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 68ba478 · on Nov 18, 2025 History
960 lines · 36.9 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.45
4
5- Add support for Azure AI Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
6- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
7- Send background tasks to Claude Code on the web by starting a message with `&`
8
9## 2.0.43
10
11- Added `permissionMode` field for custom agents
12- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
13- Added skills frontmatter field to declare skills to auto-load for subagents
14- Added the `SubagentStart` hook event
15- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
16- Fixed duplicate rendering of some messages in the UI
17- Fixed some visual flickers
18- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
19
20## 2.0.42
21
22- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
23
24## 2.0.41
25
26- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
27- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
28- Fixed incorrect labeling of user settings vs project settings in command descriptions
29- Fixed crash when plugin command hooks timeout during execution
30- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
31- Fixed broken security documentation links in trust dialogs and onboarding
32- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
33- ctrl-r history search landing on a slash command no longer cancels the search
34- SDK: Support custom timeouts for hooks
35- Allow more safe git commands to run without approval
36- Plugins: Added support for sharing and installing output styles
37- Teleporting a session from web will automatically set the upstream branch
38
39## 2.0.37
40
41- Fixed how idleness is computed for notifications
42- Hooks: Added matcher values for Notification hook events
43- Output Styles: Added `keep-coding-instructions` option to frontmatter
44
45## 2.0.36
46
47- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
48- Fixed queued messages being incorrectly executed as bash commands
49- Fixed input being lost when typing while a queued message is processed
50
51## 2.0.35
52
53- Improve fuzzy search results when searching commands
54- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
55- Added `CLAUDE_CODE_EXIT_AFTER_STOP_DELAY` environment variable to automatically exit SDK mode after a specified idle duration, useful for automated workflows and scripts
56- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
57- Fixed messages returning null `stop_reason` and `stop_sequence` values
58- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
59
60## 2.0.34
61
62- VSCode Extension: Added setting to configure the initial permission mode for new conversations
63- Improved file path suggestion performance with native Rust-based fuzzy finder
64- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
65- Fixed memory crash when reading or writing large files (especially base64-encoded images)
66
67## 2.0.33
68
69- Native binary installs now launch quicker.
70- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
71- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
72
73## 2.0.32
74
75- Un-deprecate output styles based on community feedback
76- Added `companyAnnouncements` setting for displaying announcements on startup
77- Fixed hook progress messages not updating correctly during PostToolUse hook execution
78
79## 2.0.31
80
81- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
82- Vertex: add support for Web Search on supported models
83- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
84- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
85- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
86- Fixed plugin uninstall not removing plugins
87
88## 2.0.30
89
90- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
91- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
92- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
93- Added prompt-based stop hooks
94- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
95- Enabled SSE MCP servers on native build
96- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
97- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
98- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
99- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
100- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
101- Fixed bug that saved session permissions to local settings
102- Fixed MCP tools not being available to sub-agents
103- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
104- Fixed delay when navigating through typeahead suggestions with arrow keys
105- VSCode: Restored selection indicator in input footer showing current file or code selection status
106
107## 2.0.28
108
109- Plan mode: introduced new Plan subagent
110- Subagents: claude can now choose to resume subagents
111- Subagents: claude can dynamically choose the model used by its subagents
112- SDK: added --max-budget-usd flag
113- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
114- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
115- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
116- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
117- Various other bug fixes
118
119## 2.0.27
120
121- New UI for permission prompts
122- Added current branch filtering and search to session resume screen for easier navigation
123- Fixed directory @-mention causing "No assistant message found" error
124- VSCode Extension: Add config setting to include .gitignored files in file searches
125- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
126
127## 2.0.25
128
129- 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
130
131## 2.0.24
132
133- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
134- Claude Code Web: Support for Web -> CLI teleport
135- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
136- Bedrock: Display awsAuthRefresh output when auth is required
137
138## 2.0.22
139
140- Fixed content layout shift when scrolling through slash commands
141- IDE: Add toggle to enable/disable thinking.
142- Fix bug causing duplicate permission prompts with parallel tool calls
143- Add support for enterprise managed MCP allowlist and denylist
144
145## 2.0.21
146
147- Support MCP `structuredContent` field in tool responses
148- Added an interactive question tool
149- Claude will now ask you questions more often in plan mode
150- Added Haiku 4.5 as a model option for Pro users
151- Fixed an issue where queued commands don't have access to previous messages' output
152
153## 2.0.20
154
155- Added support for Claude Skills
156
157## 2.0.19
158
159- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
160- Fixed a bug where Haiku was unnecessarily called in print mode
161
162## 2.0.17
163
164- Added Haiku 4.5 to model selector!
165- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
166- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
167- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
168- OTEL: support HTTP_PROXY and HTTPS_PROXY
169- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
170
171## 2.0.15
172
173- Fixed bug with resuming where previously created files needed to be read again before writing
174- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
175
176## 2.0.14
177
178- Fix @-mentioning MCP servers to toggle them on/off
179- Improve permission checks for bash with inline env vars
180- Fix ultrathink + thinking toggle
181- Reduce unnecessary logins
182- Document --system-prompt
183- Several improvements to rendering
184- Plugins UI polish
185
186## 2.0.13
187
188- Fixed `/plugin` not working on native build
189
190## 2.0.12
191
192- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
193- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
194- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
195- `/plugin validate` command for validating plugin structure and configuration
196- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
197- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
198- Comprehensive error messages and diagnostics via `/doctor` command
199- Avoid flickering in `/model` selector
200- Improvements to `/help`
201- Avoid mentioning hooks in `/resume` summaries
202- Changes to the "verbose" setting in `/config` now persist across sessions
203
204## 2.0.11
205
206- Reduced system prompt size by 1.4k tokens
207- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
208- Fixed Opus fallback rate limit errors appearing incorrectly
209- Fixed /add-dir command selecting wrong default tab
210
211## 2.0.10
212
213- Rewrote terminal renderer for buttery smooth UI
214- Enable/disable MCP servers by @mentioning, or in /mcp
215- Added tab completion for shell commands in bash mode
216- PreToolUse hooks can now modify tool inputs
217- Press Ctrl-G to edit your prompt in your system's configured text editor
218- Fixes for bash permission checks with environment variables in the command
219
220## 2.0.9
221
222- Fix regression where bash backgrounding stopped working
223
224## 2.0.8
225
226- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
227- IDE: Add drag-and-drop support for files and folders in chat
228- /context: Fix counting for thinking blocks
229- Improve message rendering for users with light themes on dark terminals
230- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
231
232## 2.0.5
233
234- IDE: Fix IME unintended message submission with Enter and Tab
235- IDE: Add "Open in Terminal" link in login screen
236- Fix unhandled OAuth expiration 401 API errors
237- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
238
239## 2.0.1
240
241- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
242- Various bug fixes and presentation improvements
243
244## 2.0.0
245
246- New native VS Code extension
247- Fresh coat of paint throughout the whole app
248- /rewind a conversation to undo code changes
249- /usage command to see plan limits
250- Tab to toggle thinking (sticky across sessions)
251- Ctrl-R to search history
252- Unshipped claude config command
253- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
254- SDK: The Claude Code SDK is now the Claude Agent SDK
255- Add subagents dynamically with `--agents` flag
256
257## 1.0.126
258
259- Enable /context command for Bedrock and Vertex
260- Add mTLS support for HTTP-based OpenTelemetry exporters
261
262## 1.0.124
263
264- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
265- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
266- No longer inform Claude of the list of allowed tools when permission is denied
267- Fixed security vulnerability in Bash tool permission checks
268- Improved VSCode extension performance for large files
269
270## 1.0.123
271
272- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
273- Fixed thinking mode triggering on negation phrases like "don't think"
274- Fixed rendering performance degradation during token streaming
275- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
276- Enhanced BashTool environment snapshot logging
277- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
278- Migrated --debug logging to a file, to enable easy tailing & filtering
279
280## 1.0.120
281
282- Fix input lag during typing, especially noticeable with large prompts
283- Improved VSCode extension command registry and sessions dialog user experience
284- Enhanced sessions dialog responsiveness and visual feedback
285- Fixed IDE compatibility issue by removing worktree support check
286- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
287
288## 1.0.119
289
290- Fix Windows issue where process visually freezes on entering interactive mode
291- Support dynamic headers for MCP servers via headersHelper configuration
292- Fix thinking mode not working in headless sessions
293- Fix slash commands now properly update allowed tools instead of replacing them
294
295## 1.0.117
296
297- Add Ctrl-R history search to recall previous commands like bash/zsh
