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 ace0a82 · on Dec 11, 2025 History
1,084 lines · 43.1 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.65
4
5- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
6- Added context window information to status line input
7- Added `fileSuggestion` setting for custom `@` file search commands
8- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
9- Fixed prompt not being saved to history when aborting a query with Escape
10- Fixed Read tool image handling to identify format from bytes instead of file extension
11
12## 2.0.64
13
14- Made auto-compacting instant
15- Agents and bash commands can run asynchronously and send messages to wake up the main agent
16- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
17- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
18- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
19- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
20- Fixed auto-loading .env when using native installer
21- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
22- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
23- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
24- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
25- Bedrock: Improve efficiency of token counting
26- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
27
28## 2.0.62
29
30- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
31- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
32- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
33- Fixed slash command selection not working when multiple commands share the same name
34- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
35- Fixed running versions getting removed because lock file incorrectly going stale
36- Fixed IDE diff tab not closing when rejecting file changes
37
38## 2.0.61
39
40- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
41
42## 2.0.60
43
44- Added background agent support. Agents run in the background while you work
45- Added --disable-slash-commands CLI flag to disable all slash commands
46- Added model name to "Co-Authored-By" commit messages
47- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
48- Updated Fetch to skip summarization for pre-approved websites
49- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
50
51## 2.0.59
52
53- Added --agent CLI flag to override the agent setting for the current session
54- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
55- VS Code: Fixed .claude.json config file being read from incorrect location
56
57## 2.0.58
58
59- Pro users now have access to Opus 4.5 as part of their subscription!
60- Fixed timer duration showing "11m 60s" instead of "12m 0s"
61- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
62
63## 2.0.57
64
65- Added feedback input when rejecting plans, allowing users to tell Claude what to change
66- VSCode: Added streaming message support for real-time response display
67
68## 2.0.56
69
70- Added setting to enable/disable terminal progress bar (OSC 9;4)
71- 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.
72
73## 2.0.55
74
75- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
76- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
77- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
78- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
79
80## 2.0.54
81
82- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
83- Fix issue with excessive iTerm notifications
84
85## 2.0.52
86
87- Fixed duplicate message display when starting Claude with a command line argument
88- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
89- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
90- Permit some uses of `$!` in bash commands
91
92## 2.0.51
93
94- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
95- Introducing Claude Code for Desktop: https://claude.com/download
96- 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
97- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
98- Plan Mode now builds more precise plans and executes more thoroughly
99- Usage limit notifications now easier to understand
100- Switched `/usage` back to "% used"
101- Fixed handling of thinking errors
102- Fixed performance regression
103
104## 2.0.50
105
106- Fixed bug preventing calling MCP tools that have nested references in their input schemas
107- Silenced a noisy but harmless error during upgrades
108- Improved ultrathink text display
109- Improved clarity of 5-hour session limit warning message
110
111## 2.0.49
112
113- Added readline-style ctrl-y for pasting deleted text
114- Improved clarity of usage limit warning message
115- Fixed handling of subagent permissions
116
117## 2.0.47
118
119- Improved error messages and validation for `claude --teleport`
120- Improved error handling in `/usage`
121- Fixed race condition with history entry not getting logged at exit
122- Fixed Vertex AI configuration not being applied from `settings.json`
123
124## 2.0.46
125
126- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
127
128## 2.0.45
129
130- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
131- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
132- Send background tasks to Claude Code on the web by starting a message with `&`
133
134## 2.0.43
135
136- Added `permissionMode` field for custom agents
137- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
138- Added skills frontmatter field to declare skills to auto-load for subagents
139- Added the `SubagentStart` hook event
140- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
141- Fixed duplicate rendering of some messages in the UI
142- Fixed some visual flickers
143- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
144
145## 2.0.42
146
147- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
148
149## 2.0.41
150
151- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
152- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
153- Fixed incorrect labeling of user settings vs project settings in command descriptions
154- Fixed crash when plugin command hooks timeout during execution
155- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
156- Fixed broken security documentation links in trust dialogs and onboarding
157- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
158- ctrl-r history search landing on a slash command no longer cancels the search
159- SDK: Support custom timeouts for hooks
160- Allow more safe git commands to run without approval
161- Plugins: Added support for sharing and installing output styles
162- Teleporting a session from web will automatically set the upstream branch
163
164## 2.0.37
165
166- Fixed how idleness is computed for notifications
167- Hooks: Added matcher values for Notification hook events
168- Output Styles: Added `keep-coding-instructions` option to frontmatter
169
170## 2.0.36
171
172- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
173- Fixed queued messages being incorrectly executed as bash commands
174- Fixed input being lost when typing while a queued message is processed
175
176## 2.0.35
177
178- Improve fuzzy search results when searching commands
179- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
180- 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
181- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
182- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
183
184## 2.0.34
185
186- VSCode Extension: Added setting to configure the initial permission mode for new conversations
187- Improved file path suggestion performance with native Rust-based fuzzy finder
188- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
189- Fixed memory crash when reading or writing large files (especially base64-encoded images)
190
191## 2.0.33
192
193- Native binary installs now launch quicker.
194- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
195- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
196
197## 2.0.32
198
199- Un-deprecate output styles based on community feedback
200- Added `companyAnnouncements` setting for displaying announcements on startup
201- Fixed hook progress messages not updating correctly during PostToolUse hook execution
202
203## 2.0.31
204
205- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
206- Vertex: add support for Web Search on supported models
207- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
208- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
209- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
210- Fixed plugin uninstall not removing plugins
211
212## 2.0.30
213
214- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
215- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
216- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
217- Added prompt-based stop hooks
218- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
219- Enabled SSE MCP servers on native build
220- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
221- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
222- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
223- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
224- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
225- Fixed bug that saved session permissions to local settings
226- Fixed MCP tools not being available to sub-agents
227- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
228- Fixed delay when navigating through typeahead suggestions with arrow keys
229- VSCode: Restored selection indicator in input footer showing current file or code selection status
230
231## 2.0.28
232
233- Plan mode: introduced new Plan subagent
234- Subagents: claude can now choose to resume subagents
235- Subagents: claude can dynamically choose the model used by its subagents
236- SDK: added --max-budget-usd flag
237- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
238- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
239- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
240- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
241- Various other bug fixes
242
243## 2.0.27
244
245- New UI for permission prompts
246- Added current branch filtering and search to session resume screen for easier navigation
247- Fixed directory @-mention causing "No assistant message found" error
248- VSCode Extension: Add config setting to include .gitignored files in file searches
249- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
250
251## 2.0.25
252
253- Removed legacy SDK entrypoint. Please migrate to @anthropic-ai/claude-agent-sdk for future SDK updates: https://platform.claude.com/docs/en/agent-sdk/migration-guide
254
255## 2.0.24
256
257- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
258- Claude Code Web: Support for Web -> CLI teleport
259- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
260- Bedrock: Display awsAuthRefresh output when auth is required
261
262## 2.0.22
263
264- Fixed content layout shift when scrolling through slash commands
265- IDE: Add toggle to enable/disable thinking.
266- Fix bug causing duplicate permission prompts with parallel tool calls
267- Add support for enterprise managed MCP allowlist and denylist
268
269## 2.0.21
270
271- Support MCP `structuredContent` field in tool responses
272- Added an interactive question tool
273- Claude will now ask you questions more often in plan mode
274- Added Haiku 4.5 as a model option for Pro users
275- Fixed an issue where queued commands don't have access to previous messages' output
276
277## 2.0.20
278
279- Added support for Claude Skills
280
281## 2.0.19
282
283- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
284- Fixed a bug where Haiku was unnecessarily called in print mode
285
286## 2.0.17
287
288- Added Haiku 4.5 to model selector!
289- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
290- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
291- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
292- OTEL: support HTTP_PROXY and HTTPS_PROXY
293- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
294
295## 2.0.15
296
297- Fixed bug with resuming where previously created files needed to be read again before writing
298- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
299
300## 2.0.14
301
302- Fix @-mentioning MCP servers to toggle them on/off
303- Improve permission checks for bash with inline env vars
304- Fix ultrathink + thinking toggle
305- Reduce unnecessary logins
306- Document --system-prompt
307- Several improvements to rendering
308- Plugins UI polish
309
310## 2.0.13
311
312- Fixed `/plugin` not working on native build
313
314## 2.0.12
315
316- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
317- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
318- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
319- `/plugin validate` command for validating plugin structure and configuration
320- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
321- Plugin documentation available at https://code.claude.com/docs/en/plugins
322- Comprehensive error messages and diagnostics via `/doctor` command
323- Avoid flickering in `/model` selector
324- Improvements to `/help`
325- Avoid mentioning hooks in `/resume` summaries
326- Changes to the "verbose" setting in `/config` now persist across sessions
327
328## 2.0.11
329
330- Reduced system prompt size by 1.4k tokens
331- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
332- Fixed Opus fallback rate limit errors appearing incorrectly
333- Fixed /add-dir command selecting wrong default tab
334
335## 2.0.10
336
337- Rewrote terminal renderer for buttery smooth UI
338- Enable/disable MCP servers by @mentioning, or in /mcp
339- Added tab completion for shell commands in bash mode
340- PreToolUse hooks can now modify tool inputs
341- Press Ctrl-G to edit your prompt in your system's configured text editor
342- Fixes for bash permission checks with environment variables in the command
343
344## 2.0.9
345
346- Fix regression where bash backgrounding stopped working
347
348## 2.0.8
349
350- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
351- IDE: Add drag-and-drop support for files and folders in chat
352- /context: Fix counting for thinking blocks
353- Improve message rendering for users with light themes on dark terminals
354- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
355
356## 2.0.5
357
358- IDE: Fix IME unintended message submission with Enter and Tab
359- IDE: Add "Open in Terminal" link in login screen
360- Fix unhandled OAuth expiration 401 API errors
361- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
362
363## 2.0.1
364
365- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
366- Various bug fixes and presentation improvements
367
368## 2.0.0
369
370- New native VS Code extension
371- Fresh coat of paint throughout the whole app
372- /rewind a conversation to undo code changes
373- /usage command to see plan limits
374- Tab to toggle thinking (sticky across sessions)
375- Ctrl-R to search history
376- Unshipped claude config command
377- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
378- SDK: The Claude Code SDK is now the Claude Agent SDK
379- Add subagents dynamically with `--agents` flag
380
381## 1.0.126
382
383- Enable /context command for Bedrock and Vertex
384- Add mTLS support for HTTP-based OpenTelemetry exporters
385
386## 1.0.124
387
388- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
389- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
390- No longer inform Claude of the list of allowed tools when permission is denied
391- Fixed security vulnerability in Bash tool permission checks
392- Improved VSCode extension performance for large files
393
394## 1.0.123
395
396- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
397- Fixed thinking mode triggering on negation phrases like "don't think"
398- Fixed rendering performance degradation during token streaming
399- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
400- Enhanced BashTool environment snapshot logging
401- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
402- Migrated --debug logging to a file, to enable easy tailing & filtering
403
404## 1.0.120
405
406- Fix input lag during typing, especially noticeable with large prompts
407- Improved VSCode extension command registry and sessions dialog user experience
408- Enhanced sessions dialog responsiveness and visual feedback
409- Fixed IDE compatibility issue by removing worktree support check
410- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
411
412## 1.0.119
413
414- Fix Windows issue where process visually freezes on entering interactive mode
415- Support dynamic headers for MCP servers via headersHelper configuration
416- Fix thinking mode not working in headless sessions
417- Fix slash commands now properly update allowed tools instead of replacing them
418
419## 1.0.117
420
421- Add Ctrl-R history search to recall previous commands like bash/zsh
