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 a856c62 · on Jan 12, 2026 History
1,348 lines · 64.2 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.5
4
5- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
6
7## 2.1.4
8
9- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
10- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
11
12## 2.1.3
13
14- Merged slash commands and skills, simplifying the mental model with no change in behavior
15- Added release channel (`stable` or `latest`) toggle to `/config`
16- Added detection and warnings for unreachable permission rules, with warnings in `/doctor` and after saving rules that include the source of each rule and actionable fix guidance
17- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
18- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
19- Fixed mismatch between background task count in status bar and items shown in tasks dialog
20- Fixed sub-agents using the wrong model during conversation compaction
21- Fixed web search in sub-agents using incorrect model
22- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
23- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
24- Improved slash command suggestion readability by truncating long descriptions to 2 lines
25- Changed tool hook execution timeout from 60 seconds to 10 minutes
26- [VSCode] Added clickable destination selector for permission requests, allowing you to choose where settings are saved (this project, all projects, shared with team, or session only)
27
28## 2.1.2
29
30- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
31- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
32- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
33- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
34- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
35- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
36- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
37- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
38- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
39- Fixed updates incorrectly claiming another installation is in progress
40- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
41- Fixed remote session URL and teleport being broken when using `/tasks` command
42- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
43- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
44- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
45- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
46- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
47- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
48- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
49- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
50- [SDK] Changed minimum zod peer dependency to ^4.0.0
51- [VSCode] Fixed usage display not updating after manual compact
52
53## 2.1.0
54
55- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
56- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
57- Added support for `agent` field in skills to specify agent type for execution
58- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
59- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
60- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
61- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
62- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
63- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
64- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
65- Fixed Esc key with queued prompts to only move them to input without canceling the running task
66- Reduced permission prompts for complex bash commands
67- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
68- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
69- Fixed plugin path resolution for file-based marketplace sources
70- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
71- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
72- Fixed Claude in Chrome support for WSL environments
73- Fixed Windows native installer silently failing when executable creation fails
74- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
75- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
76- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
77- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
78- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
79- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
80- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
81- Added hooks support for skill and slash command frontmatter
82- Added new Vim motions: `;` and `,` to repeat f/F/t/T motions, `y` operator for yank with `yy`/`Y`, `p`/`P` for paste, text objects (`iw`, `aw`, `iW`, `aW`, `i"`, `a"`, `i'`, `a'`, `i(`, `a(`, `i[`, `a[`, `i{`, `a{`), `>>` and `<<` for indent/dedent, and `J` to join lines
83- Added `/plan` command shortcut to enable plan mode directly from the prompt
84- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
85- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
86- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
87- Added support for `once: true` config for hooks
88- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
89- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
90- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
91- Added left/right arrow key navigation for cycling through tabs in dialogs
92- Added real-time thinking block display in Ctrl+O transcript mode
93- Added filepath to full output in background bash task details dialog
94- Added Skills as a separate category in the context visualization
95- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
96- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
97- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
98- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
99- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
100- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
101- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
102- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
103- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
104- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
105- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
106- Fixed commands with `$()` command substitution failing with parse errors
107- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
108- Fixed bash command prefix extraction to correctly identify subcommands after global options (e.g., `git -C /path log` now correctly matches `Bash(git log:*)` rules)
109- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
110- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
111- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
112- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
113- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
114- Fixed subagents sometimes not inheriting the parent's model by default
115- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
116- Fixed duplicate Bash commands appearing in permission request option labels
117- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
118- Fixed background task completion notifications to appear proactively with bullet point
119- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
120- Fixed cursor disappearing after dismissing permission dialogs
121- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
122- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
123- Fixed images being silently dropped when queueing messages while backgrounding a task
124- Fixed large pasted images failing with "Image was too large" error
125- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
126- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
127- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
128- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
129- Fixed spinner token counter to properly accumulate tokens from subagents during execution
130- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
131- Fixed race condition where LSP tool could return "no server available" during startup
132- Fixed feedback submission hanging indefinitely when network requests timeout
133- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
134- Fixed hook success message showing trailing colon when hook has no output
135- Multiple optimizations to improve startup performance
136- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
137- Improved performance when reading Jupyter notebooks with many cells
138- Improved reliability for piped input like `cat refactor.md | claude`
139- Improved reliability for AskQuestion tool
140- Improved sed in-place edit commands to render as file edits with diff preview
141- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
142- Improved compaction reliability
143- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
144- Improved skills to show progress while executing, displaying tool uses as they happen
145- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
146- Improved skill suggestions to prioritize recently and frequently used skills
147- Improved spinner feedback when waiting for the first response token
148- Improved token count display in spinner to include tokens from background agents
149- Improved incremental output for async agents to give the main thread more control and visibility
150- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
151- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
152- Improved macOS screenshot paste reliability with TIFF format support
153- Improved `/stats` output
154- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
155- Changed "Interrupted" message color from red to grey for a less alarming appearance
156- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
157- Removed underline styling from image reference links
158- [SDK] Changed minimum zod peer dependency to ^4.0.0
159- [VSCode] Added currently selected model name to the context menu
160- [VSCode] Added descriptive labels on auto-accept permission button (e.g., "Yes, allow npm for this project" instead of "Yes, and don't ask again")
161- [VSCode] Fixed paragraph breaks not rendering in markdown content
162- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
163- [Windows] Fixed issue with improper rendering
164
165## 2.0.76
166
167- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
168
169## 2.0.75
170
171- Minor bugfixes
172
173## 2.0.74
174
175- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
176- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
177- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
178- Added syntax highlighting info to theme picker
179- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
180- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
181- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
182- Fixed a potential crash when syntax highlighting isn't initialized correctly
183- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
184- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
185- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
186- [Windows] Fixed issue with improper rendering
187- [VSCode] Added gift tag pictogram for year-end promotion message
188
189## 2.0.73
190
191- Added clickable `[Image #N]` links that open attached images in the default viewer
192- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
193- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
194- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
195- Fixed slow input history cycling and race condition that could overwrite text after message submission
196- Improved `/theme` command to open theme picker directly
197- Improved theme picker UI
198- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
199- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
200
201## 2.0.72
202
203- Added Claude in Chrome (Beta) feature that works with the Chrome extension (https://claude.ai/chrome) to let you control your browser directly from Claude Code
204- Reduced terminal flickering
205- Added scannable QR code to mobile app tip for quick app downloads
206- Added loading indicator when resuming conversations for better feedback
207- Fixed `/context` command not respecting custom system prompts in non-interactive mode
208- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
209- Improved @ mention file suggestion speed (~3x faster in git repositories)
210- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
211- Improved settings validation errors to be more prominent
212- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
213
214## 2.0.71
215
216- Added /config toggle to enable/disable prompt suggestions
217- Added `/settings` as an alias for the `/config` command
218- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
219- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
220- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
221- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
222- New syntax highlighting engine for native build
223
224## 2.0.70
225
226- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
227- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
228- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
229- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
230- Fixed input being cleared when processing queued commands while the user was typing
231- Fixed prompt suggestions replacing typed input when pressing Tab
232- Fixed diff view not updating when terminal is resized
233- Improved memory usage by 3x for large conversations
234- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
235- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
236- Fix thinking mode toggle in /config not persisting correctly
237- Improve UI for file creation permission dialog
238
239## 2.0.69
240
241- Minor bugfixes
242
243## 2.0.68
244
245- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
246- Fixed a bug where disallowed MCP tools were visible to the model
247- Fixed an issue where steering messages could be lost while a subagent is working
248- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
249- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
250- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
251
252## 2.0.67
253
254- Thinking mode is now enabled by default for Opus 4.5
255- Thinking mode configuration has moved to /config
256- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
257- Show reason why autoupdater is disabled in `/doctor`
258- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
259- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
260- Fixed scroll position resetting after deleting a permission rule in `/permissions`
261- 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
262- Fixed `claude install --force` not bypassing stale lock files
263- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
264- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
265
266## 2.0.65
267
268- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
269- Added context window information to status line input
270- Added `fileSuggestion` setting for custom `@` file search commands
271- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
272- Fixed prompt not being saved to history when aborting a query with Escape
273- Fixed Read tool image handling to identify format from bytes instead of file extension
274
275## 2.0.64
276
277- Made auto-compacting instant
278- Agents and bash commands can run asynchronously and send messages to wake up the main agent
279- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
280- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
281- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
282- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
283- Fixed auto-loading .env when using native installer
284- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
285- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
286- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
287- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
288- Bedrock: Improve efficiency of token counting
289- Bedrock: Add support for `aws login` AWS Management Console credentials
290- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
291
292## 2.0.62
293
294- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
295- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
296- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
297- Fixed slash command selection not working when multiple commands share the same name
298- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
299- Fixed running versions getting removed because lock file incorrectly going stale
300- Fixed IDE diff tab not closing when rejecting file changes
301
302## 2.0.61
303
304- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
305
306## 2.0.60
307
308- Added background agent support. Agents run in the background while you work
309- Added --disable-slash-commands CLI flag to disable all slash commands
310- Added model name to "Co-Authored-By" commit messages
311- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
312- Updated Fetch to skip summarization for pre-approved websites
313- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
314
315## 2.0.59
316
317- Added --agent CLI flag to override the agent setting for the current session
318- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
319- VS Code: Fixed .claude.json config file being read from incorrect location
320
321## 2.0.58
322
323- Pro users now have access to Opus 4.5 as part of their subscription!
