• Sources: primary
  • Summary: Anthropic's prompting documentation for Claude Fable 5.2 says the model follows what the prompt says about who owns the code. When the prompt said nothing or framed the code as someone else's, the model tended to add to existing structure even when that repeated code. With the sentence saying the codebase is the model's own and that it will keep building on it, it merged repeated code in about half of small coding tasks, against under a tenth without it.
  • Why it matters: Teams running Fable 5.2 in agent harnesses can steer toward refactoring or toward minimal additive changes with one line in the system prompt or project instructions.