do my scheduled skills say Paused?

Your scheduled skills run only on your own Claude Code, and it isn't set up yet. Switch on Allow on my Claude Code in Skills and set up the routine. Run skills on a schedule.

Why did the routine stop running?

  • With a token: Skills shows Reconnect Claude when the trigger's token stopped working. Make a new token in the routine and paste it.
  • Quick setup in the Claude desktop app: it runs only while the app is open. Paste the line into Claude Code in the terminal or at claude.ai/code instead, to run it in the cloud.

Why don't any findings arrive?

  1. A quiet week is normal: a run that finds nothing posts nothing.
  2. Check the skill has a schedule on its page in Skills.
  3. Ask your agent "Run my weekly check-up now" and read what it says.
  4. Skills that need a source (an ad account, Search Console) find nothing until it's connected.

What if a finding is wrong?

Tell your agent what's wrong and ask it to remember: "Remember that /old-pricing is retired." Or dismiss it: mark_insight.

Next step

Skills FAQ

Related
  • Skills FAQ: Skeptical questions about skills: why bother, what they can touch, who runs and pays for them and whether they miss things.
  • Skills: Skills are instructions your agent follows on demand or on a schedule. They tell you what changed and what to fix, and stay quiet otherwise.
  • Run skills on a schedule: Allow scheduled skills on your Claude Code once: a cloud routine with a token, or a one-line hourly check.
  • Skills security: What a scheduled analytics run can read and do, whose Claude Code runs it, what it can never touch and the limits that keep your website safe.
  • How a run works: A scheduled run, step by step: it starts, reads context and numbers, reports a finding or stays quiet and finishes.
  • Anatomy of a finding: One finding from start to fix, told as a story: our rule notices wasted ad spend, a skill runs and explains it, and a person decides what to change.
Still have questions?
Was this page useful?