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 4928f2c · on Dec 2, 2025 History
1,019 lines · 39.5 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.56
4
5- Added setting to enable/disable terminal progress bar (OSC 9;4)
6- VSCode Extension: Added support for VS Code's secondary sidebar (VS Code 1.97+), allowing Claude Code to be displayed in the right sidebar while keeping the file explorer on the left. Requires setting sidebar as Preferred Location in the config.
7
8## 2.0.55
9
10- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
11- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
12- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
13- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
14
15## 2.0.54
16
17- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
18- Fix issue with excessive iTerm notifications
19
20## 2.0.52
21
22- Fixed duplicate message display when starting Claude with a command line argument
23- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
24- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
25- Permit some uses of `$!` in bash commands
26
27## 2.0.51
28
29- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
30- Introducing Claude Code for Desktop: https://claude.com/download
31- To give you room to try out our new model, we've updated usage limits for Claude Code users. See the Claude Opus 4.5 blog for full details
32- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
33- Plan Mode now builds more precise plans and executes more thoroughly
34- Usage limit notifications now easier to understand
35- Switched `/usage` back to "% used"
36- Fixed handling of thinking errors
37- Fixed performance regression
38
39## 2.0.50
40
41- Fixed bug preventing calling MCP tools that have nested references in their input schemas
42- Silenced a noisy but harmless error during upgrades
43- Improved ultrathink text display
44- Improved clarity of 5-hour session limit warning message
45
46## 2.0.49
47
48- Added readline-style ctrl-y for pasting deleted text
49- Improved clarity of usage limit warning message
50- Fixed handling of subagent permissions
51
52## 2.0.47
53
54- Improved error messages and validation for `claude --teleport`
55- Improved error handling in `/usage`
56- Fixed race condition with history entry not getting logged at exit
57- Fixed Vertex AI configuration not being applied from `settings.json`
58
59## 2.0.46
60
61- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
62
63## 2.0.45
64
65- Add support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
66- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
67- Send background tasks to Claude Code on the web by starting a message with `&`
68
69## 2.0.43
70
71- Added `permissionMode` field for custom agents
72- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
73- Added skills frontmatter field to declare skills to auto-load for subagents
74- Added the `SubagentStart` hook event
75- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
76- Fixed duplicate rendering of some messages in the UI
77- Fixed some visual flickers
78- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
79
80## 2.0.42
81
82- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
83
84## 2.0.41
85
86- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
87- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
88- Fixed incorrect labeling of user settings vs project settings in command descriptions
89- Fixed crash when plugin command hooks timeout during execution
90- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
91- Fixed broken security documentation links in trust dialogs and onboarding
92- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
93- ctrl-r history search landing on a slash command no longer cancels the search
94- SDK: Support custom timeouts for hooks
95- Allow more safe git commands to run without approval
96- Plugins: Added support for sharing and installing output styles
97- Teleporting a session from web will automatically set the upstream branch
98
99## 2.0.37
100
101- Fixed how idleness is computed for notifications
102- Hooks: Added matcher values for Notification hook events
103- Output Styles: Added `keep-coding-instructions` option to frontmatter
104
105## 2.0.36
106
107- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
108- Fixed queued messages being incorrectly executed as bash commands
109- Fixed input being lost when typing while a queued message is processed
110
111## 2.0.35
112
113- Improve fuzzy search results when searching commands
114- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
115- 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
116- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
117- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
118
119## 2.0.34
120
121- VSCode Extension: Added setting to configure the initial permission mode for new conversations
122- Improved file path suggestion performance with native Rust-based fuzzy finder
123- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
124- Fixed memory crash when reading or writing large files (especially base64-encoded images)
125
126## 2.0.33
127
128- Native binary installs now launch quicker.
129- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
130- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
131
132## 2.0.32
133
134- Un-deprecate output styles based on community feedback
135- Added `companyAnnouncements` setting for displaying announcements on startup
136- Fixed hook progress messages not updating correctly during PostToolUse hook execution
137
138## 2.0.31
139
140- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
141- Vertex: add support for Web Search on supported models
142- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
143- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
144- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
145- Fixed plugin uninstall not removing plugins
146
147## 2.0.30
148
149- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
150- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
151- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
152- Added prompt-based stop hooks
153- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
154- Enabled SSE MCP servers on native build
155- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
156- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
157- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
158- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
159- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
160- Fixed bug that saved session permissions to local settings
161- Fixed MCP tools not being available to sub-agents
162- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
163- Fixed delay when navigating through typeahead suggestions with arrow keys
164- VSCode: Restored selection indicator in input footer showing current file or code selection status
165
166## 2.0.28
167
168- Plan mode: introduced new Plan subagent
169- Subagents: claude can now choose to resume subagents
170- Subagents: claude can dynamically choose the model used by its subagents
171- SDK: added --max-budget-usd flag
172- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
173- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
174- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
175- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
176- Various other bug fixes
177
178## 2.0.27
179
180- New UI for permission prompts
181- Added current branch filtering and search to session resume screen for easier navigation
182- Fixed directory @-mention causing "No assistant message found" error
183- VSCode Extension: Add config setting to include .gitignored files in file searches
184- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
185
186## 2.0.25
187
188- 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
189
190## 2.0.24
191
192- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
193- Claude Code Web: Support for Web -> CLI teleport
194- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
195- Bedrock: Display awsAuthRefresh output when auth is required
196
197## 2.0.22
198
199- Fixed content layout shift when scrolling through slash commands
200- IDE: Add toggle to enable/disable thinking.
201- Fix bug causing duplicate permission prompts with parallel tool calls
202- Add support for enterprise managed MCP allowlist and denylist
203
204## 2.0.21
205
206- Support MCP `structuredContent` field in tool responses
207- Added an interactive question tool
208- Claude will now ask you questions more often in plan mode
209- Added Haiku 4.5 as a model option for Pro users
210- Fixed an issue where queued commands don't have access to previous messages' output
211
212## 2.0.20
213
214- Added support for Claude Skills
215
216## 2.0.19
217
218- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
219- Fixed a bug where Haiku was unnecessarily called in print mode
220
221## 2.0.17
222
223- Added Haiku 4.5 to model selector!
224- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
225- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
226- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
227- OTEL: support HTTP_PROXY and HTTPS_PROXY
228- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
229
230## 2.0.15
231
232- Fixed bug with resuming where previously created files needed to be read again before writing
233- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
234
235## 2.0.14
236
237- Fix @-mentioning MCP servers to toggle them on/off
238- Improve permission checks for bash with inline env vars
239- Fix ultrathink + thinking toggle
240- Reduce unnecessary logins
241- Document --system-prompt
242- Several improvements to rendering
243- Plugins UI polish
244
245## 2.0.13
246
247- Fixed `/plugin` not working on native build
248
249## 2.0.12
250
251- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
252- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
253- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
254- `/plugin validate` command for validating plugin structure and configuration
255- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
256- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
257- Comprehensive error messages and diagnostics via `/doctor` command
258- Avoid flickering in `/model` selector
259- Improvements to `/help`
260- Avoid mentioning hooks in `/resume` summaries
261- Changes to the "verbose" setting in `/config` now persist across sessions
262
263## 2.0.11
264
265- Reduced system prompt size by 1.4k tokens
266- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
267- Fixed Opus fallback rate limit errors appearing incorrectly
268- Fixed /add-dir command selecting wrong default tab
269
270## 2.0.10
271
272- Rewrote terminal renderer for buttery smooth UI
273- Enable/disable MCP servers by @mentioning, or in /mcp
