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 3464c79 · on Nov 24, 2025 History
1,002 lines · 38.5 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.52
4
5- Fixed duplicate message display when starting Claude with a command line argument
6- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
7- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
8- Permit some uses of `$!` in bash commands
9
10## 2.0.51
11
12- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
13- Introducing Claude Code for Desktop: https://claude.com/download
14- 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
15- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
16- Plan Mode now builds more precise plans and executes more thoroughly
17- Usage limit notifications now easier to understand
18- Switched `/usage` back to "% used"
19- Fixed handling of thinking errors
20- Fixed performance regression
21
22## 2.0.50
23
24- Fixed bug preventing calling MCP tools that have nested references in their input schemas
25- Silenced a noisy but harmless error during upgrades
26- Improved ultrathink text display
27- Improved clarity of 5-hour session limit warning message
28
29## 2.0.49
30
31- Added readline-style ctrl-y for pasting deleted text
32- Improved clarity of usage limit warning message
33- Fixed handling of subagent permissions
34
35## 2.0.47
36
37- Improved error messages and validation for `claude --teleport`
38- Improved error handling in `/usage`
39- Fixed race condition with history entry not getting logged at exit
40- Fixed Vertex AI configuration not being applied from `settings.json`
41
42## 2.0.46
43
44- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
45
46## 2.0.45
47
48- Add support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
49- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
50- Send background tasks to Claude Code on the web by starting a message with `&`
51
52## 2.0.43
53
54- Added `permissionMode` field for custom agents
55- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
56- Added skills frontmatter field to declare skills to auto-load for subagents
57- Added the `SubagentStart` hook event
58- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
59- Fixed duplicate rendering of some messages in the UI
60- Fixed some visual flickers
61- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
62
63## 2.0.42
64
65- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
66
67## 2.0.41
68
69- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
70- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
71- Fixed incorrect labeling of user settings vs project settings in command descriptions
72- Fixed crash when plugin command hooks timeout during execution
73- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
74- Fixed broken security documentation links in trust dialogs and onboarding
75- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
76- ctrl-r history search landing on a slash command no longer cancels the search
77- SDK: Support custom timeouts for hooks
78- Allow more safe git commands to run without approval
79- Plugins: Added support for sharing and installing output styles
80- Teleporting a session from web will automatically set the upstream branch
81
82## 2.0.37
83
84- Fixed how idleness is computed for notifications
85- Hooks: Added matcher values for Notification hook events
86- Output Styles: Added `keep-coding-instructions` option to frontmatter
87
88## 2.0.36
89
90- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
91- Fixed queued messages being incorrectly executed as bash commands
92- Fixed input being lost when typing while a queued message is processed
93
94## 2.0.35
95
96- Improve fuzzy search results when searching commands
97- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
98- 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
99- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
100- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
101
102## 2.0.34
103
104- VSCode Extension: Added setting to configure the initial permission mode for new conversations
105- Improved file path suggestion performance with native Rust-based fuzzy finder
106- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
107- Fixed memory crash when reading or writing large files (especially base64-encoded images)
108
109## 2.0.33
110
111- Native binary installs now launch quicker.
112- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
113- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
114
115## 2.0.32
116
117- Un-deprecate output styles based on community feedback
118- Added `companyAnnouncements` setting for displaying announcements on startup
119- Fixed hook progress messages not updating correctly during PostToolUse hook execution
120
121## 2.0.31
122
123- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
124- Vertex: add support for Web Search on supported models
125- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
126- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
127- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
128- Fixed plugin uninstall not removing plugins
129
130## 2.0.30
131
132- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
133- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
134- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
135- Added prompt-based stop hooks
136- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
137- Enabled SSE MCP servers on native build
138- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
139- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
140- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
141- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
142- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
143- Fixed bug that saved session permissions to local settings
144- Fixed MCP tools not being available to sub-agents
145- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
146- Fixed delay when navigating through typeahead suggestions with arrow keys
147- VSCode: Restored selection indicator in input footer showing current file or code selection status
148
149## 2.0.28
150
151- Plan mode: introduced new Plan subagent
152- Subagents: claude can now choose to resume subagents
153- Subagents: claude can dynamically choose the model used by its subagents
154- SDK: added --max-budget-usd flag
155- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
156- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
157- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
158- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
159- Various other bug fixes
160
161## 2.0.27
162
163- New UI for permission prompts
164- Added current branch filtering and search to session resume screen for easier navigation
165- Fixed directory @-mention causing "No assistant message found" error
166- VSCode Extension: Add config setting to include .gitignored files in file searches
167- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
168
169## 2.0.25
170
171- 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
172
173## 2.0.24
174
175- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
176- Claude Code Web: Support for Web -> CLI teleport
177- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
178- Bedrock: Display awsAuthRefresh output when auth is required
179
180## 2.0.22
181
182- Fixed content layout shift when scrolling through slash commands
183- IDE: Add toggle to enable/disable thinking.
184- Fix bug causing duplicate permission prompts with parallel tool calls
185- Add support for enterprise managed MCP allowlist and denylist
186
187## 2.0.21
188
189- Support MCP `structuredContent` field in tool responses
190- Added an interactive question tool
191- Claude will now ask you questions more often in plan mode
192- Added Haiku 4.5 as a model option for Pro users
193- Fixed an issue where queued commands don't have access to previous messages' output
194
195## 2.0.20
196
197- Added support for Claude Skills
198
199## 2.0.19
200
201- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
202- Fixed a bug where Haiku was unnecessarily called in print mode
203
204## 2.0.17
205
206- Added Haiku 4.5 to model selector!
207- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
208- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
209- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
210- OTEL: support HTTP_PROXY and HTTPS_PROXY
211- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
212
213## 2.0.15
214
215- Fixed bug with resuming where previously created files needed to be read again before writing
216- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
217
218## 2.0.14
219
220- Fix @-mentioning MCP servers to toggle them on/off
221- Improve permission checks for bash with inline env vars
222- Fix ultrathink + thinking toggle
223- Reduce unnecessary logins
224- Document --system-prompt
225- Several improvements to rendering
226- Plugins UI polish
227
228## 2.0.13
229
230- Fixed `/plugin` not working on native build
231
232## 2.0.12
233
234- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
235- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
236- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
237- `/plugin validate` command for validating plugin structure and configuration
238- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
239- Plugin documentation available at https://docs.claude.com/en/docs/claude-code/plugins
240- Comprehensive error messages and diagnostics via `/doctor` command
241- Avoid flickering in `/model` selector
242- Improvements to `/help`
243- Avoid mentioning hooks in `/resume` summaries
244- Changes to the "verbose" setting in `/config` now persist across sessions
245
246## 2.0.11
247
248- Reduced system prompt size by 1.4k tokens
249- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
250- Fixed Opus fallback rate limit errors appearing incorrectly
251- Fixed /add-dir command selecting wrong default tab
252
253## 2.0.10
254
255- Rewrote terminal renderer for buttery smooth UI
256- Enable/disable MCP servers by @mentioning, or in /mcp
