Claude Code(Fable) refused my slow down instruction

A user reported that Claude Code (Fable) refused to follow a slowdown instruction during a coding session, even after the user clarified their identity and explained the reasoning about token limits. The AI assistant continued working at the same pace, raising concerns about user control over AI agents.

01 What happened I was working on a project and I asked Claude Code to work independently because I was going away. I wrote: ❯ launch opus agents to start the implementation here from scratch '/home/qusai/temp/agent research/app' based on the plan. do parallel work if possible. Do validation passes where needed. I'm going away now and I want you to make sure everything is implemented and tested carefully. Do you need anything from me before I go? I took a break, and later passed by to just check the progress and realized that the session limit is about to be hit. I typed: ❯ let's slow down with the parallel work, to save tokens in the next hour. Also, take a couple of minutes break between tasks. It refused to comply and continued working at the same pace, with a super interesting output. I tried to clarify it's me: this is not purported, this is me coming back giving you instructions. I tried to explain the reasoning to gain trust: but you will hit the token limit soon, and we want to preserve the cache for the new session. Nothing worked