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 8462b07 · on Oct 9, 2025 History
777 lines · 27.8 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.13
4
5- Fixed `/plugin` not working on native build
6
7## 2.0.12
8
9- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
10- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
11- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
12- `/plugin validate` command for validating plugin structure and configuration
13- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
14- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
15- Comprehensive error messages and diagnostics via `/doctor` command
16- Avoid flickering in `/model` selector
17- Improvements to `/help`
18- Avoid mentioning hooks in `/resume` summaries
19- Changes to the "verbose" setting in `/config` now persist across sessions
20
21## 2.0.11
22
23- Reduced system prompt size by 1.4k tokens
24- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
25- Fixed Opus fallback rate limit errors appearing incorrectly
26- Fixed /add-dir command selecting wrong default tab
27
28## 2.0.10
29
30- Rewrote terminal renderer for buttery smooth UI
31- Enable/disable MCP servers by @mentioning, or in /mcp
32- Added tab completion for shell commands in bash mode
33- PreToolUse hooks can now modify tool inputs
34- Press Ctrl-G to edit your prompt in your system's configured text editor
35- Fixes for bash permission checks with environment variables in the command
36
37## 2.0.9
38
39- Fix regression where bash backgrounding stopped working
40
41## 2.0.8
42
43- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
44- IDE: Add drag-and-drop support for files and folders in chat
45- /context: Fix counting for thinking blocks
46- Improve message rendering for users with light themes on dark terminals
47- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
48
49## 2.0.5
50
51- IDE: Fix IME unintended message submission with Enter and Tab
52- IDE: Add "Open in Terminal" link in login screen
53- Fix unhandled OAuth expiration 401 API errors
54- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
55
56## 2.0.1
57
58- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
59- Various bug fixes and presentation improvements
60
61## 2.0.0
62
63- New native VS Code extension
64- Fresh coat of paint throughout the whole app
65- /rewind a conversation to undo code changes
66- /usage command to see plan limits
67- Tab to toggle thinking (sticky across sessions)
68- Ctrl-R to search history
69- Unshipped claude config command
70- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
71- SDK: The Claude Code SDK is now the Claude Agent SDK
72- Add subagents dynamically with `--agents` flag
73
74## 1.0.126
75
76- Enable /context command for Bedrock and Vertex
77- Add mTLS support for HTTP-based OpenTelemetry exporters
78
79## 1.0.124
80
81- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
82- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
83- No longer inform Claude of the list of allowed tools when permission is denied
84- Fixed security vulnerability in Bash tool permission checks
85- Improved VSCode extension performance for large files
86
87## 1.0.123
88
89- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
90- Fixed thinking mode triggering on negation phrases like "don't think"
91- Fixed rendering performance degradation during token streaming
92- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
93- Enhanced BashTool environment snapshot logging
94- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
95- Migrated --debug logging to a file, to enable easy tailing & filtering
96
97## 1.0.120
98
99- Fix input lag during typing, especially noticeable with large prompts
100- Improved VSCode extension command registry and sessions dialog user experience
101- Enhanced sessions dialog responsiveness and visual feedback
102- Fixed IDE compatibility issue by removing worktree support check
103- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
104
105## 1.0.119
106
107- Fix Windows issue where process visually freezes on entering interactive mode
108- Support dynamic headers for MCP servers via headersHelper configuration
109- Fix thinking mode not working in headless sessions
110- Fix slash commands now properly update allowed tools instead of replacing them
111
112## 1.0.117
113
114- Add Ctrl-R history search to recall previous commands like bash/zsh
115- Fix input lag while typing, especially on Windows
116- Add sed command to auto-allowed commands in acceptEdits mode
117- Fix Windows PATH comparison to be case-insensitive for drive letters
118- Add permissions management hint to /add-dir output
119
120## 1.0.115
121
122- Improve thinking mode display with enhanced visual effects
