Anthropic Reduces Claude Code System Prompts by 80%, New Model Maintains Coding Performance Without Strict Rules

TechFlow reports that on July 25, Anthropic engineers disclosed that the team reduced approximately 80% of the system prompts for Claude Code in the latest Claude model, with no performance decline observed in coding evaluations. Early versions contained many strict rules in the prompts due to insufficient model capabilities (such as prohibiting multi-line comments, etc.), but after the new model gained autonomous decision-making capabilities, these rules instead triggered instruction conflicts and thought divergence. The new strategy shifted to describing goals (such as "write code that matches the surrounding code"), and distributed instructions into tool descriptions and skill modules enabled on demand. The team explicitly views prompt bloat as technical debt, believing that as model capabilities improve, instructions should be simplified rather than stacked.