274- Added tab completion for shell commands in bash mode
275- PreToolUse hooks can now modify tool inputs
276- Press Ctrl-G to edit your prompt in your system's configured text editor
277- Fixes for bash permission checks with environment variables in the command
278
279## 2.0.9
280
281- Fix regression where bash backgrounding stopped working
282
283## 2.0.8
284
285- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
286- IDE: Add drag-and-drop support for files and folders in chat
287- /context: Fix counting for thinking blocks
288- Improve message rendering for users with light themes on dark terminals
289- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
290
291## 2.0.5
292
293- IDE: Fix IME unintended message submission with Enter and Tab
294- IDE: Add "Open in Terminal" link in login screen
295- Fix unhandled OAuth expiration 401 API errors
296- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
297
298## 2.0.1
299
300- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
301- Various bug fixes and presentation improvements
302
303## 2.0.0
304
305- New native VS Code extension
306- Fresh coat of paint throughout the whole app
307- /rewind a conversation to undo code changes
308- /usage command to see plan limits
309- Tab to toggle thinking (sticky across sessions)
310- Ctrl-R to search history
311- Unshipped claude config command
312- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
313- SDK: The Claude Code SDK is now the Claude Agent SDK
314- Add subagents dynamically with `--agents` flag
315
316## 1.0.126
317
318- Enable /context command for Bedrock and Vertex
319- Add mTLS support for HTTP-based OpenTelemetry exporters
320
321## 1.0.124
322
323- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
324- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
325- No longer inform Claude of the list of allowed tools when permission is denied
326- Fixed security vulnerability in Bash tool permission checks
327- Improved VSCode extension performance for large files
328
329## 1.0.123
330
331- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
332- Fixed thinking mode triggering on negation phrases like "don't think"
333- Fixed rendering performance degradation during token streaming
334- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
335- Enhanced BashTool environment snapshot logging
336- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
337- Migrated --debug logging to a file, to enable easy tailing & filtering
338
339## 1.0.120
340
341- Fix input lag during typing, especially noticeable with large prompts
342- Improved VSCode extension command registry and sessions dialog user experience
343- Enhanced sessions dialog responsiveness and visual feedback
344- Fixed IDE compatibility issue by removing worktree support check
345- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
346
347## 1.0.119
348
349- Fix Windows issue where process visually freezes on entering interactive mode
350- Support dynamic headers for MCP servers via headersHelper configuration
351- Fix thinking mode not working in headless sessions
352- Fix slash commands now properly update allowed tools instead of replacing them
353
354## 1.0.117
355
356- Add Ctrl-R history search to recall previous commands like bash/zsh
357- Fix input lag while typing, especially on Windows
358- Add sed command to auto-allowed commands in acceptEdits mode
359- Fix Windows PATH comparison to be case-insensitive for drive letters
360- Add permissions management hint to /add-dir output
361
362## 1.0.115
363
364- Improve thinking mode display with enhanced visual effects
365- Type /t to temporarily disable thinking mode in your prompt
366- Improve path validation for glob and grep tools
367- Show condensed output for post-tool hooks to reduce visual clutter
368- Fix visual feedback when loading state completes
369- Improve UI consistency for permission request dialogs
370
371## 1.0.113
372
373- Deprecated piped input in interactive mode
374- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
375
376## 1.0.112
377
378- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
379- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
380- Hooks: Added systemMessage support for SessionEnd hooks
381- Added `spinnerTipsEnabled` setting to disable spinner tips
382- IDE: Various improvements and bug fixes
383
384## 1.0.111
385
386- /model now validates provided model names
387- Fixed Bash tool crashes caused by malformed shell syntax parsing
388
389## 1.0.110
390
391- /terminal-setup command now supports WezTerm
392- MCP: OAuth tokens now proactively refresh before expiration
393- Fixed reliability issues with background Bash processes
394
395## 1.0.109
396
397- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
398
399## 1.0.106
400
401- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
402
403## 1.0.97
404
405- Settings: /doctor now validates permission rule syntax and suggests corrections
406
407## 1.0.94
408
409- Vertex: add support for global endpoints for supported models
410- /memory command now allows direct editing of all imported memory files
411- SDK: Add custom tools as callbacks
412- Added /todos command to list current todo items
413
414## 1.0.93
415
416- Windows: Add alt + v shortcut for pasting images from clipboard
417- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
418
419## 1.0.90
420
421- Settings file changes take effect immediately - no restart required
422
423## 1.0.88
424
425- Fixed issue causing "OAuth authentication is currently not supported"
426- Status line input now includes `exceeds_200k_tokens`
427- Fixed incorrect usage tracking in /cost.
428- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
429- Bedrock: Updated default Sonnet model to Sonnet 4
430
431## 1.0.86
432
433- Added /context to help users self-serve debug context issues
434- SDK: Added UUID support for all SDK messages
435- SDK: Added `--replay-user-messages` to replay user messages back to stdout
436
437## 1.0.85
438
439- Status line input now includes session cost info
440- Hooks: Introduced SessionEnd hook
441
442## 1.0.84
443
444- Fix tool_use/tool_result id mismatch error when network is unstable
445- Fix Claude sometimes ignoring real-time steering when wrapping up a task
446- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
447- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
448
449## 1.0.83
450
451- @-mention: Support files with spaces in path
452- New shimmering spinner
453
454## 1.0.82
455
456- SDK: Add request cancellation support
457- SDK: New additionalDirectories option to search custom paths, improved slash command processing
458- Settings: Validation prevents invalid fields in .claude/settings.json files
459- MCP: Improve tool name consistency
460- Bash: Fix crash when Claude tries to automatically read large files
461
462## 1.0.81
463
464- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
465- Agents: Fix custom agent loading when agent files are unparsable
466
467## 1.0.80
468
469- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
470
471## 1.0.77
472
473- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
474- SDK: Add session support and permission denial tracking
475- Fix token limit errors in conversation summarization
476- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
477
478## 1.0.73
479
480- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
481- MCP: Press Esc to cancel OAuth authentication flows
482- Bash: Improved command validation and reduced false security warnings
483- UI: Enhanced spinner animations and status line visual hierarchy
484- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
485
486## 1.0.72
487
488- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
489
490## 1.0.71
491
492- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
493- Customizable status line: add your terminal prompt to Claude Code with /statusline
494
495## 1.0.70
496
497- Performance: Optimized message rendering for better performance with large contexts
498- Windows: Fixed native file search, ripgrep, and subagent functionality
499- Added support for @-mentions in slash command arguments
500
501## 1.0.69
502
503- Upgraded Opus to version 4.1
504
505## 1.0.68
506
507- Fix incorrect model names being used for certain commands like `/pr-comments`
508- 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.
509- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
510- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
511- SDK: Added canUseTool callback support for tool confirmation
512- Added `disableAllHooks` setting
513- Improved file suggestions performance in large repos
514
515## 1.0.65
516
517- IDE: Fixed connection stability issues and error handling for diagnostics
518- Windows: Fixed shell environment setup for users without .bashrc files
519
520## 1.0.64
521
522- Agents: Added model customization support - you can now specify which model an agent should use
523- Agents: Fixed unintended access to the recursive agent tool
524- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
525- SDK: Fixed user input tracking across multi-turn conversations
526- Added hidden files to file search and @-mention suggestions
527
528## 1.0.63
529
530- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
531
532## 1.0.62
533
534- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
535- Hooks: Added SessionStart hook for new session initialization
536- /add-dir command now supports typeahead for directory paths
537- Improved network connectivity check reliability
538
539## 1.0.61
540
541- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
542- Settings: Added `--settings` flag to load settings from a JSON file
543- Settings: Fixed resolution of settings files paths that are symlinks
544- OTEL: Fixed reporting of wrong organization after authentication changes
545- Slash commands: Fixed permissions checking for allowed-tools with Bash
546- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
547- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
548- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
549
550## 1.0.60
551
552- You can now create custom subagents for specialized tasks! Run /agents to get started
553
554## 1.0.59
555
556- SDK: Added tool confirmation support with canUseTool callback
557- SDK: Allow specifying env for spawned process
558- Hooks: Exposed PermissionDecision to hooks (including "ask")
559- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
560- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
561