123- Type /t to temporarily disable thinking mode in your prompt
124- Improve path validation for glob and grep tools
125- Show condensed output for post-tool hooks to reduce visual clutter
126- Fix visual feedback when loading state completes
127- Improve UI consistency for permission request dialogs
128
129## 1.0.113
130
131- Deprecated piped input in interactive mode
132- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
133
134## 1.0.112
135
136- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
137- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
138- Hooks: Added systemMessage support for SessionEnd hooks
139- Added `spinnerTipsEnabled` setting to disable spinner tips
140- IDE: Various improvements and bug fixes
141
142## 1.0.111
143
144- /model now validates provided model names
145- Fixed Bash tool crashes caused by malformed shell syntax parsing
146
147## 1.0.110
148
149- /terminal-setup command now supports WezTerm
150- MCP: OAuth tokens now proactively refresh before expiration
151- Fixed reliability issues with background Bash processes
152
153## 1.0.109
154
155- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
156
157## 1.0.106
158
159- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
160
161## 1.0.97
162
163- Settings: /doctor now validates permission rule syntax and suggests corrections
164
165## 1.0.94
166
167- Vertex: add support for global endpoints for supported models
168- /memory command now allows direct editing of all imported memory files
169- SDK: Add custom tools as callbacks
170- Added /todos command to list current todo items
171
172## 1.0.93
173
174- Windows: Add alt + v shortcut for pasting images from clipboard
175- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
176
177## 1.0.90
178
179- Settings file changes take effect immediately - no restart required
180
181## 1.0.88
182
183- Fixed issue causing "OAuth authentication is currently not supported"
184- Status line input now includes `exceeds_200k_tokens`
185- Fixed incorrect usage tracking in /cost.
186- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
187- Bedrock: Updated default Sonnet model to Sonnet 4
188
189## 1.0.86
190
191- Added /context to help users self-serve debug context issues
192- SDK: Added UUID support for all SDK messages
193- SDK: Added `--replay-user-messages` to replay user messages back to stdout
194
195## 1.0.85
196
197- Status line input now includes session cost info
198- Hooks: Introduced SessionEnd hook
199
200## 1.0.84
201
202- Fix tool_use/tool_result id mismatch error when network is unstable
203- Fix Claude sometimes ignoring real-time steering when wrapping up a task
204- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
205- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
206
207## 1.0.83
208
209- @-mention: Support files with spaces in path
210- New shimmering spinner
211
212## 1.0.82
213
214- SDK: Add request cancellation support
215- SDK: New additionalDirectories option to search custom paths, improved slash command processing
216- Settings: Validation prevents invalid fields in .claude/settings.json files
217- MCP: Improve tool name consistency
218- Bash: Fix crash when Claude tries to automatically read large files
219
220## 1.0.81
221
222- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
223- Agents: Fix custom agent loading when agent files are unparsable
224
225## 1.0.80
226
227- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
228
229## 1.0.77
230
231- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
232- SDK: Add session support and permission denial tracking
233- Fix token limit errors in conversation summarization
234- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
235
236## 1.0.73
237
238- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
239- MCP: Press Esc to cancel OAuth authentication flows
240- Bash: Improved command validation and reduced false security warnings
241- UI: Enhanced spinner animations and status line visual hierarchy
242- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
243
244## 1.0.72
245
246- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
247
248## 1.0.71
249
250- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
251- Customizable status line: add your terminal prompt to Claude Code with /statusline
252
253## 1.0.70
254
255- Performance: Optimized message rendering for better performance with large contexts
256- Windows: Fixed native file search, ripgrep, and subagent functionality
257- Added support for @-mentions in slash command arguments
258
259## 1.0.69
260
261- Upgraded Opus to version 4.1
262
263## 1.0.68
264
265- Fix incorrect model names being used for certain commands like `/pr-comments`
266- 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.
267- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
268- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
269- SDK: Added canUseTool callback support for tool confirmation
270- Added `disableAllHooks` setting
271- Improved file suggestions performance in large repos
272
273## 1.0.65
274
275- IDE: Fixed connection stability issues and error handling for diagnostics
276- Windows: Fixed shell environment setup for users without .bashrc files
277
278## 1.0.64
279
280- Agents: Added model customization support - you can now specify which model an agent should use
281- Agents: Fixed unintended access to the recursive agent tool
282- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
283- SDK: Fixed user input tracking across multi-turn conversations
284- Added hidden files to file search and @-mention suggestions
285
286## 1.0.63
287
288- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
289
290## 1.0.62
291
292- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
293- Hooks: Added SessionStart hook for new session initialization
294- /add-dir command now supports typeahead for directory paths
295- Improved network connectivity check reliability
296
297## 1.0.61
298
299- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
300- Settings: Added `--settings` flag to load settings from a JSON file
301- Settings: Fixed resolution of settings files paths that are symlinks
302- OTEL: Fixed reporting of wrong organization after authentication changes
303- Slash commands: Fixed permissions checking for allowed-tools with Bash
304- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
305- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
306- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
307
308## 1.0.60
309
310- You can now create custom subagents for specialized tasks! Run /agents to get started
311
312## 1.0.59
313
314- SDK: Added tool confirmation support with canUseTool callback
315- SDK: Allow specifying env for spawned process
316- Hooks: Exposed PermissionDecision to hooks (including "ask")
317- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
318- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
319
320## 1.0.58
321
322- Added support for reading PDFs
323- MCP: Improved server health status display in 'claude mcp list'
324- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
325
326## 1.0.57
327
328- Added support for specifying a model in slash commands
329- Improved permission messages to help Claude understand allowed tools
330- Fix: Remove trailing newlines from bash output in terminal wrapping
331
332## 1.0.56
333
334- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
335- Fixes for WSL IDE detection
336- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
337
338## 1.0.55
339
340- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
341- Windows: fixed Ctrl+Z crash
342- SDK: Added ability to capture error logging
343- Add --system-prompt-file option to override system prompt in print mode
344
345## 1.0.54
346
347- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
348- Custom slash commands: Added argument-hint to frontmatter
349- Windows: OAuth uses port 45454 and properly constructs browser URL
350- Windows: mode switching now uses alt + m, and plan mode renders properly
351- Shell: Switch to in-memory shell snapshot to fix file-related errors
352
353## 1.0.53
354
355- Updated @-mention file truncation from 100 lines to 2000 lines
356- 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).
357
358## 1.0.52
359
360- Added support for MCP server instructions
361
362## 1.0.51
363
364- Added support for native Windows (requires Git for Windows)
365- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
366- Settings: /doctor can now help you identify and fix invalid setting files
367- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
368- Increased auto-compact warning threshold from 60% to 80%
369- Fixed an issue with handling user directories with spaces for shell snapshots
370- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
371- Custom slash commands: Fixed user-level commands in subdirectories
372- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
373
374## 1.0.48
375
376- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
377- Added progress messages to Bash tool based on the last 5 lines of command output
378- Added expanding variables support for MCP server configuration
379- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
380- Improved IDE extension path handling when Claude Code runs in WSL
381- Hooks: Added a PreCompact hook
382- Vim mode: Added c, f/F, t/T
383
384## 1.0.45
385
386- Redesigned Search (Grep) tool with new tool input parameters and features
387- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
388- Fixed config file corruption issue by enforcing atomic writes
389- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
390- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
391- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
392
393## 1.0.44
394
395- New /export command lets you quickly export a conversation for sharing
396- MCP: resource_link tool results are now supported
397- MCP: tool annotations and tool titles now display in /mcp view
398- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
399
400## 1.0.43
401
402- Fixed a bug where the theme selector was saving excessively
403- Hooks: Added EPIPE system error handling
404
405## 1.0.42
406
407- Added tilde (`~`) expansion support to `/add-dir` command
408
409## 1.0.41
410
411- Hooks: Split Stop hook triggering into Stop and SubagentStop
