Claude Code 2.1.128 fixes worktrees, OTEL

// 45d agoPRODUCT UPDATE

Claude Code 2.1.128 fixes worktrees, OTEL

Claude Code 2.1.128 tightens up the CLI with a better EnterWorktree default, cleaner subprocess env isolation, and a batch of polish fixes across focus mode, notifications, MCP, and plugin loading. It reads like a release aimed at reducing footguns for people using Claude Code in real terminal workflows.

// ANALYSIS

This is a small but meaningful hardening release: the headline worktree change prevents unpushed commits from disappearing, and the OTEL fix stops Claude Code from leaking its own tracing config into child processes.

–EnterWorktree now branches from local HEAD, which matters for anyone working with dirty local history or unpublished commits
–Stripping `OTEL_*` from subprocesses avoids surprising telemetry behavior in Bash, hooks, MCP, and LSP workloads
–The rest of the patch set is quality-of-life debt paydown: fewer visual glitches, better Remote Control behavior, and cleaner `/mcp` and `/plugin` handling
–Adding zip support to `--plugin-dir` makes plugin distribution a bit less awkward for teams that want to ship packaged extensions
–The update reinforces Claude Code’s positioning as a serious terminal-native agent, not just a wrapper around chat completions

// TAGS

claude-codeclidevtoolagentcoding-agentai-codingautomationobservability

DISCOVERED

45d ago

2026-05-04

PUBLISHED

45d ago

2026-05-04

RELEVANCE

9/ 10

AUTHOR

ClaudeCodeLog

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE26m ago

VS Code adds Enterprise-Managed Authorization for MCP

VS Code now supports Enterprise-Managed Authorization for Model Context Protocol (MCP), allowing developers to access organization-approved MCP servers using their work accounts. This update eliminates the need for repeated OAuth setups and connector-by-connector consent flows, ensuring that enterprise authorization remains centralized and seamless.

MODEL1h ago

GLM-5.2 rivals Claude Opus on Fireworks

Zhipu AI's newly released GLM-5.2 model is receiving praise in the developer community as an open-weights rival to Anthropic's Claude Opus 4.8. Boasting a 1M-token context window optimized on Fireworks AI for low latency, GLM-5.2 offers a compelling alternative for enterprise developers building complex AI agents.

MODEL1h ago

GLM-5.2 hits Fireworks inference

Z.ai's GLM-5.2 is now available on Fireworks serverless inference with open weights, an MIT license, 1M-token context, and coding-first positioning for long-horizon agent workflows.