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
1,139 lines · 46.8 KBPreviewCodeRaw
1# Changelog
2
3## 2.0.71
4
5- Added `blockedMarketplaces` to enterprise managed settings, allowing administrators to prevent specific plugin marketplaces from being installed
6- Added /config toggle to enable/disable prompt suggestions
7- Added `/settings` as an alias for the `/config` command
8- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
9- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
10- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
11- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
12- New syntax highlighting engine for native build
13
14## 2.0.70
15
16- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
17- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
18- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
19- Added `plan_mode_required` spawn parameter for teammates to require plan approval before implementing changes
20- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
21- Fixed input being cleared when processing queued commands while the user was typing
22- Fixed prompt suggestions replacing typed input when pressing Tab
23- Fixed diff view not updating when terminal is resized
24- Improved memory usage by 3x for large conversations
25- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
26- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
27- Fix thinking mode toggle in /config not persisting correctly
28- Improve UI for file creation permission dialog
29
30## 2.0.69
31
32- Minor bugfixes
33
34## 2.0.68
35
36- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
37- Fixed a bug where disallowed MCP tools were visible to the model
38- Fixed an issue where steering messages could be lost while a subagent is working
39- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
40- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
41- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
42
43## 2.0.67
44
45- Thinking mode is now enabled by default for Opus 4.5
46- Thinking mode configuration has moved to /config
47- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
48- Show reason why autoupdater is disabled in `/doctor`
49- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
50- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
51- Fixed scroll position resetting after deleting a permission rule in `/permissions`
52- Fixed word deletion (opt+delete) and word navigation (opt+arrow) not working correctly with non-Latin text such as Cyrillic, Greek, Arabic, Hebrew, Thai, and Chinese
53- Fixed `claude install --force` not bypassing stale lock files
54- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
55- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
56
57## 2.0.65
58
59- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
60- Added context window information to status line input
61- Added `fileSuggestion` setting for custom `@` file search commands
62- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
63- Fixed prompt not being saved to history when aborting a query with Escape
64- Fixed Read tool image handling to identify format from bytes instead of file extension
65
66## 2.0.64
67
68- Made auto-compacting instant
69- Agents and bash commands can run asynchronously and send messages to wake up the main agent
70- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
71- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
72- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
73- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
74- Fixed auto-loading .env when using native installer
75- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
76- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
77- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
78- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
79- Bedrock: Improve efficiency of token counting
80- Bedrock: Add support for `aws login` AWS Management Console credentials
81- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
82
83## 2.0.62
84
85- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
86- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
87- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
88- Fixed slash command selection not working when multiple commands share the same name
89- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
90- Fixed running versions getting removed because lock file incorrectly going stale
91- Fixed IDE diff tab not closing when rejecting file changes
92
93## 2.0.61
94
95- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
96
97## 2.0.60
98
99- Added background agent support. Agents run in the background while you work
100- Added --disable-slash-commands CLI flag to disable all slash commands
101- Added model name to "Co-Authored-By" commit messages
102- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
103- Updated Fetch to skip summarization for pre-approved websites
104- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
105
106## 2.0.59
107
108- Added --agent CLI flag to override the agent setting for the current session
109- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
110- VS Code: Fixed .claude.json config file being read from incorrect location
111
112## 2.0.58
113
114- Pro users now have access to Opus 4.5 as part of their subscription!
115- Fixed timer duration showing "11m 60s" instead of "12m 0s"
116- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
117
118## 2.0.57
119
120- Added feedback input when rejecting plans, allowing users to tell Claude what to change
121- VSCode: Added streaming message support for real-time response display
122
123## 2.0.56
124
125- Added setting to enable/disable terminal progress bar (OSC 9;4)
126- 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.