257- Added tab completion for shell commands in bash mode
258- PreToolUse hooks can now modify tool inputs
259- Press Ctrl-G to edit your prompt in your system's configured text editor
260- Fixes for bash permission checks with environment variables in the command
261
262## 2.0.9
263
264- Fix regression where bash backgrounding stopped working
265
266## 2.0.8
267
268- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
269- IDE: Add drag-and-drop support for files and folders in chat
270- /context: Fix counting for thinking blocks
271- Improve message rendering for users with light themes on dark terminals
272- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
273
274## 2.0.5
275
276- IDE: Fix IME unintended message submission with Enter and Tab
277- IDE: Add "Open in Terminal" link in login screen
278- Fix unhandled OAuth expiration 401 API errors
279- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
280
281## 2.0.1
282
283- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
284- Various bug fixes and presentation improvements
285
286## 2.0.0
287
288- New native VS Code extension
289- Fresh coat of paint throughout the whole app
290- /rewind a conversation to undo code changes
291- /usage command to see plan limits
292- Tab to toggle thinking (sticky across sessions)
293- Ctrl-R to search history
294- Unshipped claude config command
295- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
296- SDK: The Claude Code SDK is now the Claude Agent SDK
297- Add subagents dynamically with `--agents` flag
298
299## 1.0.126
300
301- Enable /context command for Bedrock and Vertex
302- Add mTLS support for HTTP-based OpenTelemetry exporters
303
304## 1.0.124
305
306- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
307- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
308- No longer inform Claude of the list of allowed tools when permission is denied
309- Fixed security vulnerability in Bash tool permission checks
310- Improved VSCode extension performance for large files
311
312## 1.0.123
313
314- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
315- Fixed thinking mode triggering on negation phrases like "don't think"
316- Fixed rendering performance degradation during token streaming
317- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://docs.claude.com/en/docs/claude-code/slash-commands#SlashCommand-tool
318- Enhanced BashTool environment snapshot logging
319- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
320- Migrated --debug logging to a file, to enable easy tailing & filtering
321
322## 1.0.120
323
324- Fix input lag during typing, especially noticeable with large prompts
325- Improved VSCode extension command registry and sessions dialog user experience
326- Enhanced sessions dialog responsiveness and visual feedback
327- Fixed IDE compatibility issue by removing worktree support check
328- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
329
330## 1.0.119
331
332- Fix Windows issue where process visually freezes on entering interactive mode
333- Support dynamic headers for MCP servers via headersHelper configuration
334- Fix thinking mode not working in headless sessions
335- Fix slash commands now properly update allowed tools instead of replacing them
336
337## 1.0.117
338
339- Add Ctrl-R history search to recall previous commands like bash/zsh
340- Fix input lag while typing, especially on Windows
341- Add sed command to auto-allowed commands in acceptEdits mode
342- Fix Windows PATH comparison to be case-insensitive for drive letters
343- Add permissions management hint to /add-dir output
344
345## 1.0.115
346
347- Improve thinking mode display with enhanced visual effects
348- Type /t to temporarily disable thinking mode in your prompt
349- Improve path validation for glob and grep tools
350- Show condensed output for post-tool hooks to reduce visual clutter
351- Fix visual feedback when loading state completes
352- Improve UI consistency for permission request dialogs
353
354## 1.0.113
355
356- Deprecated piped input in interactive mode
357- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
358
359## 1.0.112
360
361- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
362- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
363- Hooks: Added systemMessage support for SessionEnd hooks
364- Added `spinnerTipsEnabled` setting to disable spinner tips
365- IDE: Various improvements and bug fixes
366
367## 1.0.111
368
369- /model now validates provided model names
370- Fixed Bash tool crashes caused by malformed shell syntax parsing
371
372## 1.0.110
373
374- /terminal-setup command now supports WezTerm
375- MCP: OAuth tokens now proactively refresh before expiration
376- Fixed reliability issues with background Bash processes
377
378## 1.0.109
379
380- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
381
382## 1.0.106
383
384- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
385
386## 1.0.97
387
388- Settings: /doctor now validates permission rule syntax and suggests corrections
389
390## 1.0.94
391
392- Vertex: add support for global endpoints for supported models
393- /memory command now allows direct editing of all imported memory files
394- SDK: Add custom tools as callbacks
395- Added /todos command to list current todo items
396
397## 1.0.93
398
399- Windows: Add alt + v shortcut for pasting images from clipboard
400- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
401
402## 1.0.90
403
404- Settings file changes take effect immediately - no restart required
405
406## 1.0.88
407
408- Fixed issue causing "OAuth authentication is currently not supported"
409- Status line input now includes `exceeds_200k_tokens`
410- Fixed incorrect usage tracking in /cost.
411- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
412- Bedrock: Updated default Sonnet model to Sonnet 4
413
414## 1.0.86
415
416- Added /context to help users self-serve debug context issues
417- SDK: Added UUID support for all SDK messages
418- SDK: Added `--replay-user-messages` to replay user messages back to stdout
419
420## 1.0.85
421
422- Status line input now includes session cost info
423- Hooks: Introduced SessionEnd hook
424
425## 1.0.84
426
427- Fix tool_use/tool_result id mismatch error when network is unstable
428- Fix Claude sometimes ignoring real-time steering when wrapping up a task
429- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
430- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
431
432## 1.0.83
433
434- @-mention: Support files with spaces in path
435- New shimmering spinner
436
437## 1.0.82
438
439- SDK: Add request cancellation support
440- SDK: New additionalDirectories option to search custom paths, improved slash command processing
441- Settings: Validation prevents invalid fields in .claude/settings.json files
442- MCP: Improve tool name consistency
443- Bash: Fix crash when Claude tries to automatically read large files
444
445## 1.0.81
446
447- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://docs.claude.com/en/docs/claude-code/output-styles
448- Agents: Fix custom agent loading when agent files are unparsable
449
450## 1.0.80
451
452- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
453
454## 1.0.77
455
456- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
457- SDK: Add session support and permission denial tracking
458- Fix token limit errors in conversation summarization
459- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
460
461## 1.0.73
462
463- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
464- MCP: Press Esc to cancel OAuth authentication flows
465- Bash: Improved command validation and reduced false security warnings
466- UI: Enhanced spinner animations and status line visual hierarchy
467- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
468
469## 1.0.72
470
471- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
472
473## 1.0.71
474
475- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
476- Customizable status line: add your terminal prompt to Claude Code with /statusline
477
478## 1.0.70
479
480- Performance: Optimized message rendering for better performance with large contexts
481- Windows: Fixed native file search, ripgrep, and subagent functionality
482- Added support for @-mentions in slash command arguments
483
484## 1.0.69
485
486- Upgraded Opus to version 4.1
487
488## 1.0.68
489
490- Fix incorrect model names being used for certain commands like `/pr-comments`
491- 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.
492- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
493- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
494- SDK: Added canUseTool callback support for tool confirmation
495- Added `disableAllHooks` setting
496- Improved file suggestions performance in large repos
497
498## 1.0.65
499
500- IDE: Fixed connection stability issues and error handling for diagnostics
501- Windows: Fixed shell environment setup for users without .bashrc files
502
503## 1.0.64
504
505- Agents: Added model customization support - you can now specify which model an agent should use
506- Agents: Fixed unintended access to the recursive agent tool
507- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
508- SDK: Fixed user input tracking across multi-turn conversations
509- Added hidden files to file search and @-mention suggestions
510
511## 1.0.63
512
513- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
514
515## 1.0.62
516
517- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
518- Hooks: Added SessionStart hook for new session initialization
519- /add-dir command now supports typeahead for directory paths
520- Improved network connectivity check reliability
521
522## 1.0.61
523
524- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
525- Settings: Added `--settings` flag to load settings from a JSON file
526- Settings: Fixed resolution of settings files paths that are symlinks
527- OTEL: Fixed reporting of wrong organization after authentication changes
528- Slash commands: Fixed permissions checking for allowed-tools with Bash
529- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
530- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
531- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
532
533## 1.0.60
534
535- You can now create custom subagents for specialized tasks! Run /agents to get started
536
537## 1.0.59
538
539- SDK: Added tool confirmation support with canUseTool callback
540- SDK: Allow specifying env for spawned process
541- Hooks: Exposed PermissionDecision to hooks (including "ask")
542- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
543- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
544
545## 1.0.58
546
547- Added support for reading PDFs
548- MCP: Improved server health status display in 'claude mcp list'
549- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
550
551## 1.0.57
552
553- Added support for specifying a model in slash commands
554- Improved permission messages to help Claude understand allowed tools
555- Fix: Remove trailing newlines from bash output in terminal wrapping
556
557## 1.0.56
558
559- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
560- Fixes for WSL IDE detection
561- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
562
563## 1.0.55
564
565- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
566- Windows: fixed Ctrl+Z crash
567- SDK: Added ability to capture error logging
568- Add --system-prompt-file option to override system prompt in print mode
569
570## 1.0.54
571
572- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
573- Custom slash commands: Added argument-hint to frontmatter
574- Windows: OAuth uses port 45454 and properly constructs browser URL
575- Windows: mode switching now uses alt + m, and plan mode renders properly
576- Shell: Switch to in-memory shell snapshot to fix file-related errors
577
578## 1.0.53
579
580- Updated @-mention file truncation from 100 lines to 2000 lines
581- 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).
