Claude Code v2.1.158: Auto Mode Now on Bedrock, Vertex, Foundry for Opus 4.7/4.8
Claude Code v2.1.158 extends auto mode to Bedrock, Vertex, and Foundry deployments running Opus 4.7 or Opus 4.8. Auto mode allows the agent to automatically determine when to run commands, edit files, or ask for clarification without requiring manual approval for each step.
Key Details
- New platforms: Bedrock, Vertex AI, and Foundry (in addition to existing support)
- Models: Opus 4.7 and Opus 4.8
- Opt-in: Set environment variable
CLAUDE_CODE_ENABLE_AUTO_MODE=1
To enable, launch Claude Code with the environment variable set:
CLAUDE_CODE_ENABLE_AUTO_MODE=1 claude codeOnce enabled, the agent will autonomously execute actions like file edits and shell commands, reducing user interaction overhead. This is particularly useful for long-running multi-step tasks where manual confirmation would break flow.
No other changes were included in this patch release.
📖 Read the full source: GitHub Claude-Code
👀 See Also
Anam Cara-3: Advancements in Interactive AI Avatars
Anam Cara-3 introduces advanced interactive avatars with a two-stage pipeline for audio-to-video conversion, achieving impressive speed and responsiveness.
OpenClaw 2026.3.13 regression causes false unreachable status reports
OpenClaw version 2026.3.13 introduced a diagnostic regression where status commands falsely report unreachable gateways despite RPC probes working correctly. Rolling back to 2026.3.12 resolves the issue.
Merlin Research releases Qwen3.5-4B-Safety-Thinking model for structured reasoning
Merlin Research has released Qwen3.5-4B-Safety-Thinking, a 4 billion parameter safety-aligned reasoning model built on Qwen3.5. The model is designed for structured 'thinking' and safety in real-world scenarios including agent systems.
Bonsai 1.7B Ternary Model Hits 442 T/s on M4 Max with Autonomously Tuned Metal Kernels
Autonomous agent ata optimized Metal kernels for Bonsai 1.7B Q2_0, achieving 442 t/s decode (+42%) and 4622 t/s prefill (+9%) on M4 Max vs unmodified llama.cpp.