127
128## 2.0.55
129
130- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
131- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
132- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
133- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
134
135## 2.0.54
136
137- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
138- Fix issue with excessive iTerm notifications
139
140## 2.0.52
141
142- Fixed duplicate message display when starting Claude with a command line argument
143- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
144- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
145- Permit some uses of `$!` in bash commands
146
147## 2.0.51
148
149- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
150- Introducing Claude Code for Desktop: https://claude.com/download
151- 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
152- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
153- Plan Mode now builds more precise plans and executes more thoroughly
154- Usage limit notifications now easier to understand
155- Switched `/usage` back to "% used"
156- Fixed handling of thinking errors
157- Fixed performance regression
158
159## 2.0.50
160
161- Fixed bug preventing calling MCP tools that have nested references in their input schemas
162- Silenced a noisy but harmless error during upgrades
163- Improved ultrathink text display
164- Improved clarity of 5-hour session limit warning message
165
166## 2.0.49
167
168- Added readline-style ctrl-y for pasting deleted text
169- Improved clarity of usage limit warning message
170- Fixed handling of subagent permissions
171
172## 2.0.47
173
174- Improved error messages and validation for `claude --teleport`
175- Improved error handling in `/usage`
176- Fixed race condition with history entry not getting logged at exit
177- Fixed Vertex AI configuration not being applied from `settings.json`
178
179## 2.0.46
180
181- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
182
183## 2.0.45
184
185- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
186- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
187- Send background tasks to Claude Code on the web by starting a message with `&`
188
189## 2.0.43
190
191- Added `permissionMode` field for custom agents
192- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
193- Added skills frontmatter field to declare skills to auto-load for subagents
194- Added the `SubagentStart` hook event
195- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
196- Fixed duplicate rendering of some messages in the UI
197- Fixed some visual flickers
198- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
199
200## 2.0.42
201
202- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
203
204## 2.0.41
205
206- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
207- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
208- Fixed incorrect labeling of user settings vs project settings in command descriptions
209- Fixed crash when plugin command hooks timeout during execution
210- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
211- Fixed broken security documentation links in trust dialogs and onboarding
212- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
213- ctrl-r history search landing on a slash command no longer cancels the search
214- SDK: Support custom timeouts for hooks
215- Allow more safe git commands to run without approval
216- Plugins: Added support for sharing and installing output styles
217- Teleporting a session from web will automatically set the upstream branch
218
219## 2.0.37
220
221- Fixed how idleness is computed for notifications
222- Hooks: Added matcher values for Notification hook events
223- Output Styles: Added `keep-coding-instructions` option to frontmatter
224
225## 2.0.36
226
227- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
228- Fixed queued messages being incorrectly executed as bash commands
229- Fixed input being lost when typing while a queued message is processed
230
231## 2.0.35
232
233- Improve fuzzy search results when searching commands
234- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
235- 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
236- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
237- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
238
239## 2.0.34
240
241- VSCode Extension: Added setting to configure the initial permission mode for new conversations
242- Improved file path suggestion performance with native Rust-based fuzzy finder
243- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
244- Fixed memory crash when reading or writing large files (especially base64-encoded images)
245
246## 2.0.33
247
248- Native binary installs now launch quicker.
249- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
250- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
251
252## 2.0.32
253
254- Un-deprecate output styles based on community feedback
255- Added `companyAnnouncements` setting for displaying announcements on startup
256- Fixed hook progress messages not updating correctly during PostToolUse hook execution
257
258## 2.0.31
259
260- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
261- Vertex: add support for Web Search on supported models
262- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
263- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
264- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
265- Fixed plugin uninstall not removing plugins
266
267## 2.0.30
268
269- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
270- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
271- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
272- Added prompt-based stop hooks
273- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
274- Enabled SSE MCP servers on native build
275- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
276- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
277- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
278- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
279- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
280- Fixed bug that saved session permissions to local settings
281- Fixed MCP tools not being available to sub-agents
282- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
283- Fixed delay when navigating through typeahead suggestions with arrow keys
284- VSCode: Restored selection indicator in input footer showing current file or code selection status
285
286## 2.0.28
287
288- Plan mode: introduced new Plan subagent
289- Subagents: claude can now choose to resume subagents
290- Subagents: claude can dynamically choose the model used by its subagents
291- SDK: added --max-budget-usd flag
292- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
293- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
294- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
295- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
296- Various other bug fixes
297
298## 2.0.27
299
300- New UI for permission prompts
301- Added current branch filtering and search to session resume screen for easier navigation
302- Fixed directory @-mention causing "No assistant message found" error
303- VSCode Extension: Add config setting to include .gitignored files in file searches
304- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
305
306## 2.0.25
307
308- 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
309
310## 2.0.24
311
312- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
313- Claude Code Web: Support for Web -> CLI teleport
314- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
315- Bedrock: Display awsAuthRefresh output when auth is required
316
317## 2.0.22
318
319- Fixed content layout shift when scrolling through slash commands
320- IDE: Add toggle to enable/disable thinking.