562## 1.0.58
563
564- Added support for reading PDFs
565- MCP: Improved server health status display in 'claude mcp list'
566- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
567
568## 1.0.57
569
570- Added support for specifying a model in slash commands
571- Improved permission messages to help Claude understand allowed tools
572- Fix: Remove trailing newlines from bash output in terminal wrapping
573
574## 1.0.56
575
576- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
577- Fixes for WSL IDE detection
578- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
579
580## 1.0.55
581
582- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
583- Windows: fixed Ctrl+Z crash
584- SDK: Added ability to capture error logging
585- Add --system-prompt-file option to override system prompt in print mode
586
587## 1.0.54
588
589- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
590- Custom slash commands: Added argument-hint to frontmatter
591- Windows: OAuth uses port 45454 and properly constructs browser URL
592- Windows: mode switching now uses alt + m, and plan mode renders properly
593- Shell: Switch to in-memory shell snapshot to fix file-related errors
594
595## 1.0.53
596
597- Updated @-mention file truncation from 100 lines to 2000 lines
598- 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).
599
600## 1.0.52
601
602- Added support for MCP server instructions
603
604## 1.0.51
605
606- Added support for native Windows (requires Git for Windows)
607- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
608- Settings: /doctor can now help you identify and fix invalid setting files
609- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
610- Increased auto-compact warning threshold from 60% to 80%
611- Fixed an issue with handling user directories with spaces for shell snapshots
612- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
613- Custom slash commands: Fixed user-level commands in subdirectories
614- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
615
616## 1.0.48
617
618- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
619- Added progress messages to Bash tool based on the last 5 lines of command output
620- Added expanding variables support for MCP server configuration
621- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
622- Improved IDE extension path handling when Claude Code runs in WSL
623- Hooks: Added a PreCompact hook
624- Vim mode: Added c, f/F, t/T
625
626## 1.0.45
627
628- Redesigned Search (Grep) tool with new tool input parameters and features
629- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
630- Fixed config file corruption issue by enforcing atomic writes
631- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
632- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
633- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
634
635## 1.0.44
636
637- New /export command lets you quickly export a conversation for sharing
638- MCP: resource_link tool results are now supported
639- MCP: tool annotations and tool titles now display in /mcp view
640- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
641
642## 1.0.43
643
644- Fixed a bug where the theme selector was saving excessively
645- Hooks: Added EPIPE system error handling
646
647## 1.0.42
648
649- Added tilde (`~`) expansion support to `/add-dir` command
650
651## 1.0.41
652
653- Hooks: Split Stop hook triggering into Stop and SubagentStop
654- Hooks: Enabled optional timeout configuration for each command
655- Hooks: Added "hook_event_name" to hook input
656- Fixed a bug where MCP tools would display twice in tool list
657- New tool parameters JSON for Bash tool in `tool_decision` event
658
659## 1.0.40
660
661- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
662
663## 1.0.39
664
665- New Active Time metric in OpenTelemetry logging
666
667## 1.0.38
668
669- 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
670
671## 1.0.37
672
673- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
674
675## 1.0.36
676
677- Web search now takes today's date into context
678- Fixed a bug where stdio MCP servers were not terminating properly on exit
679
680## 1.0.35
681
682- Added support for MCP OAuth Authorization Server discovery
683
684## 1.0.34
685
686- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
687
688## 1.0.33
689
690- Improved logging functionality with session ID support
691- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
692- Improvements to plan mode
693
694## 1.0.32
695
696- Updated loopback config for litellm
697- Added forceLoginMethod setting to bypass login selection screen
698
699## 1.0.31
700
701- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
702
703## 1.0.30
704
705- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
706- Improved file path autocomplete with filename matching
707- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
