spacehub

surshar / claude-code

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

Built by Anthropic · github.com/anthropics

claude-code / CHANGELOG.md
GitHub Actionschore: Update CHANGELOG.md 4dc23d0 · on Oct 28, 2025 History
856 lines · 31.1 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.28
4
5- Plan mode: introduced new Plan subagent
6- Subagents: claude can now choose to resume subagents
7- Subagents: claude can dynamically choose the model used by its subagents
8- SDK: added --max-budget-usd flag
9- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
10- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
11- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
12- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
13- Various other bug fixes
14
15## 2.0.27
16
17- New UI for permission prompts
18- Added current branch filtering and search to session resume screen for easier navigation
19- Fixed directory @-mention causing "No assistant message found" error
20- VSCode Extension: Add config setting to include .gitignored files in file searches
21- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
22
23## 2.0.25
24
25- Removed legacy SDK entrypoint. Please migrate to @anthropic-ai/claude-agent-sdk for future SDK updates: https://docs.claude.com/en/docs/claude-code/sdk/migration-guide
26
27## 2.0.24
28
29- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
30- Claude Code Web: Support for Web -> CLI teleport
31- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
32- Bedrock: Display awsAuthRefresh output when auth is required
33
34## 2.0.22
35
36- Fixed content layout shift when scrolling through slash commands
37- IDE: Add toggle to enable/disable thinking.
38- Fix bug causing duplicate permission prompts with parallel tool calls
39- Add support for enterprise managed MCP allowlist and denylist
40
41## 2.0.21
42
43- Support MCP `structuredContent` field in tool responses
44- Added an interactive question tool
45- Claude will now ask you questions more often in plan mode
46- Added Haiku 4.5 as a model option for Pro users
47- Fixed an issue where queued commands don't have access to previous messages' output
48
49## 2.0.20
50
51- Added support for Claude Skills
52
53## 2.0.19
54
55- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
56- Fixed a bug where Haiku was unnecessarily called in print mode
57
58## 2.0.17
59
60- Added Haiku 4.5 to model selector!
61- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
62- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
63- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
64- OTEL: support HTTP_PROXY and HTTPS_PROXY
65- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
66
67## 2.0.15
68
69- Fixed bug with resuming where previously created files needed to be read again before writing
70- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
71
72## 2.0.14
73
74- Fix @-mentioning MCP servers to toggle them on/off
75- Improve permission checks for bash with inline env vars
76- Fix ultrathink + thinking toggle
77- Reduce unnecessary logins
78- Document --system-prompt
79- Several improvements to rendering
80- Plugins UI polish
81
82## 2.0.13
83
84- Fixed `/plugin` not working on native build
85
86## 2.0.12
87
88- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
89- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
90- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
91- `/plugin validate` command for validating plugin structure and configuration
92- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
93- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
94- Comprehensive error messages and diagnostics via `/doctor` command
95- Avoid flickering in `/model` selector
96- Improvements to `/help`
97- Avoid mentioning hooks in `/resume` summaries
98- Changes to the "verbose" setting in `/config` now persist across sessions
99
100## 2.0.11
101
102- Reduced system prompt size by 1.4k tokens
103- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
104- Fixed Opus fallback rate limit errors appearing incorrectly
105- Fixed /add-dir command selecting wrong default tab
106
107## 2.0.10
108
109- Rewrote terminal renderer for buttery smooth UI
110- Enable/disable MCP servers by @mentioning, or in /mcp
111- Added tab completion for shell commands in bash mode
112- PreToolUse hooks can now modify tool inputs
113- Press Ctrl-G to edit your prompt in your system's configured text editor
114- Fixes for bash permission checks with environment variables in the command
115
116## 2.0.9
117
118- Fix regression where bash backgrounding stopped working
119
120## 2.0.8
121
122- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
123- IDE: Add drag-and-drop support for files and folders in chat
124- /context: Fix counting for thinking blocks
125- Improve message rendering for users with light themes on dark terminals
126- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
127
128## 2.0.5
129