321- Fix bug causing duplicate permission prompts with parallel tool calls
322- Add support for enterprise managed MCP allowlist and denylist
323
324## 2.0.21
325
326- Support MCP `structuredContent` field in tool responses
327- Added an interactive question tool
328- Claude will now ask you questions more often in plan mode
329- Added Haiku 4.5 as a model option for Pro users
330- Fixed an issue where queued commands don't have access to previous messages' output
331
332## 2.0.20
333
334- Added support for Claude Skills
335
336## 2.0.19
337
338- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
339- Fixed a bug where Haiku was unnecessarily called in print mode
340
341## 2.0.17
342
343- Added Haiku 4.5 to model selector!
344- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
345- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
346- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
347- OTEL: support HTTP_PROXY and HTTPS_PROXY
348- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
349
350## 2.0.15
351
352- Fixed bug with resuming where previously created files needed to be read again before writing
353- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
354
355## 2.0.14
356
357- Fix @-mentioning MCP servers to toggle them on/off
358- Improve permission checks for bash with inline env vars
359- Fix ultrathink + thinking toggle
360- Reduce unnecessary logins
361- Document --system-prompt
362- Several improvements to rendering
363- Plugins UI polish
364
365## 2.0.13
366
367- Fixed `/plugin` not working on native build
368
369## 2.0.12
370
371- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
372- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
373- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
374- `/plugin validate` command for validating plugin structure and configuration
375- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
376- Plugin documentation available at https://code.claude.com/docs/en/plugins
377- Comprehensive error messages and diagnostics via `/doctor` command
378- Avoid flickering in `/model` selector
379- Improvements to `/help`
380- Avoid mentioning hooks in `/resume` summaries
381- Changes to the "verbose" setting in `/config` now persist across sessions
382
383## 2.0.11
384
385- Reduced system prompt size by 1.4k tokens
386- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
387- Fixed Opus fallback rate limit errors appearing incorrectly
388- Fixed /add-dir command selecting wrong default tab
389
390## 2.0.10
391
392- Rewrote terminal renderer for buttery smooth UI
393- Enable/disable MCP servers by @mentioning, or in /mcp
394- Added tab completion for shell commands in bash mode
395- PreToolUse hooks can now modify tool inputs
396- Press Ctrl-G to edit your prompt in your system's configured text editor
397- Fixes for bash permission checks with environment variables in the command
398
399## 2.0.9
400
401- Fix regression where bash backgrounding stopped working
402
403## 2.0.8
404
405- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
406- IDE: Add drag-and-drop support for files and folders in chat
407- /context: Fix counting for thinking blocks
408- Improve message rendering for users with light themes on dark terminals
409- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
410
411## 2.0.5
412
413- IDE: Fix IME unintended message submission with Enter and Tab
414- IDE: Add "Open in Terminal" link in login screen
415- Fix unhandled OAuth expiration 401 API errors
416- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
417
418## 2.0.1
419
420- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
421- Various bug fixes and presentation improvements
422
423## 2.0.0
424
425- New native VS Code extension
426- Fresh coat of paint throughout the whole app
427- /rewind a conversation to undo code changes
428- /usage command to see plan limits
429- Tab to toggle thinking (sticky across sessions)
430- Ctrl-R to search history
431- Unshipped claude config command
432- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
433- SDK: The Claude Code SDK is now the Claude Agent SDK
434- Add subagents dynamically with `--agents` flag
435
436## 1.0.126
437
438- Enable /context command for Bedrock and Vertex
439- Add mTLS support for HTTP-based OpenTelemetry exporters
440
441## 1.0.124
442
443- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
444- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
445- No longer inform Claude of the list of allowed tools when permission is denied
446- Fixed security vulnerability in Bash tool permission checks
447- Improved VSCode extension performance for large files
448
449## 1.0.123
450
451- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
452- Fixed thinking mode triggering on negation phrases like "don't think"
453- Fixed rendering performance degradation during token streaming
454- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
455- Enhanced BashTool environment snapshot logging
456- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
457- Migrated --debug logging to a file, to enable easy tailing & filtering
458
459## 1.0.120
460
461- Fix input lag during typing, especially noticeable with large prompts
462- Improved VSCode extension command registry and sessions dialog user experience
463- Enhanced sessions dialog responsiveness and visual feedback
464- Fixed IDE compatibility issue by removing worktree support check
465- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
466
467## 1.0.119