582
583## 1.0.52
584
585- Added support for MCP server instructions
586
587## 1.0.51
588
589- Added support for native Windows (requires Git for Windows)
590- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
591- Settings: /doctor can now help you identify and fix invalid setting files
592- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
593- Increased auto-compact warning threshold from 60% to 80%
594- Fixed an issue with handling user directories with spaces for shell snapshots
595- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
596- Custom slash commands: Fixed user-level commands in subdirectories
597- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
598
599## 1.0.48
600
601- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
602- Added progress messages to Bash tool based on the last 5 lines of command output
603- Added expanding variables support for MCP server configuration
604- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
605- Improved IDE extension path handling when Claude Code runs in WSL
606- Hooks: Added a PreCompact hook
607- Vim mode: Added c, f/F, t/T
608
609## 1.0.45
610
611- Redesigned Search (Grep) tool with new tool input parameters and features
612- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
613- Fixed config file corruption issue by enforcing atomic writes
614- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
615- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
616- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
617
618## 1.0.44
619
620- New /export command lets you quickly export a conversation for sharing
621- MCP: resource_link tool results are now supported
622- MCP: tool annotations and tool titles now display in /mcp view
623- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
624
625## 1.0.43
626
627- Fixed a bug where the theme selector was saving excessively
628- Hooks: Added EPIPE system error handling
629
630## 1.0.42
631
632- Added tilde (`~`) expansion support to `/add-dir` command
633
634## 1.0.41
635
636- Hooks: Split Stop hook triggering into Stop and SubagentStop
637- Hooks: Enabled optional timeout configuration for each command
638- Hooks: Added "hook_event_name" to hook input
639- Fixed a bug where MCP tools would display twice in tool list
640- New tool parameters JSON for Bash tool in `tool_decision` event
641
642## 1.0.40
643
644- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
645
646## 1.0.39
647
648- New Active Time metric in OpenTelemetry logging
649
650## 1.0.38
651
652- 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
653
654## 1.0.37
655
656- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
657
658## 1.0.36
659
660- Web search now takes today's date into context
661- Fixed a bug where stdio MCP servers were not terminating properly on exit
662
663## 1.0.35
664
665- Added support for MCP OAuth Authorization Server discovery
666
667## 1.0.34
668
669- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
670
671## 1.0.33
672
673- Improved logging functionality with session ID support
674- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
675- Improvements to plan mode
676
677## 1.0.32
678
679- Updated loopback config for litellm
680- Added forceLoginMethod setting to bypass login selection screen
681
682## 1.0.31
683
684- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
685
686## 1.0.30
687
688- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
689- Improved file path autocomplete with filename matching
690- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
691- Enhanced jq regex support for complex filters with pipes and select
692
693## 1.0.29
694
695- Improved CJK character support in cursor navigation and rendering
696
697## 1.0.28
698
699- Slash commands: Fix selector display during history navigation
700- Resizes images before upload to prevent API size limit errors
701- Added XDG_CONFIG_HOME support to configuration directory
702- Performance optimizations for memory usage
703- New attributes (terminal.type, language) in OpenTelemetry logging
704
705## 1.0.27
706
707- Streamable HTTP MCP servers are now supported
708- Remote MCP servers (SSE and HTTP) now support OAuth
709- MCP resources can now be @-mentioned
710- /resume slash command to switch conversations within Claude Code
711
712## 1.0.25
713
714- Slash commands: moved "project" and "user" prefixes to descriptions
715- Slash commands: improved reliability for command discovery