708- Enhanced jq regex support for complex filters with pipes and select
709
710## 1.0.29
711
712- Improved CJK character support in cursor navigation and rendering
713
714## 1.0.28
715
716- Slash commands: Fix selector display during history navigation
717- Resizes images before upload to prevent API size limit errors
718- Added XDG_CONFIG_HOME support to configuration directory
719- Performance optimizations for memory usage
720- New attributes (terminal.type, language) in OpenTelemetry logging
721
722## 1.0.27
723
724- Streamable HTTP MCP servers are now supported
725- Remote MCP servers (SSE and HTTP) now support OAuth
726- MCP resources can now be @-mentioned
727- /resume slash command to switch conversations within Claude Code
728
729## 1.0.25
730
731- Slash commands: moved "project" and "user" prefixes to descriptions
732- Slash commands: improved reliability for command discovery
733- Improved support for Ghostty
734- Improved web search reliability
735
736## 1.0.24
737
738- Improved /mcp output
739- Fixed a bug where settings arrays got overwritten instead of merged
740
741## 1.0.23
742
743- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
744- Released Python SDK: pip install claude-code-sdk to get started
745
746## 1.0.22
747
748- SDK: Renamed `total_cost` to `total_cost_usd`
749
750## 1.0.21
751
752- Improved editing of files with tab-based indentation
753- Fix for tool_use without matching tool_result errors
754- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
755
756## 1.0.18
757
758- Added --add-dir CLI argument for specifying additional working directories
759- Added streaming input support without require -p flag
760- Improved startup performance and session storage performance
761- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
762- Added detailed MCP server tools display (/mcp)
763- MCP authentication and permission improvements
764- Added auto-reconnection for MCP SSE connections on disconnect
765- Fixed issue where pasted content was lost when dialogs appeared
766
767## 1.0.17
768
769- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
770- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
771- MCP server list UI improvements
772- Update Claude Code process title to display "claude" instead of "node"
773
774## 1.0.11
775
776- Claude Code can now also be used with a Claude Pro subscription
777- Added /upgrade for smoother switching to Claude Max plans
778- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
779- Improved shell configuration error handling
780- Improved todo list handling during compaction
781
782## 1.0.10
783
784- Added markdown table support
785- Improved streaming performance
786
787## 1.0.8
788
789- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
790- Increased default otel interval from 1s -> 5s
791- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
792- Fixed a regression where search tools unnecessarily asked for permissions
793- Added support for triggering thinking non-English languages
794- Improved compacting UI
795
796## 1.0.7
797
798- Renamed /allowed-tools -> /permissions
799- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
800- Deprecated claude config commands in favor of editing settings.json
801- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
802- Improved error handling for /install-github-app
803- Bugfixes, UI polish, and tool reliability improvements
804
805## 1.0.6
806
807- Improved edit reliability for tab-indented files
808- Respect CLAUDE_CONFIG_DIR everywhere
809- Reduced unnecessary tool permission prompts
810- Added support for symlinks in @file typeahead
811- Bugfixes, UI polish, and tool reliability improvements
812
813## 1.0.4
814
815- Fixed a bug where MCP tool errors weren't being parsed correctly
816
817## 1.0.1
818
819- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
820- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
821- Updated documentation links and OAuth process descriptions
822
823## 1.0.0
824
825- Claude Code is now generally available
826- Introducing Sonnet 4 and Opus 4 models
827
828## 0.2.125
829
830- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
831- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
832
833## 0.2.117
834
835- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
836- Introduced settings.cleanupPeriodDays
837- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
838- Introduced --debug mode
839
840## 0.2.108
841
842- You can now send messages to Claude while it works to steer Claude in real-time
843- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
844- Fixed a bug where thinking was not working in -p mode
845- Fixed a regression in /cost reporting
846- Deprecated MCP wizard interface in favor of other MCP commands
847- Lots of other bugfixes and improvements
848
849## 0.2.107
850
851- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
852
