A rule that adds ceremony needs to get out of an agent's way, including a rule I wrote myself. It is easy to complain about a slow model while treating my configuration as sacred. If I asked it to request confirmation at every step, part of that queue of questions has my signature on it.
In July 2026, I questioned whether the rules we had created to organize models were holding newer models back. Instead of demanding another answer, I looked at how I was organizing execution. It was a suspicion about the configuration, worth investigating without promoting it to a diagnosis before doing the work.
That brings me back to the Opus 4.8 announcement on May 28, 2026. Anthropic introduced dynamic workflows in Claude Code, with the idea that execution could adapt to the job. A tool capable of choosing different routes deserves instructions that allow that choice. Otherwise I acquire new capability and tie it to an old script.
I understand how the pile grows. One mistake produces a rule. Another request needs an exception. Eventually the file says to investigate everything, execute quickly, and stop to confirm things already authorized. That is the situation I want instruction reviews to prevent. A document can sound reasonable line by line and become impossible when read as a whole.
My first question is which mistake a rule prevents. I want to name that mistake and recognize it when it happens. If the instruction protects necessary behavior, that requirement stays. If it prescribes a command sequence without explaining why, I need to examine how much of the agent's route I have decided for it.
Deleting everything would erase useful knowledge too. The review should be small enough to understand: one questionable instruction, the behavior worth preserving, and comparable executions. An occasional difference calls for another observation. It does not justify announcing that bureaucracy caused every problem in the workflow.
The announced feature does not perform this investigation for me. It does make frozen instructions harder to defend while the tool underneath changes. I will judge rules by the errors they prevent and rewrite those that merely prescribe steps. I want understandable limits, a delivery I can check, and freedom to choose the route. An instruction that cannot explain its usefulness belongs in review alongside the bad code.