324- Fixed timer duration showing "11m 60s" instead of "12m 0s"
325- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
326
327## 2.0.57
328
329- Added feedback input when rejecting plans, allowing users to tell Claude what to change
330- VSCode: Added streaming message support for real-time response display
331
332## 2.0.56
333
334- Added setting to enable/disable terminal progress bar (OSC 9;4)
335- 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.
336
337## 2.0.55
338
339- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
340- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
341- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
342- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
343
344## 2.0.54
345
346- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
347- Fix issue with excessive iTerm notifications
348
349## 2.0.52
350
351- Fixed duplicate message display when starting Claude with a command line argument
352- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
353- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
354- Permit some uses of `$!` in bash commands
355
356## 2.0.51
357
358- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
359- Introducing Claude Code for Desktop: https://claude.com/download
360- 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
361- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
362- Plan Mode now builds more precise plans and executes more thoroughly
363- Usage limit notifications now easier to understand
364- Switched `/usage` back to "% used"
365- Fixed handling of thinking errors
366- Fixed performance regression
367
368## 2.0.50
369
370- Fixed bug preventing calling MCP tools that have nested references in their input schemas
371- Silenced a noisy but harmless error during upgrades
372- Improved ultrathink text display
373- Improved clarity of 5-hour session limit warning message
374
375## 2.0.49
376
377- Added readline-style ctrl-y for pasting deleted text
378- Improved clarity of usage limit warning message
379- Fixed handling of subagent permissions
380
381## 2.0.47
382
383- Improved error messages and validation for `claude --teleport`
384- Improved error handling in `/usage`
385- Fixed race condition with history entry not getting logged at exit
386- Fixed Vertex AI configuration not being applied from `settings.json`
387
388## 2.0.46
389
390- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
391
392## 2.0.45
393
394- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
395- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
396- Send background tasks to Claude Code on the web by starting a message with `&`
397
398## 2.0.43
399
400- Added `permissionMode` field for custom agents
401- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
402- Added skills frontmatter field to declare skills to auto-load for subagents
403- Added the `SubagentStart` hook event
404- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
405- Fixed duplicate rendering of some messages in the UI
406- Fixed some visual flickers
407- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
408
409## 2.0.42
410
411- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
412
413## 2.0.41
414
415- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
416- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
417- Fixed incorrect labeling of user settings vs project settings in command descriptions
418- Fixed crash when plugin command hooks timeout during execution
419- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
420- Fixed broken security documentation links in trust dialogs and onboarding
421- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
422- ctrl-r history search landing on a slash command no longer cancels the search
423- SDK: Support custom timeouts for hooks
424- Allow more safe git commands to run without approval
425- Plugins: Added support for sharing and installing output styles
426- Teleporting a session from web will automatically set the upstream branch
427
428## 2.0.37
429
430- Fixed how idleness is computed for notifications
431- Hooks: Added matcher values for Notification hook events
432- Output Styles: Added `keep-coding-instructions` option to frontmatter
433
434## 2.0.36
435
436- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
437- Fixed queued messages being incorrectly executed as bash commands
438- Fixed input being lost when typing while a queued message is processed
439
440## 2.0.35
441
442- Improve fuzzy search results when searching commands
443- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
444- 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
445- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
446- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
447
448## 2.0.34
449
450- VSCode Extension: Added setting to configure the initial permission mode for new conversations
451- Improved file path suggestion performance with native Rust-based fuzzy finder
452- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
453- Fixed memory crash when reading or writing large files (especially base64-encoded images)
454
455## 2.0.33
456
457- Native binary installs now launch quicker.
458- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
459- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
460
461## 2.0.32
462
463- Un-deprecate output styles based on community feedback
464- Added `companyAnnouncements` setting for displaying announcements on startup
465- Fixed hook progress messages not updating correctly during PostToolUse hook execution
466
467## 2.0.31
468
469- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
470- Vertex: add support for Web Search on supported models
471- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
472- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
473- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
474- Fixed plugin uninstall not removing plugins
475
476## 2.0.30
477
478- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
479- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
480- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
481- Added prompt-based stop hooks
482- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
483- Enabled SSE MCP servers on native build
484- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
485- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
486- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
487- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
488- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
489- Fixed bug that saved session permissions to local settings
490- Fixed MCP tools not being available to sub-agents
491- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
492- Fixed delay when navigating through typeahead suggestions with arrow keys
493- VSCode: Restored selection indicator in input footer showing current file or code selection status
494
495## 2.0.28
496
497- Plan mode: introduced new Plan subagent
498- Subagents: claude can now choose to resume subagents
499- Subagents: claude can dynamically choose the model used by its subagents
500- SDK: added --max-budget-usd flag
501- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
502- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
503- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
504- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
505- Various other bug fixes
506
507## 2.0.27
508
509- New UI for permission prompts
510- Added current branch filtering and search to session resume screen for easier navigation
511- Fixed directory @-mention causing "No assistant message found" error
512- VSCode Extension: Add config setting to include .gitignored files in file searches
513- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
514
515## 2.0.25
516
517- 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
518
519## 2.0.24
520
521- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
522- Claude Code Web: Support for Web -> CLI teleport
523- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
524- Bedrock: Display awsAuthRefresh output when auth is required
525
526## 2.0.22
527
528- Fixed content layout shift when scrolling through slash commands
529- IDE: Add toggle to enable/disable thinking.