468
469- Fix Windows issue where process visually freezes on entering interactive mode
470- Support dynamic headers for MCP servers via headersHelper configuration
471- Fix thinking mode not working in headless sessions
472- Fix slash commands now properly update allowed tools instead of replacing them
473
474## 1.0.117
475
476- Add Ctrl-R history search to recall previous commands like bash/zsh
477- Fix input lag while typing, especially on Windows
478- Add sed command to auto-allowed commands in acceptEdits mode
479- Fix Windows PATH comparison to be case-insensitive for drive letters
480- Add permissions management hint to /add-dir output
481
482## 1.0.115
483
484- Improve thinking mode display with enhanced visual effects
485- Type /t to temporarily disable thinking mode in your prompt
486- Improve path validation for glob and grep tools
487- Show condensed output for post-tool hooks to reduce visual clutter
488- Fix visual feedback when loading state completes
489- Improve UI consistency for permission request dialogs
490
491## 1.0.113
492
493- Deprecated piped input in interactive mode
494- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
495
496## 1.0.112
497
498- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
499- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
500- Hooks: Added systemMessage support for SessionEnd hooks
501- Added `spinnerTipsEnabled` setting to disable spinner tips
502- IDE: Various improvements and bug fixes
503
504## 1.0.111
505
506- /model now validates provided model names
507- Fixed Bash tool crashes caused by malformed shell syntax parsing
508
509## 1.0.110
510
511- /terminal-setup command now supports WezTerm
512- MCP: OAuth tokens now proactively refresh before expiration
513- Fixed reliability issues with background Bash processes
514
515## 1.0.109
516
517- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
518
519## 1.0.106
520
521- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
522
523## 1.0.97
524
525- Settings: /doctor now validates permission rule syntax and suggests corrections
526
527## 1.0.94
528
529- Vertex: add support for global endpoints for supported models
530- /memory command now allows direct editing of all imported memory files
531- SDK: Add custom tools as callbacks
532- Added /todos command to list current todo items
533
534## 1.0.93
535
536- Windows: Add alt + v shortcut for pasting images from clipboard
537- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
538
539## 1.0.90
540
541- Settings file changes take effect immediately - no restart required
542
543## 1.0.88
544
545- Fixed issue causing "OAuth authentication is currently not supported"
546- Status line input now includes `exceeds_200k_tokens`
547- Fixed incorrect usage tracking in /cost.
548- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
549- Bedrock: Updated default Sonnet model to Sonnet 4
550
551## 1.0.86
552
553- Added /context to help users self-serve debug context issues
554- SDK: Added UUID support for all SDK messages
555- SDK: Added `--replay-user-messages` to replay user messages back to stdout
556
557## 1.0.85
558
559- Status line input now includes session cost info
560- Hooks: Introduced SessionEnd hook
561
562## 1.0.84
563
564- Fix tool_use/tool_result id mismatch error when network is unstable
565- Fix Claude sometimes ignoring real-time steering when wrapping up a task
566- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
567- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
568
569## 1.0.83
570
571- @-mention: Support files with spaces in path
572- New shimmering spinner
573
574## 1.0.82
575
576- SDK: Add request cancellation support
577- SDK: New additionalDirectories option to search custom paths, improved slash command processing
578- Settings: Validation prevents invalid fields in .claude/settings.json files
579- MCP: Improve tool name consistency
580- Bash: Fix crash when Claude tries to automatically read large files
581
582## 1.0.81
583
584- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
585- Agents: Fix custom agent loading when agent files are unparsable
586
587## 1.0.80
588
589- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
590
591## 1.0.77
592
593- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
594- SDK: Add session support and permission denial tracking
595- Fix token limit errors in conversation summarization
596- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
597
598## 1.0.73
599
600- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
601- MCP: Press Esc to cancel OAuth authentication flows
602- Bash: Improved command validation and reduced false security warnings
603- UI: Enhanced spinner animations and status line visual hierarchy
604- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
605
606## 1.0.72
607
608- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
609
610## 1.0.71
611
612- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
613- Customizable status line: add your terminal prompt to Claude Code with /statusline
614
615## 1.0.70
616
617- Performance: Optimized message rendering for better performance with large contexts
618- Windows: Fixed native file search, ripgrep, and subagent functionality
619- Added support for @-mentions in slash command arguments
620
621## 1.0.69
622
623- Upgraded Opus to version 4.1
624
625## 1.0.68
626
627- Fix incorrect model names being used for certain commands like `/pr-comments`
628- 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.
629- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
630- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
631- SDK: Added canUseTool callback support for tool confirmation
632- Added `disableAllHooks` setting
633- Improved file suggestions performance in large repos
634
635## 1.0.65
636
637- IDE: Fixed connection stability issues and error handling for diagnostics
638- Windows: Fixed shell environment setup for users without .bashrc files
639
640## 1.0.64
641
642- Agents: Added model customization support - you can now specify which model an agent should use
643- Agents: Fixed unintended access to the recursive agent tool
644- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
645- SDK: Fixed user input tracking across multi-turn conversations
646- Added hidden files to file search and @-mention suggestions
647
648## 1.0.63
649
650- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
651
652## 1.0.62
653
654- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
655- Hooks: Added SessionStart hook for new session initialization
656- /add-dir command now supports typeahead for directory paths
657- Improved network connectivity check reliability
658
659## 1.0.61
660
661- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
662- Settings: Added `--settings` flag to load settings from a JSON file
663- Settings: Fixed resolution of settings files paths that are symlinks
664- OTEL: Fixed reporting of wrong organization after authentication changes
665- Slash commands: Fixed permissions checking for allowed-tools with Bash
666- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
667- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
668- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
669
670## 1.0.60
671
672- You can now create custom subagents for specialized tasks! Run /agents to get started
673
674## 1.0.59
675
676- SDK: Added tool confirmation support with canUseTool callback
677- SDK: Allow specifying env for spawned process
678- Hooks: Exposed PermissionDecision to hooks (including "ask")
679- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
680- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
681
682## 1.0.58
683
684- Added support for reading PDFs
685- MCP: Improved server health status display in 'claude mcp list'
686- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
687
688## 1.0.57
689
690- Added support for specifying a model in slash commands
691- Improved permission messages to help Claude understand allowed tools
692- Fix: Remove trailing newlines from bash output in terminal wrapping
693
694## 1.0.56
695
696- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
697- Fixes for WSL IDE detection
698- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
699
700## 1.0.55
701
702- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
703- Windows: fixed Ctrl+Z crash
704- SDK: Added ability to capture error logging
705- Add --system-prompt-file option to override system prompt in print mode
706
707## 1.0.54
708
709- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
710- Custom slash commands: Added argument-hint to frontmatter
711- Windows: OAuth uses port 45454 and properly constructs browser URL
712- Windows: mode switching now uses alt + m, and plan mode renders properly
713- Shell: Switch to in-memory shell snapshot to fix file-related errors
714
715## 1.0.53
716
717- Updated @-mention file truncation from 100 lines to 2000 lines
718- 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).
719
720## 1.0.52
721
722- Added support for MCP server instructions
723
724## 1.0.51
725
726- Added support for native Windows (requires Git for Windows)
727- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
728- Settings: /doctor can now help you identify and fix invalid setting files
729- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
730- Increased auto-compact warning threshold from 60% to 80%
731- Fixed an issue with handling user directories with spaces for shell snapshots
732- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
733- Custom slash commands: Fixed user-level commands in subdirectories
734- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
735
736## 1.0.48
737
738- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
739- Added progress messages to Bash tool based on the last 5 lines of command output
740- Added expanding variables support for MCP server configuration
741- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
742- Improved IDE extension path handling when Claude Code runs in WSL
743- Hooks: Added a PreCompact hook
744- Vim mode: Added c, f/F, t/T
745
746## 1.0.45
747
748- Redesigned Search (Grep) tool with new tool input parameters and features
749- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
750- Fixed config file corruption issue by enforcing atomic writes
751- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
752- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
753- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
754
755## 1.0.44
756
757- New /export command lets you quickly export a conversation for sharing
758- MCP: resource_link tool results are now supported
759- MCP: tool annotations and tool titles now display in /mcp view
760- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
761
762## 1.0.43
763
764- Fixed a bug where the theme selector was saving excessively
765- Hooks: Added EPIPE system error handling
766
767## 1.0.42
768
769- Added tilde (`~`) expansion support to `/add-dir` command
770
771## 1.0.41
772
773- Hooks: Split Stop hook triggering into Stop and SubagentStop
