releaseintermediate
[Release] anthropics/claude-code v2.1.126: v2.1.126
By ashwin-antgithub
View original on githubClaude Code v2.1.126 introduces gateway model listing, project purge functionality, improved OAuth authentication for WSL/SSH/containers, and numerous bug fixes. Key improvements include permission bypass options, better shell detection on Windows, security fixes for managed settings, and stability enhancements for streaming, image handling, and remote sessions. The release also refines telemetry events, fixes UI issues, and resolves edge cases in tool availability and clipboard handling.
Key Points
- •Model picker now lists models from gateway's `/v1/models` endpoint when using Anthropic-compatible gateways via `ANTHROPIC_BASE_URL`
- •New `claude project purge [path]` command deletes all Claude Code state (transcripts, tasks, file history, config) with `--dry-run`, `-y/--yes`, `-i/--interactive`, and `--all` flags
- •`--dangerously-skip-permissions` flag bypasses prompts for protected paths (`.claude/`, `.git/`, `.vscode/`, shell configs) while maintaining safety for catastrophic commands
- •OAuth login now accepts pasted codes in terminal when browser callback can't reach localhost, improving WSL2, SSH, and container support
- •`claude_code.skill_activated` OpenTelemetry event now includes `invocation_trigger` attribute tracking user-slash, claude-proactive, or nested-skill invocations
- •Security fix: `allowManagedDomainsOnly` and `allowManagedReadPathsOnly` settings now properly enforced when higher-priority sources lack sandbox blocks
- •Windows PowerShell 7 detection improved for Microsoft Store, MSI, and .NET global tool installations; PowerShell now treated as primary shell when enabled
- •Fixed critical streaming issues: idle timeouts during Mac sleep, long thinking pauses, empty turns, and IPv6-only devcontainers
- •Image handling improved: automatic downscaling on paste (>2000px), removal of oversized images from history, and session recovery
- •Fixed deferred tools (WebSearch, WebFetch) availability for skills with `context: fork` and subagents on first turn
Found this useful? Add it to a playbook for a step-by-step implementation guide.
Workflow Diagram
Start Process
Step A
Step B
Step C
Complete
Concepts
Artifacts (5)
Project Purge Commandcommand
claude project purge [path] [--dry-run] [-y|--yes] [-i|--interactive] [--all]Permission Bypass Flagcommand
--dangerously-skip-permissionsOAuth Login Commandcommand
claude auth loginGateway Model Configurationbashconfig
ANTHROPIC_BASE_URL=<gateway-url>
# Model picker will list models from gateway's /v1/models endpointOpenTelemetry Event Attributesconfig
Event: claude_code.skill_activated
Attributes:
- invocation_trigger: "user-slash" | "claude-proactive" | "nested-skill"