298- Fix input lag while typing, especially on Windows
299- Add sed command to auto-allowed commands in acceptEdits mode
300- Fix Windows PATH comparison to be case-insensitive for drive letters
301- Add permissions management hint to /add-dir output
302
303## 1.0.115
304
305- Improve thinking mode display with enhanced visual effects
306- Type /t to temporarily disable thinking mode in your prompt
307- Improve path validation for glob and grep tools
308- Show condensed output for post-tool hooks to reduce visual clutter
309- Fix visual feedback when loading state completes
310- Improve UI consistency for permission request dialogs
311
312## 1.0.113
313
314- Deprecated piped input in interactive mode
315- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
316
317## 1.0.112
318
319- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
320- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
321- Hooks: Added systemMessage support for SessionEnd hooks
322- Added `spinnerTipsEnabled` setting to disable spinner tips
323- IDE: Various improvements and bug fixes
324
325## 1.0.111
326
327- /model now validates provided model names
328- Fixed Bash tool crashes caused by malformed shell syntax parsing
329
330## 1.0.110
331
332- /terminal-setup command now supports WezTerm
333- MCP: OAuth tokens now proactively refresh before expiration
334- Fixed reliability issues with background Bash processes
335
336## 1.0.109
337
338- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
339
340## 1.0.106
341
342- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
343
344## 1.0.97
345
346- Settings: /doctor now validates permission rule syntax and suggests corrections
347
348## 1.0.94
349
350- Vertex: add support for global endpoints for supported models
351- /memory command now allows direct editing of all imported memory files
352- SDK: Add custom tools as callbacks
353- Added /todos command to list current todo items
354
355## 1.0.93
356
357- Windows: Add alt + v shortcut for pasting images from clipboard
358- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
359
360## 1.0.90
361
362- Settings file changes take effect immediately - no restart required
363
364## 1.0.88
365
366- Fixed issue causing "OAuth authentication is currently not supported"
367- Status line input now includes `exceeds_200k_tokens`
368- Fixed incorrect usage tracking in /cost.
369- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
370- Bedrock: Updated default Sonnet model to Sonnet 4
371
372## 1.0.86
373
374- Added /context to help users self-serve debug context issues
375- SDK: Added UUID support for all SDK messages
376- SDK: Added `--replay-user-messages` to replay user messages back to stdout
377
378## 1.0.85
379
380- Status line input now includes session cost info
381- Hooks: Introduced SessionEnd hook
382
383## 1.0.84
384
385- Fix tool_use/tool_result id mismatch error when network is unstable
386- Fix Claude sometimes ignoring real-time steering when wrapping up a task
387- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
388- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
389
390## 1.0.83
391
392- @-mention: Support files with spaces in path
393- New shimmering spinner
394
395## 1.0.82
396
397- SDK: Add request cancellation support
398- SDK: New additionalDirectories option to search custom paths, improved slash command processing
399- Settings: Validation prevents invalid fields in .claude/settings.json files
400- MCP: Improve tool name consistency
401- Bash: Fix crash when Claude tries to automatically read large files
402
403## 1.0.81
404
405- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
406- Agents: Fix custom agent loading when agent files are unparsable
407
408## 1.0.80
409
410- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
411
412## 1.0.77
413
414- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
415- SDK: Add session support and permission denial tracking
416- Fix token limit errors in conversation summarization
417- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
418
419## 1.0.73
420
421- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
422- MCP: Press Esc to cancel OAuth authentication flows
423- Bash: Improved command validation and reduced false security warnings
424- UI: Enhanced spinner animations and status line visual hierarchy
425- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
426
427## 1.0.72
428
429- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
430
431## 1.0.71
432
433- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
434- Customizable status line: add your terminal prompt to Claude Code with /statusline
435
436## 1.0.70
437
438- Performance: Optimized message rendering for better performance with large contexts
439- Windows: Fixed native file search, ripgrep, and subagent functionality
440- Added support for @-mentions in slash command arguments
441
442## 1.0.69
443
444- Upgraded Opus to version 4.1
445
446## 1.0.68
447
448- Fix incorrect model names being used for certain commands like `/pr-comments`
449- 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.
450- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
451- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
452- SDK: Added canUseTool callback support for tool confirmation
453- Added `disableAllHooks` setting
454- Improved file suggestions performance in large repos
455
456## 1.0.65
457
458- IDE: Fixed connection stability issues and error handling for diagnostics
459- Windows: Fixed shell environment setup for users without .bashrc files
460
461## 1.0.64
462
463- Agents: Added model customization support - you can now specify which model an agent should use
464- Agents: Fixed unintended access to the recursive agent tool
465- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
466- SDK: Fixed user input tracking across multi-turn conversations
467- Added hidden files to file search and @-mention suggestions
468
469## 1.0.63
470
471- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
472
473## 1.0.62
474
475- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
476- Hooks: Added SessionStart hook for new session initialization
477- /add-dir command now supports typeahead for directory paths
478- Improved network connectivity check reliability
479
480## 1.0.61
481
482- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
483- Settings: Added `--settings` flag to load settings from a JSON file
484- Settings: Fixed resolution of settings files paths that are symlinks
485- OTEL: Fixed reporting of wrong organization after authentication changes
486- Slash commands: Fixed permissions checking for allowed-tools with Bash
487- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
488- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
489- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
490
491## 1.0.60
492
493- You can now create custom subagents for specialized tasks! Run /agents to get started
494
495## 1.0.59
496
497- SDK: Added tool confirmation support with canUseTool callback
498- SDK: Allow specifying env for spawned process
499- Hooks: Exposed PermissionDecision to hooks (including "ask")
500- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
501- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
502
503## 1.0.58
504
505- Added support for reading PDFs
506- MCP: Improved server health status display in 'claude mcp list'
507- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
508
509## 1.0.57
510
511- Added support for specifying a model in slash commands
512- Improved permission messages to help Claude understand allowed tools
513- Fix: Remove trailing newlines from bash output in terminal wrapping
514
515## 1.0.56
516
517- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
518- Fixes for WSL IDE detection
519- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
520
521## 1.0.55
522
523- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
524- Windows: fixed Ctrl+Z crash
525- SDK: Added ability to capture error logging
526- Add --system-prompt-file option to override system prompt in print mode
527
528## 1.0.54
529
530- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
531- Custom slash commands: Added argument-hint to frontmatter
532- Windows: OAuth uses port 45454 and properly constructs browser URL
533- Windows: mode switching now uses alt + m, and plan mode renders properly
534- Shell: Switch to in-memory shell snapshot to fix file-related errors
535
536## 1.0.53
537
538- Updated @-mention file truncation from 100 lines to 2000 lines
539- 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).
540
541## 1.0.52
542
543- Added support for MCP server instructions
544
545## 1.0.51
546
547- Added support for native Windows (requires Git for Windows)
548- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
549- Settings: /doctor can now help you identify and fix invalid setting files
550- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
551- Increased auto-compact warning threshold from 60% to 80%
552- Fixed an issue with handling user directories with spaces for shell snapshots
553- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
554- Custom slash commands: Fixed user-level commands in subdirectories
555- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
556
557## 1.0.48
558
559- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
560- Added progress messages to Bash tool based on the last 5 lines of command output
561- Added expanding variables support for MCP server configuration
562- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
563- Improved IDE extension path handling when Claude Code runs in WSL
564- Hooks: Added a PreCompact hook
565- Vim mode: Added c, f/F, t/T
566
567## 1.0.45
568
569- Redesigned Search (Grep) tool with new tool input parameters and features
570- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
571- Fixed config file corruption issue by enforcing atomic writes
572- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
573- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
574- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
575
576## 1.0.44
577
578- New /export command lets you quickly export a conversation for sharing
579- MCP: resource_link tool results are now supported
580- MCP: tool annotations and tool titles now display in /mcp view
581- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
582
583## 1.0.43
584
585- Fixed a bug where the theme selector was saving excessively
586- Hooks: Added EPIPE system error handling
587
588## 1.0.42
589
590- Added tilde (`~`) expansion support to `/add-dir` command
591
592## 1.0.41
593
594- Hooks: Split Stop hook triggering into Stop and SubagentStop