422- Fix input lag while typing, especially on Windows
423- Add sed command to auto-allowed commands in acceptEdits mode
424- Fix Windows PATH comparison to be case-insensitive for drive letters
425- Add permissions management hint to /add-dir output
426
427## 1.0.115
428
429- Improve thinking mode display with enhanced visual effects
430- Type /t to temporarily disable thinking mode in your prompt
431- Improve path validation for glob and grep tools
432- Show condensed output for post-tool hooks to reduce visual clutter
433- Fix visual feedback when loading state completes
434- Improve UI consistency for permission request dialogs
435
436## 1.0.113
437
438- Deprecated piped input in interactive mode
439- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
440
441## 1.0.112
442
443- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
444- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
445- Hooks: Added systemMessage support for SessionEnd hooks
446- Added `spinnerTipsEnabled` setting to disable spinner tips
447- IDE: Various improvements and bug fixes
448
449## 1.0.111
450
451- /model now validates provided model names
452- Fixed Bash tool crashes caused by malformed shell syntax parsing
453
454## 1.0.110
455
456- /terminal-setup command now supports WezTerm
457- MCP: OAuth tokens now proactively refresh before expiration
458- Fixed reliability issues with background Bash processes
459
460## 1.0.109
461
462- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
463
464## 1.0.106
465
466- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
467
468## 1.0.97
469
470- Settings: /doctor now validates permission rule syntax and suggests corrections
471
472## 1.0.94
473
474- Vertex: add support for global endpoints for supported models
475- /memory command now allows direct editing of all imported memory files
476- SDK: Add custom tools as callbacks
477- Added /todos command to list current todo items
478
479## 1.0.93
480
481- Windows: Add alt + v shortcut for pasting images from clipboard
482- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
483
484## 1.0.90
485
486- Settings file changes take effect immediately - no restart required
487
488## 1.0.88
489
490- Fixed issue causing "OAuth authentication is currently not supported"
491- Status line input now includes `exceeds_200k_tokens`
492- Fixed incorrect usage tracking in /cost.
493- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
494- Bedrock: Updated default Sonnet model to Sonnet 4
495
496## 1.0.86
497
498- Added /context to help users self-serve debug context issues
499- SDK: Added UUID support for all SDK messages
500- SDK: Added `--replay-user-messages` to replay user messages back to stdout
501
502## 1.0.85
503
504- Status line input now includes session cost info
505- Hooks: Introduced SessionEnd hook
506
507## 1.0.84
508
509- Fix tool_use/tool_result id mismatch error when network is unstable
510- Fix Claude sometimes ignoring real-time steering when wrapping up a task
511- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
512- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
513
514## 1.0.83
515
516- @-mention: Support files with spaces in path
517- New shimmering spinner
518
519## 1.0.82
520
521- SDK: Add request cancellation support
522- SDK: New additionalDirectories option to search custom paths, improved slash command processing
523- Settings: Validation prevents invalid fields in .claude/settings.json files
524- MCP: Improve tool name consistency
525- Bash: Fix crash when Claude tries to automatically read large files
526
527## 1.0.81
528
529- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
530- Agents: Fix custom agent loading when agent files are unparsable
531
532## 1.0.80
533
534- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
535
536## 1.0.77
537
538- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
539- SDK: Add session support and permission denial tracking
540- Fix token limit errors in conversation summarization
541- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
542
543## 1.0.73
544
545- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
546- MCP: Press Esc to cancel OAuth authentication flows
547- Bash: Improved command validation and reduced false security warnings
548- UI: Enhanced spinner animations and status line visual hierarchy
549- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
550
551## 1.0.72
552
553- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
554
555## 1.0.71
556
557- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
558- Customizable status line: add your terminal prompt to Claude Code with /statusline
559
560## 1.0.70
561
562- Performance: Optimized message rendering for better performance with large contexts
563- Windows: Fixed native file search, ripgrep, and subagent functionality
564- Added support for @-mentions in slash command arguments
565
566## 1.0.69
567
568- Upgraded Opus to version 4.1
569
570## 1.0.68
571
572- Fix incorrect model names being used for certain commands like `/pr-comments`
573- 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.
574- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
575- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
576- SDK: Added canUseTool callback support for tool confirmation
577- Added `disableAllHooks` setting
578- Improved file suggestions performance in large repos
579
580## 1.0.65
581
582- IDE: Fixed connection stability issues and error handling for diagnostics
583- Windows: Fixed shell environment setup for users without .bashrc files
584
585## 1.0.64
586
587- Agents: Added model customization support - you can now specify which model an agent should use
588- Agents: Fixed unintended access to the recursive agent tool
589- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
590- SDK: Fixed user input tracking across multi-turn conversations
591- Added hidden files to file search and @-mention suggestions
592
593## 1.0.63
594
595- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
596
597## 1.0.62
598
599- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
600- Hooks: Added SessionStart hook for new session initialization
601- /add-dir command now supports typeahead for directory paths
602- Improved network connectivity check reliability
603
604## 1.0.61
605
606- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
607- Settings: Added `--settings` flag to load settings from a JSON file
608- Settings: Fixed resolution of settings files paths that are symlinks
609- OTEL: Fixed reporting of wrong organization after authentication changes
610- Slash commands: Fixed permissions checking for allowed-tools with Bash
611- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
612- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
613- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
614
615## 1.0.60
616
617- You can now create custom subagents for specialized tasks! Run /agents to get started
618
619## 1.0.59
620
621- SDK: Added tool confirmation support with canUseTool callback
622- SDK: Allow specifying env for spawned process
623- Hooks: Exposed PermissionDecision to hooks (including "ask")
624- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
625- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
626
627## 1.0.58
628
629- Added support for reading PDFs
630- MCP: Improved server health status display in 'claude mcp list'
631- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
632
633## 1.0.57
634
635- Added support for specifying a model in slash commands
636- Improved permission messages to help Claude understand allowed tools
637- Fix: Remove trailing newlines from bash output in terminal wrapping
638
639## 1.0.56
640
641- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
642- Fixes for WSL IDE detection
643- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
644
645## 1.0.55
646
647- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
648- Windows: fixed Ctrl+Z crash
649- SDK: Added ability to capture error logging
650- Add --system-prompt-file option to override system prompt in print mode
651
652## 1.0.54
653
654- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
655- Custom slash commands: Added argument-hint to frontmatter
656- Windows: OAuth uses port 45454 and properly constructs browser URL
657- Windows: mode switching now uses alt + m, and plan mode renders properly
658- Shell: Switch to in-memory shell snapshot to fix file-related errors
659
660## 1.0.53
661
662- Updated @-mention file truncation from 100 lines to 2000 lines
663- 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).
664
665## 1.0.52
666
667- Added support for MCP server instructions
668
669## 1.0.51
670
671- Added support for native Windows (requires Git for Windows)
672- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
673- Settings: /doctor can now help you identify and fix invalid setting files
674- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
675- Increased auto-compact warning threshold from 60% to 80%
676- Fixed an issue with handling user directories with spaces for shell snapshots
677- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
678- Custom slash commands: Fixed user-level commands in subdirectories
679- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
680
681## 1.0.48
682
683- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
684- Added progress messages to Bash tool based on the last 5 lines of command output
685- Added expanding variables support for MCP server configuration
686- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
687- Improved IDE extension path handling when Claude Code runs in WSL
688- Hooks: Added a PreCompact hook
689- Vim mode: Added c, f/F, t/T
690
691## 1.0.45
692
693- Redesigned Search (Grep) tool with new tool input parameters and features
694- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
695- Fixed config file corruption issue by enforcing atomic writes
696- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
697- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
698- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
699
700## 1.0.44
701
702- New /export command lets you quickly export a conversation for sharing
703- MCP: resource_link tool results are now supported
704- MCP: tool annotations and tool titles now display in /mcp view
705- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
706
707## 1.0.43
708
709- Fixed a bug where the theme selector was saving excessively
710- Hooks: Added EPIPE system error handling
711
712## 1.0.42
713
714- Added tilde (`~`) expansion support to `/add-dir` command
715
716## 1.0.41
717
718- Hooks: Split Stop hook triggering into Stop and SubagentStop