530- Fix bug causing duplicate permission prompts with parallel tool calls
531- Add support for enterprise managed MCP allowlist and denylist
532
533## 2.0.21
534
535- Support MCP `structuredContent` field in tool responses
536- Added an interactive question tool
537- Claude will now ask you questions more often in plan mode
538- Added Haiku 4.5 as a model option for Pro users
539- Fixed an issue where queued commands don't have access to previous messages' output
540
541## 2.0.20
542
543- Added support for Claude Skills
544
545## 2.0.19
546
547- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
548- Fixed a bug where Haiku was unnecessarily called in print mode
549
550## 2.0.17
551
552- Added Haiku 4.5 to model selector!
553- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
554- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
555- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
556- OTEL: support HTTP_PROXY and HTTPS_PROXY
557- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
558
559## 2.0.15
560
561- Fixed bug with resuming where previously created files needed to be read again before writing
562- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
563
564## 2.0.14
565
566- Fix @-mentioning MCP servers to toggle them on/off
567- Improve permission checks for bash with inline env vars
568- Fix ultrathink + thinking toggle
569- Reduce unnecessary logins
570- Document --system-prompt
571- Several improvements to rendering
572- Plugins UI polish
573
574## 2.0.13
575
576- Fixed `/plugin` not working on native build
577
578## 2.0.12
579
580- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
581- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
582- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
583- `/plugin validate` command for validating plugin structure and configuration
584- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
585- Plugin documentation available at https://code.claude.com/docs/en/plugins
586- Comprehensive error messages and diagnostics via `/doctor` command
587- Avoid flickering in `/model` selector
588- Improvements to `/help`
589- Avoid mentioning hooks in `/resume` summaries
590- Changes to the "verbose" setting in `/config` now persist across sessions
591
592## 2.0.11
593
594- Reduced system prompt size by 1.4k tokens
595- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
596- Fixed Opus fallback rate limit errors appearing incorrectly
597- Fixed /add-dir command selecting wrong default tab
598
599## 2.0.10
600
601- Rewrote terminal renderer for buttery smooth UI
602- Enable/disable MCP servers by @mentioning, or in /mcp
603- Added tab completion for shell commands in bash mode
604- PreToolUse hooks can now modify tool inputs
605- Press Ctrl-G to edit your prompt in your system's configured text editor
606- Fixes for bash permission checks with environment variables in the command
607
608## 2.0.9
609
610- Fix regression where bash backgrounding stopped working
611
612## 2.0.8
613
614- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
615- IDE: Add drag-and-drop support for files and folders in chat
616- /context: Fix counting for thinking blocks
617- Improve message rendering for users with light themes on dark terminals
618- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
619
620## 2.0.5
621
622- IDE: Fix IME unintended message submission with Enter and Tab
623- IDE: Add "Open in Terminal" link in login screen
624- Fix unhandled OAuth expiration 401 API errors
625- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
626
627## 2.0.1
628
629- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
630- Various bug fixes and presentation improvements
631
632## 2.0.0
633
634- New native VS Code extension
635- Fresh coat of paint throughout the whole app
636- /rewind a conversation to undo code changes
637- /usage command to see plan limits
638- Tab to toggle thinking (sticky across sessions)
639- Ctrl-R to search history
640- Unshipped claude config command
641- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
642- SDK: The Claude Code SDK is now the Claude Agent SDK
643- Add subagents dynamically with `--agents` flag
644
645## 1.0.126
646
647- Enable /context command for Bedrock and Vertex
648- Add mTLS support for HTTP-based OpenTelemetry exporters
649
650## 1.0.124
651
652- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
653- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
654- No longer inform Claude of the list of allowed tools when permission is denied
655- Fixed security vulnerability in Bash tool permission checks
656- Improved VSCode extension performance for large files
657
658## 1.0.123
659
660- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
661- Fixed thinking mode triggering on negation phrases like "don't think"
662- Fixed rendering performance degradation during token streaming
663- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
664- Enhanced BashTool environment snapshot logging
665- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
666- Migrated --debug logging to a file, to enable easy tailing & filtering
667
668## 1.0.120
669
670- Fix input lag during typing, especially noticeable with large prompts
671- Improved VSCode extension command registry and sessions dialog user experience
672- Enhanced sessions dialog responsiveness and visual feedback
673- Fixed IDE compatibility issue by removing worktree support check
674- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
675
676## 1.0.119