595- Hooks: Enabled optional timeout configuration for each command
596- Hooks: Added "hook_event_name" to hook input
597- Fixed a bug where MCP tools would display twice in tool list
598- New tool parameters JSON for Bash tool in `tool_decision` event
599
600## 1.0.40
601
602- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
603
604## 1.0.39
605
606- New Active Time metric in OpenTelemetry logging
607
608## 1.0.38
609
610- 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
611
612## 1.0.37
613
614- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
615
616## 1.0.36
617
618- Web search now takes today's date into context
619- Fixed a bug where stdio MCP servers were not terminating properly on exit
620
621## 1.0.35
622
623- Added support for MCP OAuth Authorization Server discovery
624
625## 1.0.34
626
627- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
628
629## 1.0.33
630
631- Improved logging functionality with session ID support
632- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
633- Improvements to plan mode
634
635## 1.0.32
636
637- Updated loopback config for litellm
638- Added forceLoginMethod setting to bypass login selection screen
639
640## 1.0.31
641
642- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
643
644## 1.0.30
645
646- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
647- Improved file path autocomplete with filename matching
648- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
649- Enhanced jq regex support for complex filters with pipes and select
650
651## 1.0.29
652
653- Improved CJK character support in cursor navigation and rendering
654
655## 1.0.28
656
657- Slash commands: Fix selector display during history navigation
658- Resizes images before upload to prevent API size limit errors
659- Added XDG_CONFIG_HOME support to configuration directory
660- Performance optimizations for memory usage
661- New attributes (terminal.type, language) in OpenTelemetry logging
662
663## 1.0.27
664
665- Streamable HTTP MCP servers are now supported
666- Remote MCP servers (SSE and HTTP) now support OAuth
667- MCP resources can now be @-mentioned
668- /resume slash command to switch conversations within Claude Code
669
670## 1.0.25
671
672- Slash commands: moved "project" and "user" prefixes to descriptions
673- Slash commands: improved reliability for command discovery
674- Improved support for Ghostty
675- Improved web search reliability
676
677## 1.0.24
678
679- Improved /mcp output
680- Fixed a bug where settings arrays got overwritten instead of merged
681
682## 1.0.23
683
684- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
685- Released Python SDK: pip install claude-code-sdk to get started
686
687## 1.0.22
688
689- SDK: Renamed `total_cost` to `total_cost_usd`
690
691## 1.0.21
692
693- Improved editing of files with tab-based indentation
694- Fix for tool_use without matching tool_result errors
695- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
696
697## 1.0.18
698
699- Added --add-dir CLI argument for specifying additional working directories
700- Added streaming input support without require -p flag
701- Improved startup performance and session storage performance
702- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
703- Added detailed MCP server tools display (/mcp)
704- MCP authentication and permission improvements
705- Added auto-reconnection for MCP SSE connections on disconnect
706- Fixed issue where pasted content was lost when dialogs appeared
707
708## 1.0.17
709
710- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
711- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
712- MCP server list UI improvements
713- Update Claude Code process title to display "claude" instead of "node"
714
715## 1.0.11
716
717- Claude Code can now also be used with a Claude Pro subscription
718- Added /upgrade for smoother switching to Claude Max plans
719- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
720- Improved shell configuration error handling
721- Improved todo list handling during compaction
722
723## 1.0.10
724
725- Added markdown table support
726- Improved streaming performance
727
728## 1.0.8
729
730- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
731- Increased default otel interval from 1s -> 5s
732- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
733- Fixed a regression where search tools unnecessarily asked for permissions
734- Added support for triggering thinking non-English languages
735- Improved compacting UI
736
737## 1.0.7
738
739- Renamed /allowed-tools -> /permissions
740- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
741- Deprecated claude config commands in favor of editing settings.json
742- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
743- Improved error handling for /install-github-app
744- Bugfixes, UI polish, and tool reliability improvements
745
746## 1.0.6
747
748- Improved edit reliability for tab-indented files
749- Respect CLAUDE_CONFIG_DIR everywhere
750- Reduced unnecessary tool permission prompts
751- Added support for symlinks in @file typeahead
752- Bugfixes, UI polish, and tool reliability improvements
753
754## 1.0.4
755
756- Fixed a bug where MCP tool errors weren't being parsed correctly
757
758## 1.0.1
759
760- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
761- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
762- Updated documentation links and OAuth process descriptions
763
764## 1.0.0
765
766- Claude Code is now generally available
767- Introducing Sonnet 4 and Opus 4 models
768
769## 0.2.125
770
771- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
772- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
773
774## 0.2.117
775
776- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
777- Introduced settings.cleanupPeriodDays
778- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
779- Introduced --debug mode
780
781## 0.2.108
782
783- You can now send messages to Claude while it works to steer Claude in real-time
784- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
785- Fixed a bug where thinking was not working in -p mode
786- Fixed a regression in /cost reporting
787- Deprecated MCP wizard interface in favor of other MCP commands
788- Lots of other bugfixes and improvements
789
790## 0.2.107
791
792- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
793
794## 0.2.106
795
796- MCP SSE server configs can now specify custom headers
797- Fixed a bug where MCP permission prompt didn't always show correctly
798
799## 0.2.105
800
801- Claude can now search the web
802- Moved system & account status to /status
803- Added word movement keybindings for Vim
804- Improved latency for startup, todo tool, and file edits
805
806## 0.2.102
807
808- Improved thinking triggering reliability
809- Improved @mention reliability for images and folders
810- You can now paste multiple large chunks into one prompt
811
812## 0.2.100
813
814- Fixed a crash caused by a stack overflow error
815- Made db storage optional; missing db support disables --continue and --resume
816
817## 0.2.98
818
819- Fixed an issue where auto-compact was running twice
820
821## 0.2.96
822
823- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
824
825## 0.2.93
826
827- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
828- Claude now has access to a Todo list that helps it stay on track and be more organized
829
830## 0.2.82
831
832- Added support for --disallowedTools
833- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
834
835## 0.2.75
836
837- Hit Enter to queue up additional messages while Claude is working
838- Drag in or copy/paste image files directly into the prompt
839- @-mention files to directly add them to context
840- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
841- Improved performance for filename auto-complete
842
843## 0.2.74
844
845- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
846- Task tool can now perform writes and run bash commands
847
848## 0.2.72
849
850- Updated spinner to indicate tokens loaded and tool usage
851
852## 0.2.70
853
854- Network commands like curl are now available for Claude to use
855- Claude can now run multiple web queries in parallel
856- Pressing ESC once immediately interrupts Claude in Auto-accept mode
857
858## 0.2.69
859
860- Fixed UI glitches with improved Select component behavior
861- Enhanced terminal output display with better text truncation logic
862
863## 0.2.67
864
865- Shared project permission rules can be saved in .claude/settings.json
866
867## 0.2.66
868
869- Print mode (-p) now supports streaming output via --output-format=stream-json
870- Fixed issue where pasting could trigger memory or bash mode unexpectedly
871
872## 0.2.63
873
874- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
875
876## 0.2.61
877
878- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
879- Enhanced image detection for more reliable clipboard paste functionality
880- Fixed an issue where ESC key could crash the conversation history selector
881
882## 0.2.59
883
884- Copy+paste images directly into your prompt
885- Improved progress indicators for bash and fetch tools
886- Bugfixes for non-interactive mode (-p)
887
888## 0.2.54
889
890- Quickly add to Memory by starting your message with '#'
891- Press ctrl+r to see full output for long tool results
892- Added support for MCP SSE transport
893
894## 0.2.53
895
896- New web fetch tool lets Claude view URLs that you paste in
897- Fixed a bug with JPEG detection
898
899## 0.2.50
900
901- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
902
903## 0.2.49
904
905- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
906
907## 0.2.47
908
909- Press Tab to auto-complete file and folder names
910- Press Shift + Tab to toggle auto-accept for file edits
911- Automatic conversation compaction for infinite conversation length (toggle with /config)
912
913## 0.2.44
914
915- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
916
917## 0.2.41
918
919- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
920- MCP server startup no longer blocks the app from starting up
921
922## 0.2.37
923
924- New /release-notes command lets you view release notes at any time
925- `claude config add/remove` commands now accept multiple values separated by commas or spaces
926
927## 0.2.36
928
929- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
930- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
931
932## 0.2.34
933
934- Vim bindings for text input - enable with /vim or /config
935
936## 0.2.32
937
938- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
939- Fix for some PersistentShell issues
940
941## 0.2.31
942
943- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
944- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
945
946## 0.2.30
947
948- Added ANSI color theme for better terminal compatibility
949- Fixed issue where slash command arguments weren't being sent properly
950- (Mac-only) API keys are now stored in macOS Keychain
951
952## 0.2.26
953
954- New /approved-tools command for managing tool permissions
955- Word-level diff display for improved code readability
956- Fuzzy matching for slash commands
957
958## 0.2.21
959
960- Fuzzy matching for /commands