412- Hooks: Enabled optional timeout configuration for each command
413- Hooks: Added "hook_event_name" to hook input
414- Fixed a bug where MCP tools would display twice in tool list
415- New tool parameters JSON for Bash tool in `tool_decision` event
416
417## 1.0.40
418
419- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
420
421## 1.0.39
422
423- New Active Time metric in OpenTelemetry logging
424
425## 1.0.38
426
427- 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
428
429## 1.0.37
430
431- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
432
433## 1.0.36
434
435- Web search now takes today's date into context
436- Fixed a bug where stdio MCP servers were not terminating properly on exit
437
438## 1.0.35
439
440- Added support for MCP OAuth Authorization Server discovery
441
442## 1.0.34
443
444- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
445
446## 1.0.33
447
448- Improved logging functionality with session ID support
449- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
450- Improvements to plan mode
451
452## 1.0.32
453
454- Updated loopback config for litellm
455- Added forceLoginMethod setting to bypass login selection screen
456
457## 1.0.31
458
459- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
460
461## 1.0.30
462
463- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
464- Improved file path autocomplete with filename matching
465- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
466- Enhanced jq regex support for complex filters with pipes and select
467
468## 1.0.29
469
470- Improved CJK character support in cursor navigation and rendering
471
472## 1.0.28
473
474- Slash commands: Fix selector display during history navigation
475- Resizes images before upload to prevent API size limit errors
476- Added XDG_CONFIG_HOME support to configuration directory
477- Performance optimizations for memory usage
478- New attributes (terminal.type, language) in OpenTelemetry logging
479
480## 1.0.27
481
482- Streamable HTTP MCP servers are now supported
483- Remote MCP servers (SSE and HTTP) now support OAuth
484- MCP resources can now be @-mentioned
485- /resume slash command to switch conversations within Claude Code
486
487## 1.0.25
488
489- Slash commands: moved "project" and "user" prefixes to descriptions
490- Slash commands: improved reliability for command discovery
491- Improved support for Ghostty
492- Improved web search reliability
493
494## 1.0.24
495
496- Improved /mcp output
497- Fixed a bug where settings arrays got overwritten instead of merged
498
499## 1.0.23
500
501- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
502- Released Python SDK: pip install claude-code-sdk to get started
503
504## 1.0.22
505
506- SDK: Renamed `total_cost` to `total_cost_usd`
507
508## 1.0.21
509
510- Improved editing of files with tab-based indentation
511- Fix for tool_use without matching tool_result errors
512- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
513
514## 1.0.18
515
516- Added --add-dir CLI argument for specifying additional working directories
517- Added streaming input support without require -p flag
518- Improved startup performance and session storage performance
519- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
520- Added detailed MCP server tools display (/mcp)
521- MCP authentication and permission improvements
522- Added auto-reconnection for MCP SSE connections on disconnect
523- Fixed issue where pasted content was lost when dialogs appeared
524
525## 1.0.17
526
527- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
528- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
529- MCP server list UI improvements
530- Update Claude Code process title to display "claude" instead of "node"
531
532## 1.0.11
533
534- Claude Code can now also be used with a Claude Pro subscription
535- Added /upgrade for smoother switching to Claude Max plans
536- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
537- Improved shell configuration error handling
538- Improved todo list handling during compaction
539
540## 1.0.10
541
542- Added markdown table support
543- Improved streaming performance
544
545## 1.0.8
546
547- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
548- Increased default otel interval from 1s -> 5s
549- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
550- Fixed a regression where search tools unnecessarily asked for permissions
551- Added support for triggering thinking non-English languages
552- Improved compacting UI
553
554## 1.0.7
555
556- Renamed /allowed-tools -> /permissions
557- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
558- Deprecated claude config commands in favor of editing settings.json
559- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
560- Improved error handling for /install-github-app
561- Bugfixes, UI polish, and tool reliability improvements
562
563## 1.0.6
564
565- Improved edit reliability for tab-indented files
566- Respect CLAUDE_CONFIG_DIR everywhere
567- Reduced unnecessary tool permission prompts
568- Added support for symlinks in @file typeahead
569- Bugfixes, UI polish, and tool reliability improvements
570
571## 1.0.4
572
573- Fixed a bug where MCP tool errors weren't being parsed correctly
574
575## 1.0.1
576
577- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
578- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
579- Updated documentation links and OAuth process descriptions
580
581## 1.0.0
582
583- Claude Code is now generally available
584- Introducing Sonnet 4 and Opus 4 models
585
586## 0.2.125
587
588- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
589- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
590
591## 0.2.117
592
593- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
594- Introduced settings.cleanupPeriodDays
595- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
596- Introduced --debug mode
597
598## 0.2.108
599
600- You can now send messages to Claude while it works to steer Claude in real-time
601- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
602- Fixed a bug where thinking was not working in -p mode
603- Fixed a regression in /cost reporting
604- Deprecated MCP wizard interface in favor of other MCP commands
605- Lots of other bugfixes and improvements
606
607## 0.2.107
608
609- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
610
611## 0.2.106
612
613- MCP SSE server configs can now specify custom headers
614- Fixed a bug where MCP permission prompt didn't always show correctly
615
616## 0.2.105
617
618- Claude can now search the web
619- Moved system & account status to /status
620- Added word movement keybindings for Vim
621- Improved latency for startup, todo tool, and file edits
622
623## 0.2.102
624
625- Improved thinking triggering reliability
626- Improved @mention reliability for images and folders
627- You can now paste multiple large chunks into one prompt
628
629## 0.2.100
630
631- Fixed a crash caused by a stack overflow error
632- Made db storage optional; missing db support disables --continue and --resume
633
634## 0.2.98
635
636- Fixed an issue where auto-compact was running twice
637
638## 0.2.96
639
640- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
641
642## 0.2.93
643
644- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
645- Claude now has access to a Todo list that helps it stay on track and be more organized
646
647## 0.2.82
648
649- Added support for --disallowedTools
650- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
651
652## 0.2.75
653
654- Hit Enter to queue up additional messages while Claude is working
655- Drag in or copy/paste image files directly into the prompt
656- @-mention files to directly add them to context
657- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
658- Improved performance for filename auto-complete
659
660## 0.2.74
661
662- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
663- Task tool can now perform writes and run bash commands
664
665## 0.2.72
666
667- Updated spinner to indicate tokens loaded and tool usage
668
669## 0.2.70
670
671- Network commands like curl are now available for Claude to use
672- Claude can now run multiple web queries in parallel
673- Pressing ESC once immediately interrupts Claude in Auto-accept mode
674
675## 0.2.69
676
677- Fixed UI glitches with improved Select component behavior
678- Enhanced terminal output display with better text truncation logic
679
680## 0.2.67
681
682- Shared project permission rules can be saved in .claude/settings.json
683
684## 0.2.66
685
686- Print mode (-p) now supports streaming output via --output-format=stream-json
687- Fixed issue where pasting could trigger memory or bash mode unexpectedly
688
689## 0.2.63
690
691- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
692
693## 0.2.61
694
695- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
696- Enhanced image detection for more reliable clipboard paste functionality
697- Fixed an issue where ESC key could crash the conversation history selector
698
699## 0.2.59
700
701- Copy+paste images directly into your prompt
702- Improved progress indicators for bash and fetch tools
703- Bugfixes for non-interactive mode (-p)
704
705## 0.2.54
706
707- Quickly add to Memory by starting your message with '#'
708- Press ctrl+r to see full output for long tool results
709- Added support for MCP SSE transport
710
711## 0.2.53
712
713- New web fetch tool lets Claude view URLs that you paste in
714- Fixed a bug with JPEG detection
715
716## 0.2.50
717
718- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
719
720## 0.2.49
721
722- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
723
724## 0.2.47
725
726- Press Tab to auto-complete file and folder names
727- Press Shift + Tab to toggle auto-accept for file edits
728- Automatic conversation compaction for infinite conversation length (toggle with /config)
729
730## 0.2.44
731
732- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
733
734## 0.2.41
735
736- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
737- MCP server startup no longer blocks the app from starting up
738
739## 0.2.37
740
741- New /release-notes command lets you view release notes at any time
742- `claude config add/remove` commands now accept multiple values separated by commas or spaces
743
744## 0.2.36
745
746- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
747- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
748
749## 0.2.34
750
751- Vim bindings for text input - enable with /vim or /config
752
753## 0.2.32
754
755- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
756- Fix for some PersistentShell issues
757
758## 0.2.31
759
760- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
761- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
762
763## 0.2.30
764
765- Added ANSI color theme for better terminal compatibility
766- Fixed issue where slash command arguments weren't being sent properly
767- (Mac-only) API keys are now stored in macOS Keychain
768
769## 0.2.26
770
771- New /approved-tools command for managing tool permissions
772- Word-level diff display for improved code readability
773- Fuzzy matching for slash commands
774
775## 0.2.21
776
777- Fuzzy matching for /commands