677
678- Fix Windows issue where process visually freezes on entering interactive mode
679- Support dynamic headers for MCP servers via headersHelper configuration
680- Fix thinking mode not working in headless sessions
681- Fix slash commands now properly update allowed tools instead of replacing them
682
683## 1.0.117
684
685- Add Ctrl-R history search to recall previous commands like bash/zsh
686- Fix input lag while typing, especially on Windows
687- Add sed command to auto-allowed commands in acceptEdits mode
688- Fix Windows PATH comparison to be case-insensitive for drive letters
689- Add permissions management hint to /add-dir output
690
691## 1.0.115
692
693- Improve thinking mode display with enhanced visual effects
694- Type /t to temporarily disable thinking mode in your prompt
695- Improve path validation for glob and grep tools
696- Show condensed output for post-tool hooks to reduce visual clutter
697- Fix visual feedback when loading state completes
698- Improve UI consistency for permission request dialogs
699
700## 1.0.113
701
702- Deprecated piped input in interactive mode
703- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
704
705## 1.0.112
706
707- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
708- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
709- Hooks: Added systemMessage support for SessionEnd hooks
710- Added `spinnerTipsEnabled` setting to disable spinner tips
711- IDE: Various improvements and bug fixes
712
713## 1.0.111
714
715- /model now validates provided model names
716- Fixed Bash tool crashes caused by malformed shell syntax parsing
717
718## 1.0.110
719
720- /terminal-setup command now supports WezTerm
721- MCP: OAuth tokens now proactively refresh before expiration
722- Fixed reliability issues with background Bash processes
723
724## 1.0.109
725
726- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
727
728## 1.0.106
729
730- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
731
732## 1.0.97
733
734- Settings: /doctor now validates permission rule syntax and suggests corrections
735
736## 1.0.94
737
738- Vertex: add support for global endpoints for supported models
739- /memory command now allows direct editing of all imported memory files
740- SDK: Add custom tools as callbacks
741- Added /todos command to list current todo items
742
743## 1.0.93
744
745- Windows: Add alt + v shortcut for pasting images from clipboard
746- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
747
748## 1.0.90
749
750- Settings file changes take effect immediately - no restart required
751
752## 1.0.88
753
754- Fixed issue causing "OAuth authentication is currently not supported"
755- Status line input now includes `exceeds_200k_tokens`
756- Fixed incorrect usage tracking in /cost.
757- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
758- Bedrock: Updated default Sonnet model to Sonnet 4
759
760## 1.0.86
761
762- Added /context to help users self-serve debug context issues
763- SDK: Added UUID support for all SDK messages
764- SDK: Added `--replay-user-messages` to replay user messages back to stdout
765
766## 1.0.85
767
768- Status line input now includes session cost info
769- Hooks: Introduced SessionEnd hook
770
771## 1.0.84
772
773- Fix tool_use/tool_result id mismatch error when network is unstable
774- Fix Claude sometimes ignoring real-time steering when wrapping up a task
775- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
776- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
777
778## 1.0.83
779
780- @-mention: Support files with spaces in path
781- New shimmering spinner
782
783## 1.0.82
784
785- SDK: Add request cancellation support
786- SDK: New additionalDirectories option to search custom paths, improved slash command processing
787- Settings: Validation prevents invalid fields in .claude/settings.json files
788- MCP: Improve tool name consistency
789- Bash: Fix crash when Claude tries to automatically read large files
790
791## 1.0.81
792
793- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
794- Agents: Fix custom agent loading when agent files are unparsable
795
796## 1.0.80
797
798- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
799
800## 1.0.77
801
802- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
803- SDK: Add session support and permission denial tracking
804- Fix token limit errors in conversation summarization
805- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
806
807## 1.0.73
808
809- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
810- MCP: Press Esc to cancel OAuth authentication flows
811- Bash: Improved command validation and reduced false security warnings
812- UI: Enhanced spinner animations and status line visual hierarchy
813- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
814
815## 1.0.72
816
817- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
818
819## 1.0.71
820
821- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
822- Customizable status line: add your terminal prompt to Claude Code with /statusline
823
824## 1.0.70
825
826- Performance: Optimized message rendering for better performance with large contexts
827- Windows: Fixed native file search, ripgrep, and subagent functionality
828- Added support for @-mentions in slash command arguments
829
830## 1.0.69
831
832- Upgraded Opus to version 4.1
833
834## 1.0.68
835
836- Fix incorrect model names being used for certain commands like `/pr-comments`
837- 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.
838- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
839- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
840- SDK: Added canUseTool callback support for tool confirmation
841- Added `disableAllHooks` setting
842- Improved file suggestions performance in large repos
843
844## 1.0.65
845
846- IDE: Fixed connection stability issues and error handling for diagnostics
847- Windows: Fixed shell environment setup for users without .bashrc files
848
849## 1.0.64
850
851- Agents: Added model customization support - you can now specify which model an agent should use
852- Agents: Fixed unintended access to the recursive agent tool
853- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
854- SDK: Fixed user input tracking across multi-turn conversations
855- Added hidden files to file search and @-mention suggestions
856
857## 1.0.63
858
859- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
860
861## 1.0.62
862
863- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
864- Hooks: Added SessionStart hook for new session initialization
865- /add-dir command now supports typeahead for directory paths
866- Improved network connectivity check reliability
867
868## 1.0.61
869
870- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
871- Settings: Added `--settings` flag to load settings from a JSON file
872- Settings: Fixed resolution of settings files paths that are symlinks
873- OTEL: Fixed reporting of wrong organization after authentication changes
874- Slash commands: Fixed permissions checking for allowed-tools with Bash
875- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
876- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
877- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
878
879## 1.0.60
880
881- You can now create custom subagents for specialized tasks! Run /agents to get started
882
883## 1.0.59
884
885- SDK: Added tool confirmation support with canUseTool callback
886- SDK: Allow specifying env for spawned process
887- Hooks: Exposed PermissionDecision to hooks (including "ask")
888- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
889- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
890
891## 1.0.58
892
893- Added support for reading PDFs
894- MCP: Improved server health status display in 'claude mcp list'
895- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
896
897## 1.0.57
898
899- Added support for specifying a model in slash commands
900- Improved permission messages to help Claude understand allowed tools
901- Fix: Remove trailing newlines from bash output in terminal wrapping
902
903## 1.0.56
904
905- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
906- Fixes for WSL IDE detection
907- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
908
909## 1.0.55
910
911- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
912- Windows: fixed Ctrl+Z crash
913- SDK: Added ability to capture error logging
914- Add --system-prompt-file option to override system prompt in print mode
915
916## 1.0.54
917
918- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
919- Custom slash commands: Added argument-hint to frontmatter
920- Windows: OAuth uses port 45454 and properly constructs browser URL
921- Windows: mode switching now uses alt + m, and plan mode renders properly
922- Shell: Switch to in-memory shell snapshot to fix file-related errors
923
924## 1.0.53
925
926- Updated @-mention file truncation from 100 lines to 2000 lines
927- 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).
928
929## 1.0.52
930
931- Added support for MCP server instructions
932
933## 1.0.51
934
935- Added support for native Windows (requires Git for Windows)
936- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
937- Settings: /doctor can now help you identify and fix invalid setting files
938- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
939- Increased auto-compact warning threshold from 60% to 80%
940- Fixed an issue with handling user directories with spaces for shell snapshots
941- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
942- Custom slash commands: Fixed user-level commands in subdirectories
943- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
944
945## 1.0.48
946
947- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
948- Added progress messages to Bash tool based on the last 5 lines of command output
949- Added expanding variables support for MCP server configuration
950- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
951- Improved IDE extension path handling when Claude Code runs in WSL
952- Hooks: Added a PreCompact hook
953- Vim mode: Added c, f/F, t/T
954
955## 1.0.45
956
957- Redesigned Search (Grep) tool with new tool input parameters and features
958- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
959- Fixed config file corruption issue by enforcing atomic writes
960- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
961- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
962- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
963
964## 1.0.44
965
966- New /export command lets you quickly export a conversation for sharing
967- MCP: resource_link tool results are now supported
968- MCP: tool annotations and tool titles now display in /mcp view
969- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
970
971## 1.0.43
972
973- Fixed a bug where the theme selector was saving excessively
974- Hooks: Added EPIPE system error handling
975
976## 1.0.42
977
978- Added tilde (`~`) expansion support to `/add-dir` command
979
980## 1.0.41
981
982- Hooks: Split Stop hook triggering into Stop and SubagentStop
983- Hooks: Enabled optional timeout configuration for each command