130- IDE: Fix IME unintended message submission with Enter and Tab
131- IDE: Add "Open in Terminal" link in login screen
132- Fix unhandled OAuth expiration 401 API errors
133- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
134
135## 2.0.1
136
137- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
138- Various bug fixes and presentation improvements
139
140## 2.0.0
141
142- New native VS Code extension
143- Fresh coat of paint throughout the whole app
144- /rewind a conversation to undo code changes
145- /usage command to see plan limits
146- Tab to toggle thinking (sticky across sessions)
147- Ctrl-R to search history
148- Unshipped claude config command
149- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
150- SDK: The Claude Code SDK is now the Claude Agent SDK
151- Add subagents dynamically with `--agents` flag
152
153## 1.0.126
154
155- Enable /context command for Bedrock and Vertex
156- Add mTLS support for HTTP-based OpenTelemetry exporters
157
158## 1.0.124
159
160- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
161- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
162- No longer inform Claude of the list of allowed tools when permission is denied
163- Fixed security vulnerability in Bash tool permission checks
164- Improved VSCode extension performance for large files
165
166## 1.0.123
167
168- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
169- Fixed thinking mode triggering on negation phrases like "don't think"
170- Fixed rendering performance degradation during token streaming
171- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
172- Enhanced BashTool environment snapshot logging
173- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
174- Migrated --debug logging to a file, to enable easy tailing & filtering
175
176## 1.0.120
177
178- Fix input lag during typing, especially noticeable with large prompts
179- Improved VSCode extension command registry and sessions dialog user experience
180- Enhanced sessions dialog responsiveness and visual feedback
181- Fixed IDE compatibility issue by removing worktree support check
182- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
183
184## 1.0.119
185
186- Fix Windows issue where process visually freezes on entering interactive mode
187- Support dynamic headers for MCP servers via headersHelper configuration
188- Fix thinking mode not working in headless sessions
189- Fix slash commands now properly update allowed tools instead of replacing them
190
191## 1.0.117
192
193- Add Ctrl-R history search to recall previous commands like bash/zsh
194- Fix input lag while typing, especially on Windows
195- Add sed command to auto-allowed commands in acceptEdits mode
196- Fix Windows PATH comparison to be case-insensitive for drive letters
197- Add permissions management hint to /add-dir output
198
199## 1.0.115
200
201- Improve thinking mode display with enhanced visual effects
202- Type /t to temporarily disable thinking mode in your prompt
203- Improve path validation for glob and grep tools
204- Show condensed output for post-tool hooks to reduce visual clutter
205- Fix visual feedback when loading state completes
206- Improve UI consistency for permission request dialogs
207
208## 1.0.113
209
210- Deprecated piped input in interactive mode
211- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
212
213## 1.0.112
214
215- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
216- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
217- Hooks: Added systemMessage support for SessionEnd hooks
218- Added `spinnerTipsEnabled` setting to disable spinner tips
219- IDE: Various improvements and bug fixes
220
221## 1.0.111
222
223- /model now validates provided model names
224- Fixed Bash tool crashes caused by malformed shell syntax parsing
225
226## 1.0.110
227
228- /terminal-setup command now supports WezTerm
229- MCP: OAuth tokens now proactively refresh before expiration
230- Fixed reliability issues with background Bash processes
231
232## 1.0.109
233
234- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
235
236## 1.0.106
237
238- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
239
240## 1.0.97
241
242- Settings: /doctor now validates permission rule syntax and suggests corrections
243
244## 1.0.94
245
246- Vertex: add support for global endpoints for supported models
247- /memory command now allows direct editing of all imported memory files
248- SDK: Add custom tools as callbacks
249- Added /todos command to list current todo items
250
251## 1.0.93
252
253- Windows: Add alt + v shortcut for pasting images from clipboard
254- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
255
256## 1.0.90
257
258- Settings file changes take effect immediately - no restart required
259
260## 1.0.88
261
262- Fixed issue causing "OAuth authentication is currently not supported"
263- Status line input now includes `exceeds_200k_tokens`
264- Fixed incorrect usage tracking in /cost.
265- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
266- Bedrock: Updated default Sonnet model to Sonnet 4
267
268## 1.0.86
269
270- Added /context to help users self-serve debug context issues
271- SDK: Added UUID support for all SDK messages
272- SDK: Added `--replay-user-messages` to replay user messages back to stdout
273
274## 1.0.85
275
276- Status line input now includes session cost info
277- Hooks: Introduced SessionEnd hook
278
279## 1.0.84
280
281- Fix tool_use/tool_result id mismatch error when network is unstable
282- Fix Claude sometimes ignoring real-time steering when wrapping up a task
283- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
284- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
285
286## 1.0.83
287
288- @-mention: Support files with spaces in path
289- New shimmering spinner
290
291## 1.0.82
292
293- SDK: Add request cancellation support
294- SDK: New additionalDirectories option to search custom paths, improved slash command processing
295- Settings: Validation prevents invalid fields in .claude/settings.json files
296- MCP: Improve tool name consistency
297- Bash: Fix crash when Claude tries to automatically read large files
298
299## 1.0.81
300
301- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
302- Agents: Fix custom agent loading when agent files are unparsable
303
304## 1.0.80
305
306- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
307
308## 1.0.77
309
310- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
311- SDK: Add session support and permission denial tracking
312- Fix token limit errors in conversation summarization
313- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
314
315## 1.0.73
316
317- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
318- MCP: Press Esc to cancel OAuth authentication flows
319- Bash: Improved command validation and reduced false security warnings
320- UI: Enhanced spinner animations and status line visual hierarchy
321- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
322
323## 1.0.72
324
325- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
326
327## 1.0.71
328
329- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
330- Customizable status line: add your terminal prompt to Claude Code with /statusline
331
332## 1.0.70
333
334- Performance: Optimized message rendering for better performance with large contexts
335- Windows: Fixed native file search, ripgrep, and subagent functionality
336- Added support for @-mentions in slash command arguments
337
338## 1.0.69
339
340- Upgraded Opus to version 4.1
341
342## 1.0.68
343
344- Fix incorrect model names being used for certain commands like `/pr-comments`
345- Windows: improve permissions checks for allow / deny tools and project trust. This may create a new project entry in `.claude.json` - manually merge the history field if desired.
346- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
347- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
348- SDK: Added canUseTool callback support for tool confirmation
349- Added `disableAllHooks` setting
350- Improved file suggestions performance in large repos
351
352## 1.0.65
353
354- IDE: Fixed connection stability issues and error handling for diagnostics
355- Windows: Fixed shell environment setup for users without .bashrc files
356
357## 1.0.64
358
359- Agents: Added model customization support - you can now specify which model an agent should use
360- Agents: Fixed unintended access to the recursive agent tool
361- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
362- SDK: Fixed user input tracking across multi-turn conversations
363- Added hidden files to file search and @-mention suggestions
364
365## 1.0.63
366
367- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
368
369## 1.0.62
370
371- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
372- Hooks: Added SessionStart hook for new session initialization
373- /add-dir command now supports typeahead for directory paths
374- Improved network connectivity check reliability
375
376## 1.0.61
377
378- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
379- Settings: Added `--settings` flag to load settings from a JSON file
380- Settings: Fixed resolution of settings files paths that are symlinks
381- OTEL: Fixed reporting of wrong organization after authentication changes
382- Slash commands: Fixed permissions checking for allowed-tools with Bash
383- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
384- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
385- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
386
387## 1.0.60
388
389- You can now create custom subagents for specialized tasks! Run /agents to get started
390
391## 1.0.59
392
393- SDK: Added tool confirmation support with canUseTool callback
394- SDK: Allow specifying env for spawned process
395- Hooks: Exposed PermissionDecision to hooks (including "ask")
396- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
397- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
398
399## 1.0.58
400
401- Added support for reading PDFs
402- MCP: Improved server health status display in 'claude mcp list'
403- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
404
405## 1.0.57
406
407- Added support for specifying a model in slash commands
408- Improved permission messages to help Claude understand allowed tools
409- Fix: Remove trailing newlines from bash output in terminal wrapping
410
411## 1.0.56
412
413- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
414- Fixes for WSL IDE detection
415- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
416
417## 1.0.55
418
419- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
420- Windows: fixed Ctrl+Z crash
421- SDK: Added ability to capture error logging
422- Add --system-prompt-file option to override system prompt in print mode
423
424## 1.0.54
425
426- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
427- Custom slash commands: Added argument-hint to frontmatter
428- Windows: OAuth uses port 45454 and properly constructs browser URL
429- Windows: mode switching now uses alt + m, and plan mode renders properly
430- Shell: Switch to in-memory shell snapshot to fix file-related errors
431
432## 1.0.53
433
434- Updated @-mention file truncation from 100 lines to 2000 lines
435- Add helper script settings for AWS token refresh: awsAuthRefresh (for foreground operations like aws sso login) and awsCredentialExport (for background operation with STS-like response).
436
437## 1.0.52
438
439- Added support for MCP server instructions
440
441## 1.0.51
442
443- Added support for native Windows (requires Git for Windows)
444- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
445- Settings: /doctor can now help you identify and fix invalid setting files
446- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
447- Increased auto-compact warning threshold from 60% to 80%
448- Fixed an issue with handling user directories with spaces for shell snapshots
449- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
450- Custom slash commands: Fixed user-level commands in subdirectories
451- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
452
453## 1.0.48
454
455- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
456- Added progress messages to Bash tool based on the last 5 lines of command output
457- Added expanding variables support for MCP server configuration
458- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
459- Improved IDE extension path handling when Claude Code runs in WSL
460- Hooks: Added a PreCompact hook
461- Vim mode: Added c, f/F, t/T
462
463## 1.0.45
464
465- Redesigned Search (Grep) tool with new tool input parameters and features
466- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
467- Fixed config file corruption issue by enforcing atomic writes
468- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
469- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
470- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
471
472## 1.0.44
473
474- New /export command lets you quickly export a conversation for sharing
475- MCP: resource_link tool results are now supported
476- MCP: tool annotations and tool titles now display in /mcp view
477- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
478
479## 1.0.43
480
481- Fixed a bug where the theme selector was saving excessively
482- Hooks: Added EPIPE system error handling
483
484## 1.0.42
485
486- Added tilde (`~`) expansion support to `/add-dir` command
487
488## 1.0.41
489
490- Hooks: Split Stop hook triggering into Stop and SubagentStop
491- Hooks: Enabled optional timeout configuration for each command
492- Hooks: Added "hook_event_name" to hook input
493- Fixed a bug where MCP tools would display twice in tool list
494- New tool parameters JSON for Bash tool in `tool_decision` event
495
496## 1.0.40
497
498- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
499
500## 1.0.39
501
502- New Active Time metric in OpenTelemetry logging
503
504## 1.0.38
505
506- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://docs.claude.com/en/docs/claude-code/hooks
507
508## 1.0.37
509
510- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
511
512## 1.0.36
513
514- Web search now takes today's date into context
515- Fixed a bug where stdio MCP servers were not terminating properly on exit
516
517## 1.0.35
518
519- Added support for MCP OAuth Authorization Server discovery
520
521## 1.0.34
522
523- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
524
525## 1.0.33
526
527- Improved logging functionality with session ID support
528- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
529- Improvements to plan mode
530
531## 1.0.32
532
533- Updated loopback config for litellm
534- Added forceLoginMethod setting to bypass login selection screen
535
536## 1.0.31
537
538- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
539
540## 1.0.30
541
542- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
543- Improved file path autocomplete with filename matching
544- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
545- Enhanced jq regex support for complex filters with pipes and select
546
547## 1.0.29
548
549- Improved CJK character support in cursor navigation and rendering
550
551## 1.0.28
552
553- Slash commands: Fix selector display during history navigation
554- Resizes images before upload to prevent API size limit errors
555- Added XDG_CONFIG_HOME support to configuration directory
556- Performance optimizations for memory usage
557- New attributes (terminal.type, language) in OpenTelemetry logging
558
559## 1.0.27
560
561- Streamable HTTP MCP servers are now supported
562- Remote MCP servers (SSE and HTTP) now support OAuth
563- MCP resources can now be @-mentioned
564- /resume slash command to switch conversations within Claude Code
565
566## 1.0.25
567
568- Slash commands: moved "project" and "user" prefixes to descriptions
569- Slash commands: improved reliability for command discovery
570- Improved support for Ghostty
571- Improved web search reliability
572
573## 1.0.24
574
575- Improved /mcp output
576- Fixed a bug where settings arrays got overwritten instead of merged
577
578## 1.0.23
579
580- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
581- Released Python SDK: pip install claude-code-sdk to get started
582
583## 1.0.22
584
585- SDK: Renamed `total_cost` to `total_cost_usd`
586
587## 1.0.21
588
589- Improved editing of files with tab-based indentation
590- Fix for tool_use without matching tool_result errors
591- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
592
593## 1.0.18
594
595- Added --add-dir CLI argument for specifying additional working directories
596- Added streaming input support without require -p flag
597- Improved startup performance and session storage performance
598- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
599- Added detailed MCP server tools display (/mcp)
600- MCP authentication and permission improvements
601- Added auto-reconnection for MCP SSE connections on disconnect
602- Fixed issue where pasted content was lost when dialogs appeared
603
604## 1.0.17
605
606- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
607- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
608- MCP server list UI improvements
609- Update Claude Code process title to display "claude" instead of "node"
610
611## 1.0.11
612
613- Claude Code can now also be used with a Claude Pro subscription
614- Added /upgrade for smoother switching to Claude Max plans
615- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
616- Improved shell configuration error handling
617- Improved todo list handling during compaction
618
619## 1.0.10
620
621- Added markdown table support
622- Improved streaming performance
623
624## 1.0.8
625
626- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
627- Increased default otel interval from 1s -> 5s
628- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
629- Fixed a regression where search tools unnecessarily asked for permissions
630- Added support for triggering thinking non-English languages
631- Improved compacting UI
632
633## 1.0.7
634
635- Renamed /allowed-tools -> /permissions
636- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
637- Deprecated claude config commands in favor of editing settings.json
638- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
639- Improved error handling for /install-github-app
640- Bugfixes, UI polish, and tool reliability improvements
641
642## 1.0.6
643
644- Improved edit reliability for tab-indented files
645- Respect CLAUDE_CONFIG_DIR everywhere
646- Reduced unnecessary tool permission prompts
647- Added support for symlinks in @file typeahead
648- Bugfixes, UI polish, and tool reliability improvements
649
650## 1.0.4
651
652- Fixed a bug where MCP tool errors weren't being parsed correctly
653
654## 1.0.1
655
656- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
657- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
658- Updated documentation links and OAuth process descriptions
659
660## 1.0.0
661
662- Claude Code is now generally available
663- Introducing Sonnet 4 and Opus 4 models
664
665## 0.2.125
666
667- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
668- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
669
670## 0.2.117
671
672- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
673- Introduced settings.cleanupPeriodDays
674- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
675- Introduced --debug mode
676
677## 0.2.108
678
679- You can now send messages to Claude while it works to steer Claude in real-time
680- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
681- Fixed a bug where thinking was not working in -p mode
682- Fixed a regression in /cost reporting
683- Deprecated MCP wizard interface in favor of other MCP commands
684- Lots of other bugfixes and improvements