853## 0.2.106
854
855- MCP SSE server configs can now specify custom headers
856- Fixed a bug where MCP permission prompt didn't always show correctly
857
858## 0.2.105
859
860- Claude can now search the web
861- Moved system & account status to /status
862- Added word movement keybindings for Vim
863- Improved latency for startup, todo tool, and file edits
864
865## 0.2.102
866
867- Improved thinking triggering reliability
868- Improved @mention reliability for images and folders
869- You can now paste multiple large chunks into one prompt
870
871## 0.2.100
872
873- Fixed a crash caused by a stack overflow error
874- Made db storage optional; missing db support disables --continue and --resume
875
876## 0.2.98
877
878- Fixed an issue where auto-compact was running twice
879
880## 0.2.96
881
882- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
883
884## 0.2.93
885
886- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
887- Claude now has access to a Todo list that helps it stay on track and be more organized
888
889## 0.2.82
890
891- Added support for --disallowedTools
892- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
893
894## 0.2.75
895
896- Hit Enter to queue up additional messages while Claude is working
897- Drag in or copy/paste image files directly into the prompt
898- @-mention files to directly add them to context
899- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
900- Improved performance for filename auto-complete
901
902## 0.2.74
903
904- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
905- Task tool can now perform writes and run bash commands
906
907## 0.2.72
908
909- Updated spinner to indicate tokens loaded and tool usage
910
911## 0.2.70
912
913- Network commands like curl are now available for Claude to use
914- Claude can now run multiple web queries in parallel
915- Pressing ESC once immediately interrupts Claude in Auto-accept mode
916
917## 0.2.69
918
919- Fixed UI glitches with improved Select component behavior
920- Enhanced terminal output display with better text truncation logic
921
922## 0.2.67
923
924- Shared project permission rules can be saved in .claude/settings.json
925
926## 0.2.66
927
928- Print mode (-p) now supports streaming output via --output-format=stream-json
929- Fixed issue where pasting could trigger memory or bash mode unexpectedly
930
931## 0.2.63
932
933- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
934
935## 0.2.61
936
937- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
938- Enhanced image detection for more reliable clipboard paste functionality
939- Fixed an issue where ESC key could crash the conversation history selector
940
941## 0.2.59
942
943- Copy+paste images directly into your prompt
944- Improved progress indicators for bash and fetch tools
945- Bugfixes for non-interactive mode (-p)
946
947## 0.2.54
948
949- Quickly add to Memory by starting your message with '#'
950- Press ctrl+r to see full output for long tool results
951- Added support for MCP SSE transport
952
953## 0.2.53
954
955- New web fetch tool lets Claude view URLs that you paste in
956- Fixed a bug with JPEG detection
957
958## 0.2.50
959
960- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
961
962## 0.2.49
963
964- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
965
966## 0.2.47
967
968- Press Tab to auto-complete file and folder names
969- Press Shift + Tab to toggle auto-accept for file edits
970- Automatic conversation compaction for infinite conversation length (toggle with /config)
971
972## 0.2.44
973
974- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
975
976## 0.2.41
977
978- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
979- MCP server startup no longer blocks the app from starting up
980
981## 0.2.37
982
983- New /release-notes command lets you view release notes at any time
984- `claude config add/remove` commands now accept multiple values separated by commas or spaces
985
986## 0.2.36
987
988- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
989- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
990
991## 0.2.34
992
993- Vim bindings for text input - enable with /vim or /config
994
995## 0.2.32
996
997- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
998- Fix for some PersistentShell issues
999
1000## 0.2.31
1001
1002- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1003- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1004
1005## 0.2.30
1006
1007- Added ANSI color theme for better terminal compatibility
1008- Fixed issue where slash command arguments weren't being sent properly
1009- (Mac-only) API keys are now stored in macOS Keychain
1010
1011## 0.2.26
1012
1013- New /approved-tools command for managing tool permissions
1014- Word-level diff display for improved code readability
1015- Fuzzy matching for slash commands
1016
1017## 0.2.21
1018
1019- Fuzzy matching for /commands