984- Hooks: Added "hook_event_name" to hook input
985- Fixed a bug where MCP tools would display twice in tool list
986- New tool parameters JSON for Bash tool in `tool_decision` event
987
988## 1.0.40
989
990- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
991
992## 1.0.39
993
994- New Active Time metric in OpenTelemetry logging
995
996## 1.0.38
997
998- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
999
1000## 1.0.37
1001
1002- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1003
1004## 1.0.36
1005
1006- Web search now takes today's date into context
1007- Fixed a bug where stdio MCP servers were not terminating properly on exit
1008
1009## 1.0.35
1010
1011- Added support for MCP OAuth Authorization Server discovery
1012
1013## 1.0.34
1014
1015- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1016
1017## 1.0.33
1018
1019- Improved logging functionality with session ID support
1020- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1021- Improvements to plan mode
1022
1023## 1.0.32
1024
1025- Updated loopback config for litellm
1026- Added forceLoginMethod setting to bypass login selection screen
1027
1028## 1.0.31
1029
1030- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1031
1032## 1.0.30
1033
1034- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1035- Improved file path autocomplete with filename matching
1036- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1037- Enhanced jq regex support for complex filters with pipes and select
1038
1039## 1.0.29
1040
1041- Improved CJK character support in cursor navigation and rendering
1042
1043## 1.0.28
1044
1045- Slash commands: Fix selector display during history navigation
1046- Resizes images before upload to prevent API size limit errors
1047- Added XDG_CONFIG_HOME support to configuration directory
1048- Performance optimizations for memory usage
1049- New attributes (terminal.type, language) in OpenTelemetry logging
1050
1051## 1.0.27
1052
1053- Streamable HTTP MCP servers are now supported
1054- Remote MCP servers (SSE and HTTP) now support OAuth
1055- MCP resources can now be @-mentioned
1056- /resume slash command to switch conversations within Claude Code
1057
1058## 1.0.25
1059
1060- Slash commands: moved "project" and "user" prefixes to descriptions
1061- Slash commands: improved reliability for command discovery
1062- Improved support for Ghostty
1063- Improved web search reliability
1064
1065## 1.0.24
1066
1067- Improved /mcp output
1068- Fixed a bug where settings arrays got overwritten instead of merged
1069
1070## 1.0.23
1071
1072- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1073- Released Python SDK: pip install claude-code-sdk to get started
1074
1075## 1.0.22
1076
1077- SDK: Renamed `total_cost` to `total_cost_usd`
1078
1079## 1.0.21
1080
1081- Improved editing of files with tab-based indentation
1082- Fix for tool_use without matching tool_result errors
1083- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1084
1085## 1.0.18
1086
1087- Added --add-dir CLI argument for specifying additional working directories
1088- Added streaming input support without require -p flag
1089- Improved startup performance and session storage performance
1090- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1091- Added detailed MCP server tools display (/mcp)
1092- MCP authentication and permission improvements
1093- Added auto-reconnection for MCP SSE connections on disconnect
1094- Fixed issue where pasted content was lost when dialogs appeared
1095
1096## 1.0.17
1097
1098- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1099- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1100- MCP server list UI improvements
1101- Update Claude Code process title to display "claude" instead of "node"
1102
1103## 1.0.11
1104
1105- Claude Code can now also be used with a Claude Pro subscription
1106- Added /upgrade for smoother switching to Claude Max plans
1107- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1108- Improved shell configuration error handling
1109- Improved todo list handling during compaction
1110
1111## 1.0.10
1112
1113- Added markdown table support
1114- Improved streaming performance
1115
1116## 1.0.8
1117
1118- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1119- Increased default otel interval from 1s -> 5s
1120- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1121- Fixed a regression where search tools unnecessarily asked for permissions
1122- Added support for triggering thinking non-English languages
1123- Improved compacting UI
1124
1125## 1.0.7
1126
1127- Renamed /allowed-tools -> /permissions
1128- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1129- Deprecated claude config commands in favor of editing settings.json
1130- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1131- Improved error handling for /install-github-app
1132- Bugfixes, UI polish, and tool reliability improvements
1133
1134## 1.0.6
1135
1136- Improved edit reliability for tab-indented files
1137- Respect CLAUDE_CONFIG_DIR everywhere
1138- Reduced unnecessary tool permission prompts
1139- Added support for symlinks in @file typeahead
1140- Bugfixes, UI polish, and tool reliability improvements
1141
1142## 1.0.4
1143
1144- Fixed a bug where MCP tool errors weren't being parsed correctly
1145
1146## 1.0.1
1147
1148- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1149- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1150- Updated documentation links and OAuth process descriptions
1151
1152## 1.0.0
1153
1154- Claude Code is now generally available
1155- Introducing Sonnet 4 and Opus 4 models
1156
1157## 0.2.125
1158
1159- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1160- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1161
