This release adds 3 notable features for engineering teams evaluating rollout.
✓ No known CVEs patched in this version
Summary
AI summaryFixed policies blocking third‑party provider sessions alongside org pin.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Low |
OTEL_RESOURCE_ATTRIBUTES values are now included as labels on metric datapoints. OTEL_RESOURCE_ATTRIBUTES values are now included as labels on metric datapoints. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
| Feature | Low |
`claude agents` rows display `done/total` before detail when work is fanned out; peek shows longest‑running item. `claude agents` rows display `done/total` before detail when work is fanned out; peek shows longest‑running item. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
| Feature | Low |
/mcp collapses unused claude.ai connectors behind a "Show unused connectors" row. /mcp collapses unused claude.ai connectors behind a "Show unused connectors" row. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
| Feature | Low |
Fullscreen mode on Linux uses `wl-copy`, `xclip`, or `xsel` to copy to both clipboard and PRIMARY selection for middle‑click paste, and shows the correct key hint. Fullscreen mode on Linux uses `wl-copy`, `xclip`, or `xsel` to copy to both clipboard and PRIMARY selection for middle‑click paste, and shows the correct key hint. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
| Feature | Low |
VSCode tip suggests disabling terminal GPU acceleration (or running `/terminal-setup`) to fix garbled glyphs. VSCode tip suggests disabling terminal GPU acceleration (or running `/terminal-setup`) to fix garbled glyphs. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
| Performance | Low |
Terminal rendering performance improved by stabilizing the layout engine's JIT compilation profile. Terminal rendering performance improved by stabilizing the layout engine's JIT compilation profile. Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Performance | Low |
Rendering performance for large file writes has been improved. Rendering performance for large file writes has been improved. Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
/usage-credits now points to organization's usage settings page for Team and Enterprise admins instead of triggering re‑login. /usage-credits now points to organization's usage settings page for Team and Enterprise admins instead of triggering re‑login. Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
/autofix-pr no longer reports "cannot run on the default branch" when session is inside a git worktree or another repository. /autofix-pr no longer reports "cannot run on the default branch" when session is inside a git worktree or another repository. Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
`--resume` picker now shows sessions from the current directory even when it isn't a git worktree (e.g., jj workspaces). `--resume` picker now shows sessions from the current directory even when it isn't a git worktree (e.g., jj workspaces). Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
Windows hooks that invoke bash explicitly no longer fail with "command not found" or "cannot execute binary file". Windows hooks that invoke bash explicitly no longer fail with "command not found" or "cannot execute binary file". Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
OpenTelemetry log events are no longer silently dropped when emitted before telemetry initialization completes. OpenTelemetry log events are no longer silently dropped when emitted before telemetry initialization completes. Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
Background sessions dispatched from `claude agents` now use the model specified in `settings.json` instead of a stale daemon model. Background sessions dispatched from `claude agents` now use the model specified in `settings.json` instead of a stale daemon model. Source: llm_adapter@2026-06-03 Confidence: high |
— |
| Bugfix | Medium |
`forceLoginOrgUUID`/`forceLoginMethod` policies no longer block third‑party provider sessions alongside org pin. `forceLoginOrgUUID`/`forceLoginMethod` policies no longer block third‑party provider sessions alongside org pin. Source: llm_adapter@2026-06-03 Confidence: low |
— |
| Bugfix | Medium |
`claude mcp` list/get/add no longer print secrets; `${VAR}` references are not expanded and credential headers/URLs are redacted. `claude mcp` list/get/add no longer print secrets; `${VAR}` references are not expanded and credential headers/URLs are redacted. Source: llm_adapter@2026-06-03 Confidence: low |
— |
| Bugfix | Low |