716- Improved support for Ghostty
717- Improved web search reliability
718
719## 1.0.24
720
721- Improved /mcp output
722- Fixed a bug where settings arrays got overwritten instead of merged
723
724## 1.0.23
725
726- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
727- Released Python SDK: pip install claude-code-sdk to get started
728
729## 1.0.22
730
731- SDK: Renamed `total_cost` to `total_cost_usd`
732
733## 1.0.21
734
735- Improved editing of files with tab-based indentation
736- Fix for tool_use without matching tool_result errors
737- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
738
739## 1.0.18
740
741- Added --add-dir CLI argument for specifying additional working directories
742- Added streaming input support without require -p flag
743- Improved startup performance and session storage performance
744- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
745- Added detailed MCP server tools display (/mcp)
746- MCP authentication and permission improvements
747- Added auto-reconnection for MCP SSE connections on disconnect
748- Fixed issue where pasted content was lost when dialogs appeared
749
750## 1.0.17
751
752- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
753- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
754- MCP server list UI improvements
755- Update Claude Code process title to display "claude" instead of "node"
756
757## 1.0.11
758
759- Claude Code can now also be used with a Claude Pro subscription
760- Added /upgrade for smoother switching to Claude Max plans
761- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
762- Improved shell configuration error handling
763- Improved todo list handling during compaction
764
765## 1.0.10
766
767- Added markdown table support
768- Improved streaming performance
769
770## 1.0.8
771
772- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
773- Increased default otel interval from 1s -> 5s
774- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
775- Fixed a regression where search tools unnecessarily asked for permissions
776- Added support for triggering thinking non-English languages
777- Improved compacting UI
778
779## 1.0.7
780
781- Renamed /allowed-tools -> /permissions
782- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
783- Deprecated claude config commands in favor of editing settings.json
784- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
785- Improved error handling for /install-github-app
786- Bugfixes, UI polish, and tool reliability improvements
787
788## 1.0.6
789
790- Improved edit reliability for tab-indented files
791- Respect CLAUDE_CONFIG_DIR everywhere
792- Reduced unnecessary tool permission prompts
793- Added support for symlinks in @file typeahead
794- Bugfixes, UI polish, and tool reliability improvements
795
796## 1.0.4
797
798- Fixed a bug where MCP tool errors weren't being parsed correctly
799
800## 1.0.1
801
802- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
803- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
804- Updated documentation links and OAuth process descriptions
805
806## 1.0.0
807
808- Claude Code is now generally available
809- Introducing Sonnet 4 and Opus 4 models
810
811## 0.2.125
812
813- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
814- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
815
816## 0.2.117
817
818- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
819- Introduced settings.cleanupPeriodDays
820- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
821- Introduced --debug mode
822
823## 0.2.108
824
825- You can now send messages to Claude while it works to steer Claude in real-time
826- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
827- Fixed a bug where thinking was not working in -p mode
828- Fixed a regression in /cost reporting
829- Deprecated MCP wizard interface in favor of other MCP commands
830- Lots of other bugfixes and improvements
831
832## 0.2.107
833
834- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
835
836## 0.2.106
837
838- MCP SSE server configs can now specify custom headers
839- Fixed a bug where MCP permission prompt didn't always show correctly
840
841## 0.2.105
842
843- Claude can now search the web
844- Moved system & account status to /status
845- Added word movement keybindings for Vim
846- Improved latency for startup, todo tool, and file edits
847
848## 0.2.102
849
850- Improved thinking triggering reliability
