Claude Opus 4.6 to 4.7 Prompt Changes
Key Questions
What changes occurred in Claude Opus from 4.6 to 4.7?
Anthropic made tweaks to the system prompt between Claude Opus 4.6 and 4.7. These changes sparked a 365-point debate on Hacker News.
Why did the prompt changes cause debate on HN?
The updates relate to Chain-of-Thought (CoT) control issues in frontier models. They impact safety and enterprise prompting reliability.
What is Anthropic's stance on OpenClaw-style Claude CLI?
Anthropic has allowed OpenClaw-style Claude CLI usage again, as noted in a 427-point HN discussion. This ties into evolving prompt and access policies.
Anthropic's prompt tweaks between versions spark HN debate (365pts); tied to CoT control issues in frontier models, impacting safety/enterprise prompting.
Sources (2)
Updated Apr 21, 2026