Parallel tool calls no longer cancel other calls when a Bash command fails; each returns its own result independently. Parallel tool calls no longer cancel other calls when a Bash command fails; each returns its own result independently. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
| Bugfix | Low |
/effort dialog, workflow animations, and prompt keyword shimmer now honor the "Reduce motion" setting. /effort dialog, workflow animations, and prompt keyword shimmer now honor the "Reduce motion" setting. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
| Bugfix | Low |
Fixed potential crash when rendering Write tool results after resuming a session. Fixed potential crash when rendering Write tool results after resuming a session. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
| Bugfix | Low |
Completed subagents no longer get stuck showing as running when finalizing their result fails. Completed subagents no longer get stuck showing as running when finalizing their result fails. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
| Bugfix | Low |
Resolved `EADDRINUSE` errors from tools binding Unix sockets under `$TMPDIR` when `CLAUDE_CODE_TMPDIR` is a deep path. Resolved `EADDRINUSE` errors from tools binding Unix sockets under `$TMPDIR` when `CLAUDE_CODE_TMPDIR` is a deep path. Source: granite4.1:30b@2026-06-03-audit Confidence: low |
— |
Full changelog
What's changed
OTEL_RESOURCE_ATTRIBUTESvalues are now included as labels on metric datapoints, so you can slice usage metrics by custom dimensions like team or repoclaude agentsrows now showdone/totalbefore the detail when work is fanned out; peek shows the longest-running item/mcpnow collapses claude.ai connectors you've never signed in to behind a "Show unused connectors" row- Parallel tool calls: a failed Bash command no longer cancels other calls in the same batch — each tool returns its own result independently
- Fullscreen mode: clipboard now uses
wl-copy/xclip/xselon Linux when available, copies to both the clipboard and PRIMARY selection for middle-click paste, and the "hold {key} for native selection" hint now shows the correct key per terminal - Fixed the
/effortdialog, workflow animations, and prompt keyword shimmer not honoring the "Reduce motion" setting - Fixed
forceLoginOrgUUID/forceLoginMethodmanaged-settings policies blocking third-party provider sessions (Bedrock, Vertex, Foundry, Mantle) alongside the org pin (regression in 2.1.146) - Fixed background subagent output corrupting
claude -pstdout when using--output-format textorjson - Fixed
/usage-creditsstarting a re-login for Team and Enterprise admins instead of pointing to the organization's usage settings page - Fixed
/autofix-prreporting "cannot run on the default branch" when the session is inside a git worktree or another repository - Fixed
--resumepicker not showing sessions from the current directory when it isn't a git worktree (e.g., jj workspaces) - Fixed Windows hooks that invoke bash explicitly (e.g.,
/usr/bin/bash script.sh) failing with "command not found" or "cannot execute binary file" - Fixed OpenTelemetry log events (
user_prompt,api_request,tool_result,tool_decision) being silently dropped when emitted before telemetry initialization completed - Fixed
claude mcplist/get/add printing secrets to the terminal:${VAR}references are no longer expanded, and credential headers and URL secrets are redacted - Fixed Workflow agents spawned with
isolation: "worktree"in background sessions being blocked from editing files inside their own worktree - Fixed background sessions dispatched from
claude agentsbooting on a stale model from the daemon's environment instead of the model insettings.json - Fixed a potential crash when rendering Write tool results after resuming a session
- Fixed completed subagents getting stuck showing as running when an error occurs while finalizing their result
- Fixed
EADDRINUSEerrors from tools that bind Unix sockets under$TMPDIRwhenCLAUDE_CODE_TMPDIRis set to a deep path - Improved terminal rendering performance by stabilizing the layout engine's JIT compilation profile
- Improved rendering performance for large file writes
- [VSCode] Added a tip suggesting disabling terminal GPU acceleration (or running
/terminal-setup) to fix garbled glyphs
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
Share this release
About claude-code
All releases →Related context
Related tools
Earlier breaking changes
- v2.1.160 Renames dynamic‑workflow trigger keyword from `workflow` to `ultracode`; `workflow` no longer triggers a run
- v2.1.160 Deprecates and removes the `CLAUDE_CODE_OPUS_4_6_FAST_MODE_OVERRIDE` environment variable; it is now a no‑op
- v2.1.147 Renames /simplify to /code-review; removes cleanup-and-fix behavior.
Beta — feedback welcome: [email protected]