Anthropic cuts 80% of Claude Code’s system prompt with no eval drop
rohanpaul_ai · x · 2026-07-26
Anthropic says it removed more than 80% of Claude Code’s system prompt without hurting coding evals.
The post argues that the old prompt reflected the needs of earlier models: strict guardrails to prevent bad file operations or unprofessional output. As models improved, those rules started to conflict with each other and waste context on reconciling instructions instead of solving the task.
The new guidance shifts from long rule lists to:
- giving Claude judgment instead of exhaustive constraints
- designing interfaces and tools more carefully
- using progressive disclosure
- relying on simple tool descriptions
- leaning on auto-memory and richer references
The broader takeaway is that context engineering for capable models should be less about over-specifying behavior and more about structuring the task so the model can act intelligently.
Related event: Anthropic Slashes Claude Code System Prompts by Over 80%(11 posts)→
More from coding & agent
- SF 'Code Engineering' event picks up where 1,000-person harness engineering meetup left off — LangChain · 2026-10-27
- Prompt turns Claude Code into an iPhone backup forensics tool exposing 30-day 'deleted' photo copies — alex_verem · 2026-10-27
- Prompt turns Claude Code into a forensics tool finding deleted iPhone photos in backups — alex_verem · 2026-10-27
- Phones won't be mobile IDEs but control planes for agent fleets, argues Yuchen Jin — Yuchenj_UW · 2026-10-27
- FrontierCS 2 launches on Kaggle: 48 tasks make agents beat published CS papers — MeganRisdal · 2026-10-27
- Stop remaking old games with AI: a call for more ambitious agentic software — rschu · 2026-10-27