1162## 0.2.117
1163
1164- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1165- Introduced settings.cleanupPeriodDays
1166- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1167- Introduced --debug mode
1168
1169## 0.2.108
1170
1171- You can now send messages to Claude while it works to steer Claude in real-time
1172- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1173- Fixed a bug where thinking was not working in -p mode
1174- Fixed a regression in /cost reporting
1175- Deprecated MCP wizard interface in favor of other MCP commands
1176- Lots of other bugfixes and improvements
1177
1178## 0.2.107
1179
1180- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1181
1182## 0.2.106
1183
1184- MCP SSE server configs can now specify custom headers
1185- Fixed a bug where MCP permission prompt didn't always show correctly
1186
1187## 0.2.105
1188
1189- Claude can now search the web
1190- Moved system & account status to /status
1191- Added word movement keybindings for Vim
1192- Improved latency for startup, todo tool, and file edits
1193
1194## 0.2.102
1195
1196- Improved thinking triggering reliability
1197- Improved @mention reliability for images and folders
1198- You can now paste multiple large chunks into one prompt
1199
1200## 0.2.100
1201
1202- Fixed a crash caused by a stack overflow error
1203- Made db storage optional; missing db support disables --continue and --resume
1204
1205## 0.2.98
1206
1207- Fixed an issue where auto-compact was running twice
1208
1209## 0.2.96
1210
1211- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1212
1213## 0.2.93
1214
1215- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1216- Claude now has access to a Todo list that helps it stay on track and be more organized
1217
1218## 0.2.82
1219
1220- Added support for --disallowedTools
1221- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1222
1223## 0.2.75
1224
1225- Hit Enter to queue up additional messages while Claude is working
1226- Drag in or copy/paste image files directly into the prompt
1227- @-mention files to directly add them to context
1228- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1229- Improved performance for filename auto-complete
1230
1231## 0.2.74
1232
1233- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1234- Task tool can now perform writes and run bash commands
1235
1236## 0.2.72
1237
1238- Updated spinner to indicate tokens loaded and tool usage
1239
1240## 0.2.70
1241
1242- Network commands like curl are now available for Claude to use
1243- Claude can now run multiple web queries in parallel
1244- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1245
1246## 0.2.69
1247
1248- Fixed UI glitches with improved Select component behavior
1249- Enhanced terminal output display with better text truncation logic
1250
1251## 0.2.67
1252
1253- Shared project permission rules can be saved in .claude/settings.json
1254
1255## 0.2.66
1256
1257- Print mode (-p) now supports streaming output via --output-format=stream-json
1258- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1259
1260## 0.2.63
1261
1262- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1263
1264## 0.2.61
1265
1266- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1267- Enhanced image detection for more reliable clipboard paste functionality
1268- Fixed an issue where ESC key could crash the conversation history selector
1269
1270## 0.2.59
1271
1272- Copy+paste images directly into your prompt
1273- Improved progress indicators for bash and fetch tools
1274- Bugfixes for non-interactive mode (-p)
1275
1276## 0.2.54
1277
1278- Quickly add to Memory by starting your message with '#'
1279- Press ctrl+r to see full output for long tool results
1280- Added support for MCP SSE transport
1281
1282## 0.2.53
1283
1284- New web fetch tool lets Claude view URLs that you paste in
1285- Fixed a bug with JPEG detection
1286
1287## 0.2.50
1288
1289- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1290
1291## 0.2.49
1292
1293- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1294
1295## 0.2.47
1296
1297- Press Tab to auto-complete file and folder names
1298- Press Shift + Tab to toggle auto-accept for file edits
1299- Automatic conversation compaction for infinite conversation length (toggle with /config)
1300
1301## 0.2.44
1302
1303- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1304
1305## 0.2.41
1306
1307- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1308- MCP server startup no longer blocks the app from starting up
1309
1310## 0.2.37
1311
1312- New /release-notes command lets you view release notes at any time
1313- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1314
1315## 0.2.36
1316
1317- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1318- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1319
1320## 0.2.34
1321
1322- Vim bindings for text input - enable with /vim or /config
1323
1324## 0.2.32
1325
1326- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1327- Fix for some PersistentShell issues
1328
1329## 0.2.31
1330
1331- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1332- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1333
1334## 0.2.30
1335
1336- Added ANSI color theme for better terminal compatibility
1337- Fixed issue where slash command arguments weren't being sent properly
1338- (Mac-only) API keys are now stored in macOS Keychain
1339
1340## 0.2.26
1341
1342- New /approved-tools command for managing tool permissions
1343- Word-level diff display for improved code readability
1344- Fuzzy matching for slash commands
1345
1346## 0.2.21
1347
1348- Fuzzy matching for /commands