One sentence makes Claude Fable 5.2 refactor boldly: say the codebase is its own
brandon_galang · x · 2026-11-11
Anthropic's testing found Claude Fable 5.2 is highly sensitive to how a prompt frames code ownership:
- When the prompt says nothing about ownership or frames the codebase as someone else's, the model adds to the existing structure — duplicating code rather than deleting — and reports nearby problems instead of fixing them.
- When the prompt says the codebase is the model's own, it often merges duplicate code and makes related fixes unprompted.
The trick is adding one sentence:
> This codebase is yours: what it's for and how it's built are your call, and you'll keep building on it.
Numbers: on small coding tasks, the model merged repeated code in about half of tries with the sentence vs. under a tenth without; the effect held inside a long system prompt, and feature completion was near-perfect either way.
Related event: Prompting Trick: Tell Claude the Codebase Is Its Own for Bolder Refactors(2 posts)→
More from coding & agent
- Gemini 4 Carbon Developer Guide: When to Pick Carbon, Thinking Levels & Verify Loops — rseroter · 2026-11-11
- LaunchDarkly launches Release Autopilot; already ships 20% of PRs agent-direct — annetgriffin · 2026-11-11
- Mercor pays experts $5M/day as it bets big on RL environments for vertical models — ccerrato147 · 2026-11-11
- Mercor acquired Deeptune, building hundreds of simulated apps for agents to practice in — ccerrato147 · 2026-11-11
- Peter Yang: building a kart racer with Fable all morning — actually a good game, not one-shot slop — petergyang · 2026-11-11
- Anthropic's real bet is MCP connectors embedding HubSpot, Adobe and Atlassian inside Claude — petrusenko_max · 2026-11-11