774- Hooks: Enabled optional timeout configuration for each command
775- Hooks: Added "hook_event_name" to hook input
776- Fixed a bug where MCP tools would display twice in tool list
777- New tool parameters JSON for Bash tool in `tool_decision` event
778
779## 1.0.40
780
781- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
782
783## 1.0.39
784
785- New Active Time metric in OpenTelemetry logging
786
787## 1.0.38
788
789- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
790
791## 1.0.37
792
793- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
794
795## 1.0.36
796
797- Web search now takes today's date into context
798- Fixed a bug where stdio MCP servers were not terminating properly on exit
799
800## 1.0.35
801
802- Added support for MCP OAuth Authorization Server discovery
803
804## 1.0.34
805
806- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
807
808## 1.0.33
809
810- Improved logging functionality with session ID support
811- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
812- Improvements to plan mode
813
814## 1.0.32
815
816- Updated loopback config for litellm
817- Added forceLoginMethod setting to bypass login selection screen
818
819## 1.0.31
820
821- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
822
823## 1.0.30
824
825- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
826- Improved file path autocomplete with filename matching
827- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
828- Enhanced jq regex support for complex filters with pipes and select
829
830## 1.0.29
831
832- Improved CJK character support in cursor navigation and rendering
833
834## 1.0.28
835
836- Slash commands: Fix selector display during history navigation
837- Resizes images before upload to prevent API size limit errors
838- Added XDG_CONFIG_HOME support to configuration directory
839- Performance optimizations for memory usage
840- New attributes (terminal.type, language) in OpenTelemetry logging
841
842## 1.0.27
843
844- Streamable HTTP MCP servers are now supported
845- Remote MCP servers (SSE and HTTP) now support OAuth
846- MCP resources can now be @-mentioned
847- /resume slash command to switch conversations within Claude Code
848
849## 1.0.25
850
851- Slash commands: moved "project" and "user" prefixes to descriptions
852- Slash commands: improved reliability for command discovery
853- Improved support for Ghostty
854- Improved web search reliability
855
856## 1.0.24
857
858- Improved /mcp output
859- Fixed a bug where settings arrays got overwritten instead of merged
860
861## 1.0.23
862
863- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
864- Released Python SDK: pip install claude-code-sdk to get started
865
866## 1.0.22
867
868- SDK: Renamed `total_cost` to `total_cost_usd`
869
870## 1.0.21
871
872- Improved editing of files with tab-based indentation
873- Fix for tool_use without matching tool_result errors
874- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
875
876## 1.0.18
877
878- Added --add-dir CLI argument for specifying additional working directories
879- Added streaming input support without require -p flag
880- Improved startup performance and session storage performance
881- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
882- Added detailed MCP server tools display (/mcp)
883- MCP authentication and permission improvements
884- Added auto-reconnection for MCP SSE connections on disconnect
885- Fixed issue where pasted content was lost when dialogs appeared
886
887## 1.0.17
888
889- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
890- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
891- MCP server list UI improvements
892- Update Claude Code process title to display "claude" instead of "node"
893
894## 1.0.11
895
896- Claude Code can now also be used with a Claude Pro subscription
897- Added /upgrade for smoother switching to Claude Max plans
898- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
899- Improved shell configuration error handling
900- Improved todo list handling during compaction
901
902## 1.0.10
903
904- Added markdown table support
905- Improved streaming performance
906
907## 1.0.8
908
909- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
910- Increased default otel interval from 1s -> 5s
911- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
912- Fixed a regression where search tools unnecessarily asked for permissions
913- Added support for triggering thinking non-English languages
914- Improved compacting UI
915
916## 1.0.7
917
918- Renamed /allowed-tools -> /permissions
919- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
920- Deprecated claude config commands in favor of editing settings.json
921- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
922- Improved error handling for /install-github-app
923- Bugfixes, UI polish, and tool reliability improvements
924
925## 1.0.6
926
927- Improved edit reliability for tab-indented files
928- Respect CLAUDE_CONFIG_DIR everywhere
929- Reduced unnecessary tool permission prompts
930- Added support for symlinks in @file typeahead
931- Bugfixes, UI polish, and tool reliability improvements
932
933## 1.0.4
934
935- Fixed a bug where MCP tool errors weren't being parsed correctly
936
937## 1.0.1
938
939- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
940- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
941- Updated documentation links and OAuth process descriptions
942
943## 1.0.0
944
945- Claude Code is now generally available
946- Introducing Sonnet 4 and Opus 4 models
947
948## 0.2.125
949
950- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
951- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