719- Hooks: Enabled optional timeout configuration for each command
720- Hooks: Added "hook_event_name" to hook input
721- Fixed a bug where MCP tools would display twice in tool list
722- New tool parameters JSON for Bash tool in `tool_decision` event
723
724## 1.0.40
725
726- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
727
728## 1.0.39
729
730- New Active Time metric in OpenTelemetry logging
731
732## 1.0.38
733
734- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
735
736## 1.0.37
737
738- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
739
740## 1.0.36
741
742- Web search now takes today's date into context
743- Fixed a bug where stdio MCP servers were not terminating properly on exit
744
745## 1.0.35
746
747- Added support for MCP OAuth Authorization Server discovery
748
749## 1.0.34
750
751- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
752
753## 1.0.33
754
755- Improved logging functionality with session ID support
756- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
757- Improvements to plan mode
758
759## 1.0.32
760
761- Updated loopback config for litellm
762- Added forceLoginMethod setting to bypass login selection screen
763
764## 1.0.31
765
766- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
767
768## 1.0.30
769
770- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
771- Improved file path autocomplete with filename matching
772- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
773- Enhanced jq regex support for complex filters with pipes and select
774
775## 1.0.29
776
777- Improved CJK character support in cursor navigation and rendering
778
779## 1.0.28
780
781- Slash commands: Fix selector display during history navigation
782- Resizes images before upload to prevent API size limit errors
783- Added XDG_CONFIG_HOME support to configuration directory
784- Performance optimizations for memory usage
785- New attributes (terminal.type, language) in OpenTelemetry logging
786
787## 1.0.27
788
789- Streamable HTTP MCP servers are now supported
790- Remote MCP servers (SSE and HTTP) now support OAuth
791- MCP resources can now be @-mentioned
792- /resume slash command to switch conversations within Claude Code
793
794## 1.0.25
795
796- Slash commands: moved "project" and "user" prefixes to descriptions
797- Slash commands: improved reliability for command discovery
798- Improved support for Ghostty
799- Improved web search reliability
800
801## 1.0.24
802
803- Improved /mcp output
804- Fixed a bug where settings arrays got overwritten instead of merged
805
806## 1.0.23
807
808- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
809- Released Python SDK: pip install claude-code-sdk to get started
810
811## 1.0.22
812
813- SDK: Renamed `total_cost` to `total_cost_usd`
814
815## 1.0.21
816
817- Improved editing of files with tab-based indentation
818- Fix for tool_use without matching tool_result errors
819- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
820
821## 1.0.18
822
823- Added --add-dir CLI argument for specifying additional working directories
824- Added streaming input support without require -p flag
825- Improved startup performance and session storage performance
826- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
827- Added detailed MCP server tools display (/mcp)
828- MCP authentication and permission improvements
829- Added auto-reconnection for MCP SSE connections on disconnect
830- Fixed issue where pasted content was lost when dialogs appeared
831
832## 1.0.17
833
834- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
835- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
836- MCP server list UI improvements
837- Update Claude Code process title to display "claude" instead of "node"
838
839## 1.0.11
840
841- Claude Code can now also be used with a Claude Pro subscription
842- Added /upgrade for smoother switching to Claude Max plans
843- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
844- Improved shell configuration error handling
845- Improved todo list handling during compaction
846
847## 1.0.10
848
849- Added markdown table support
850- Improved streaming performance
851
852## 1.0.8
853
854- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
855- Increased default otel interval from 1s -> 5s
856- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
857- Fixed a regression where search tools unnecessarily asked for permissions
858- Added support for triggering thinking non-English languages
859- Improved compacting UI
860
861## 1.0.7
862
863- Renamed /allowed-tools -> /permissions
864- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
865- Deprecated claude config commands in favor of editing settings.json
866- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
867- Improved error handling for /install-github-app
868- Bugfixes, UI polish, and tool reliability improvements
869
870## 1.0.6
871
872- Improved edit reliability for tab-indented files
873- Respect CLAUDE_CONFIG_DIR everywhere
874- Reduced unnecessary tool permission prompts
875- Added support for symlinks in @file typeahead
876- Bugfixes, UI polish, and tool reliability improvements
877
878## 1.0.4
879
880- Fixed a bug where MCP tool errors weren't being parsed correctly
881
882## 1.0.1
883
884- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
885- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
886- Updated documentation links and OAuth process descriptions
887
888## 1.0.0
889
890- Claude Code is now generally available
891- Introducing Sonnet 4 and Opus 4 models
892
893## 0.2.125
894
895- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
896- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
897
898## 0.2.117
899
900- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
901- Introduced settings.cleanupPeriodDays
902- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
903- Introduced --debug mode
904
905## 0.2.108
906
907- You can now send messages to Claude while it works to steer Claude in real-time
908- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
909- Fixed a bug where thinking was not working in -p mode
910- Fixed a regression in /cost reporting
911- Deprecated MCP wizard interface in favor of other MCP commands
912- Lots of other bugfixes and improvements
913
914## 0.2.107
915
916- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
917
918## 0.2.106
919
920- MCP SSE server configs can now specify custom headers
921- Fixed a bug where MCP permission prompt didn't always show correctly
922
923## 0.2.105
924
925- Claude can now search the web
926- Moved system & account status to /status
927- Added word movement keybindings for Vim
928- Improved latency for startup, todo tool, and file edits
929
930## 0.2.102
931
932- Improved thinking triggering reliability
933- Improved @mention reliability for images and folders
934- You can now paste multiple large chunks into one prompt
935
936## 0.2.100
937
938- Fixed a crash caused by a stack overflow error
939- Made db storage optional; missing db support disables --continue and --resume
940
941## 0.2.98
942
943- Fixed an issue where auto-compact was running twice
944
945## 0.2.96
946
947- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
948
949## 0.2.93
950
951- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
952- Claude now has access to a Todo list that helps it stay on track and be more organized
953
954## 0.2.82
955
956- Added support for --disallowedTools
957- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
958
959## 0.2.75
960
961- Hit Enter to queue up additional messages while Claude is working
962- Drag in or copy/paste image files directly into the prompt
963- @-mention files to directly add them to context
964- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
965- Improved performance for filename auto-complete
966
967## 0.2.74
968
969- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
970- Task tool can now perform writes and run bash commands
971
972## 0.2.72
973
974- Updated spinner to indicate tokens loaded and tool usage
975
976## 0.2.70
977
978- Network commands like curl are now available for Claude to use
979- Claude can now run multiple web queries in parallel
980- Pressing ESC once immediately interrupts Claude in Auto-accept mode
981
982## 0.2.69
983
984- Fixed UI glitches with improved Select component behavior
985- Enhanced terminal output display with better text truncation logic
986
987## 0.2.67
988
989- Shared project permission rules can be saved in .claude/settings.json
990
991## 0.2.66
992
993- Print mode (-p) now supports streaming output via --output-format=stream-json
994- Fixed issue where pasting could trigger memory or bash mode unexpectedly
995
996## 0.2.63
997
998- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
999
1000## 0.2.61
1001
1002- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1003- Enhanced image detection for more reliable clipboard paste functionality
1004- Fixed an issue where ESC key could crash the conversation history selector
1005
1006## 0.2.59
1007
1008- Copy+paste images directly into your prompt
1009- Improved progress indicators for bash and fetch tools
1010- Bugfixes for non-interactive mode (-p)
1011
1012## 0.2.54
1013
1014- Quickly add to Memory by starting your message with '#'
1015- Press ctrl+r to see full output for long tool results
1016- Added support for MCP SSE transport
1017
1018## 0.2.53
1019
1020- New web fetch tool lets Claude view URLs that you paste in
1021- Fixed a bug with JPEG detection
1022
1023## 0.2.50
1024
1025- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1026
1027## 0.2.49
1028
1029- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1030
1031## 0.2.47
1032
1033- Press Tab to auto-complete file and folder names
1034- Press Shift + Tab to toggle auto-accept for file edits
1035- Automatic conversation compaction for infinite conversation length (toggle with /config)
1036
1037## 0.2.44
1038
1039- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1040
1041## 0.2.41
1042
1043- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1044- MCP server startup no longer blocks the app from starting up
1045
1046## 0.2.37
1047
1048- New /release-notes command lets you view release notes at any time
1049- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1050
1051## 0.2.36
1052
1053- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1054- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1055
1056## 0.2.34
1057
1058- Vim bindings for text input - enable with /vim or /config
1059
1060## 0.2.32
1061
1062- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1063- Fix for some PersistentShell issues
1064
1065## 0.2.31
1066
1067- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1068- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1069
1070## 0.2.30
1071
1072- Added ANSI color theme for better terminal compatibility
1073- Fixed issue where slash command arguments weren't being sent properly
1074- (Mac-only) API keys are now stored in macOS Keychain
1075
1076## 0.2.26
1077
1078- New /approved-tools command for managing tool permissions
1079- Word-level diff display for improved code readability
1080- Fuzzy matching for slash commands
1081
1082## 0.2.21
1083
1084- Fuzzy matching for /commands