685
686## 0.2.107
687
688- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
689
690## 0.2.106
691
692- MCP SSE server configs can now specify custom headers
693- Fixed a bug where MCP permission prompt didn't always show correctly
694
695## 0.2.105
696
697- Claude can now search the web
698- Moved system & account status to /status
699- Added word movement keybindings for Vim
700- Improved latency for startup, todo tool, and file edits
701
702## 0.2.102
703
704- Improved thinking triggering reliability
705- Improved @mention reliability for images and folders
706- You can now paste multiple large chunks into one prompt
707
708## 0.2.100
709
710- Fixed a crash caused by a stack overflow error
711- Made db storage optional; missing db support disables --continue and --resume
712
713## 0.2.98
714
715- Fixed an issue where auto-compact was running twice
716
717## 0.2.96
718
719- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
720
721## 0.2.93
722
723- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
724- Claude now has access to a Todo list that helps it stay on track and be more organized
725
726## 0.2.82
727
728- Added support for --disallowedTools
729- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
730
731## 0.2.75
732
733- Hit Enter to queue up additional messages while Claude is working
734- Drag in or copy/paste image files directly into the prompt
735- @-mention files to directly add them to context
736- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
737- Improved performance for filename auto-complete
738
739## 0.2.74
740
741- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
742- Task tool can now perform writes and run bash commands
743
744## 0.2.72
745
746- Updated spinner to indicate tokens loaded and tool usage
747
748## 0.2.70
749
750- Network commands like curl are now available for Claude to use
751- Claude can now run multiple web queries in parallel
752- Pressing ESC once immediately interrupts Claude in Auto-accept mode
753
754## 0.2.69
755
756- Fixed UI glitches with improved Select component behavior
757- Enhanced terminal output display with better text truncation logic
758
759## 0.2.67
760
761- Shared project permission rules can be saved in .claude/settings.json
762
763## 0.2.66
764
765- Print mode (-p) now supports streaming output via --output-format=stream-json
766- Fixed issue where pasting could trigger memory or bash mode unexpectedly
767
768## 0.2.63
769
770- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
771
772## 0.2.61
773
774- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
775- Enhanced image detection for more reliable clipboard paste functionality
776- Fixed an issue where ESC key could crash the conversation history selector
777
778## 0.2.59
779
780- Copy+paste images directly into your prompt
781- Improved progress indicators for bash and fetch tools
782- Bugfixes for non-interactive mode (-p)
783
784## 0.2.54
785
786- Quickly add to Memory by starting your message with '#'
787- Press ctrl+r to see full output for long tool results
788- Added support for MCP SSE transport
789
790## 0.2.53
791
792- New web fetch tool lets Claude view URLs that you paste in
793- Fixed a bug with JPEG detection
794
795## 0.2.50
796
797- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
798
799## 0.2.49
800
801- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
802
803## 0.2.47
804
805- Press Tab to auto-complete file and folder names
806- Press Shift + Tab to toggle auto-accept for file edits
807- Automatic conversation compaction for infinite conversation length (toggle with /config)
808
809## 0.2.44
810
811- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
812
813## 0.2.41
814
815- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
816- MCP server startup no longer blocks the app from starting up
817
818## 0.2.37
819
820- New /release-notes command lets you view release notes at any time
821- `claude config add/remove` commands now accept multiple values separated by commas or spaces
822
823## 0.2.36
824
825- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
826- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
827
828## 0.2.34
829
830- Vim bindings for text input - enable with /vim or /config
831
832## 0.2.32
833
834- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
835- Fix for some PersistentShell issues
836
837## 0.2.31
838
839- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
840- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
841
842## 0.2.30
843
844- Added ANSI color theme for better terminal compatibility
845- Fixed issue where slash command arguments weren't being sent properly
846- (Mac-only) API keys are now stored in macOS Keychain
847
848## 0.2.26
849
850- New /approved-tools command for managing tool permissions
851- Word-level diff display for improved code readability
852- Fuzzy matching for slash commands
853
854## 0.2.21
855
856- Fuzzy matching for /commands