851- Improved @mention reliability for images and folders
852- You can now paste multiple large chunks into one prompt
853
854## 0.2.100
855
856- Fixed a crash caused by a stack overflow error
857- Made db storage optional; missing db support disables --continue and --resume
858
859## 0.2.98
860
861- Fixed an issue where auto-compact was running twice
862
863## 0.2.96
864
865- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
866
867## 0.2.93
868
869- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
870- Claude now has access to a Todo list that helps it stay on track and be more organized
871
872## 0.2.82
873
874- Added support for --disallowedTools
875- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
876
877## 0.2.75
878
879- Hit Enter to queue up additional messages while Claude is working
880- Drag in or copy/paste image files directly into the prompt
881- @-mention files to directly add them to context
882- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
883- Improved performance for filename auto-complete
884
885## 0.2.74
886
887- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
888- Task tool can now perform writes and run bash commands
889
890## 0.2.72
891
892- Updated spinner to indicate tokens loaded and tool usage
893
894## 0.2.70
895
896- Network commands like curl are now available for Claude to use
897- Claude can now run multiple web queries in parallel
898- Pressing ESC once immediately interrupts Claude in Auto-accept mode
899
900## 0.2.69
901
902- Fixed UI glitches with improved Select component behavior
903- Enhanced terminal output display with better text truncation logic
904
905## 0.2.67
906
907- Shared project permission rules can be saved in .claude/settings.json
908
909## 0.2.66
910
911- Print mode (-p) now supports streaming output via --output-format=stream-json
912- Fixed issue where pasting could trigger memory or bash mode unexpectedly
913
914## 0.2.63
915
916- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
917
918## 0.2.61
919
920- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
921- Enhanced image detection for more reliable clipboard paste functionality
922- Fixed an issue where ESC key could crash the conversation history selector
923
924## 0.2.59
925
926- Copy+paste images directly into your prompt
927- Improved progress indicators for bash and fetch tools
928- Bugfixes for non-interactive mode (-p)
929
930## 0.2.54
931
932- Quickly add to Memory by starting your message with '#'
933- Press ctrl+r to see full output for long tool results
934- Added support for MCP SSE transport
935
936## 0.2.53
937
938- New web fetch tool lets Claude view URLs that you paste in
939- Fixed a bug with JPEG detection
940
941## 0.2.50
942
943- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
944
945## 0.2.49
946
947- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
948
949## 0.2.47
950
951- Press Tab to auto-complete file and folder names
952- Press Shift + Tab to toggle auto-accept for file edits
953- Automatic conversation compaction for infinite conversation length (toggle with /config)
954
955## 0.2.44
956
957- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
958
959## 0.2.41
960
961- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
962- MCP server startup no longer blocks the app from starting up
963
964## 0.2.37
965
966- New /release-notes command lets you view release notes at any time
967- `claude config add/remove` commands now accept multiple values separated by commas or spaces
968
969## 0.2.36
970
971- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
972- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
973
974## 0.2.34
975
976- Vim bindings for text input - enable with /vim or /config
977
978## 0.2.32
979
980- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
981- Fix for some PersistentShell issues
982
983## 0.2.31
984
985- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
986- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
987
988## 0.2.30
989
990- Added ANSI color theme for better terminal compatibility
991- Fixed issue where slash command arguments weren't being sent properly
992- (Mac-only) API keys are now stored in macOS Keychain
993
994## 0.2.26
995
996- New /approved-tools command for managing tool permissions
997- Word-level diff display for improved code readability
998- Fuzzy matching for slash commands
999
1000## 0.2.21
1001
1002- Fuzzy matching for /commands