952
953## 0.2.117
954
955- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
956- Introduced settings.cleanupPeriodDays
957- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
958- Introduced --debug mode
959
960## 0.2.108
961
962- You can now send messages to Claude while it works to steer Claude in real-time
963- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
964- Fixed a bug where thinking was not working in -p mode
965- Fixed a regression in /cost reporting
966- Deprecated MCP wizard interface in favor of other MCP commands
967- Lots of other bugfixes and improvements
968
969## 0.2.107
970
971- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
972
973## 0.2.106
974
975- MCP SSE server configs can now specify custom headers
976- Fixed a bug where MCP permission prompt didn't always show correctly
977
978## 0.2.105
979
980- Claude can now search the web
981- Moved system & account status to /status
982- Added word movement keybindings for Vim
983- Improved latency for startup, todo tool, and file edits
984
985## 0.2.102
986
987- Improved thinking triggering reliability
988- Improved @mention reliability for images and folders
989- You can now paste multiple large chunks into one prompt
990
991## 0.2.100
992
993- Fixed a crash caused by a stack overflow error
994- Made db storage optional; missing db support disables --continue and --resume
995
996## 0.2.98
997
998- Fixed an issue where auto-compact was running twice
999
1000## 0.2.96
1001
1002- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1003
1004## 0.2.93
1005
1006- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1007- Claude now has access to a Todo list that helps it stay on track and be more organized
1008
1009## 0.2.82
1010
1011- Added support for --disallowedTools
1012- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1013
1014## 0.2.75
1015
1016- Hit Enter to queue up additional messages while Claude is working
1017- Drag in or copy/paste image files directly into the prompt
1018- @-mention files to directly add them to context
1019- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1020- Improved performance for filename auto-complete
1021
1022## 0.2.74
1023
1024- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1025- Task tool can now perform writes and run bash commands
1026
1027## 0.2.72
1028
1029- Updated spinner to indicate tokens loaded and tool usage
1030
1031## 0.2.70
1032
1033- Network commands like curl are now available for Claude to use
1034- Claude can now run multiple web queries in parallel
1035- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1036
1037## 0.2.69
1038
1039- Fixed UI glitches with improved Select component behavior
1040- Enhanced terminal output display with better text truncation logic
1041
1042## 0.2.67
1043
1044- Shared project permission rules can be saved in .claude/settings.json
1045
1046## 0.2.66
1047
1048- Print mode (-p) now supports streaming output via --output-format=stream-json
1049- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1050
1051## 0.2.63
1052
1053- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1054
1055## 0.2.61
1056
1057- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1058- Enhanced image detection for more reliable clipboard paste functionality
1059- Fixed an issue where ESC key could crash the conversation history selector
1060
1061## 0.2.59
1062
1063- Copy+paste images directly into your prompt
1064- Improved progress indicators for bash and fetch tools
1065- Bugfixes for non-interactive mode (-p)
1066
1067## 0.2.54
1068
1069- Quickly add to Memory by starting your message with '#'
1070- Press ctrl+r to see full output for long tool results
1071- Added support for MCP SSE transport
1072
1073## 0.2.53
1074
1075- New web fetch tool lets Claude view URLs that you paste in
1076- Fixed a bug with JPEG detection
1077
1078## 0.2.50
1079
1080- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1081
1082## 0.2.49
1083
1084- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1085
1086## 0.2.47
1087
1088- Press Tab to auto-complete file and folder names
1089- Press Shift + Tab to toggle auto-accept for file edits
1090- Automatic conversation compaction for infinite conversation length (toggle with /config)
1091
1092## 0.2.44
1093
1094- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1095
1096## 0.2.41
1097
1098- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1099- MCP server startup no longer blocks the app from starting up
1100
1101## 0.2.37
1102
1103- New /release-notes command lets you view release notes at any time
1104- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1105
1106## 0.2.36
1107
1108- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1109- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1110
1111## 0.2.34
1112
1113- Vim bindings for text input - enable with /vim or /config
1114
1115## 0.2.32
1116
1117- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1118- Fix for some PersistentShell issues
1119
1120## 0.2.31
1121
1122- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1123- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1124
1125## 0.2.30
1126
1127- Added ANSI color theme for better terminal compatibility
1128- Fixed issue where slash command arguments weren't being sent properly
1129- (Mac-only) API keys are now stored in macOS Keychain
1130
1131## 0.2.26
1132
1133- New /approved-tools command for managing tool permissions
1134- Word-level diff display for improved code readability
1135- Fuzzy matching for slash commands
1136
1137## 0.2.21
1138
1139- Fuzzy matching for /commands