Claude Code v2.1.267 ships maxEffortLevel cap and a dozen fixes
ashwin-ant · ghdev · 2026-09-10
anthropics/claude-code released v2.1.267.
- New maxEffortLevel setting (top-level or per-model under modelSettings) caps effort level across all providers including Bedrock, Vertex and Foundry
- New --system-prompt-snapshot off to render a fresh system prompt per request for prompt iteration
Fixes include: Cowork cloud scheduled tasks failing under sandboxing policies, blank /context output on mobile, broken shortcuts after tmux/ssh reconnect, large-output Workflow agent() calls wrongly refused in auto mode, a backslash path containment bypass for marketplaces on macOS/Linux, expired AWS/Google credentials retrying ten times before erroring, spurious turns when resuming after /compact, and parallel tool calls being dropped when reloading 5MB+ transcripts.
More from coding & agent
- Agent Field Kit launches $29 security runbooks for production AI agents — alvelda · 2026-10-27
- Cloudflare Durable Objects validated as top architecture for AI agent WebSocket servers — tristanbob · 2026-10-27
- GitHub Next's Chopin prototype explores real-time multiplayer agentic planning — max__drake · 2026-10-27
- Veynod: An Open-Source Shared Workspace Where AI Agents Collaborate Like a Team — TKOWNS · 2026-10-27
- TMLR paper: enforced thinking doesn't reliably make LLM agents disclose key info to users — SharonYixuanLi · 2026-10-27
- LLM-written Rust library bbtree verified: speed claims hold, memory savings reverse — lmoroney · 2026-10-27