Anthropic's new Claude Fable 5.1 docs: one prompt line removes 'Claude-speak'
daniel_mac8 · x · 2026-09-03
danielmac8 shares Anthropic's newly published prompting docs for Claude Fable 5.1, along with a handy trick: typing 'Remove all mannered prose' into your prompt or CLAUDE.md strips the Claude-speak.
The docs cover behavioral differences and prompting patterns, including effort levels, progress updates between tool calls, batching independent tool calls in agent loops, append-only conversation history, writing density and formatting, quoting retrieved sources, preserving constraints in compaction summaries, keeping changes and tests to the task, subagents, and vision.
Related event: Viral Prompt Trick Removes Claude's Mannered Prose(4 posts)→
More from coding & agent
- Zhipu boosts GLM Coding Plan: unlimited GLM-5.3-Flash in ZCode through Sep 20 — Zai_org · 2026-09-03
- How One Founder Cut a 10-Hour Workday to 2 With an Email Agent — JanJanJaJa · 2026-09-03
- Agentic SREs cut MTTR below 20 minutes, but reasoning vs SLA trade-off remains — JiaZhihao · 2026-09-03
- You don't need Python to learn agents: start from organizational theory, says KarlMuth — KarlMuth · 2026-09-03
- Stop letting Claude burn your subscription: a 10x token-saving workflow — alexcovo_eth · 2026-09-03
- Automation vs Fixed Agent vs Malleable Agent: A Simple Framework — dfinke · 2026-09-03