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 052a131 · on Jan 21, 2026 History
1,450 lines · 72.2 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.15
4
5- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
6- Improved UI rendering performance with React Compiler
7- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
8- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
9
10## 2.1.14
11
12- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
13- Added search to installed plugins list - type to filter by name or description
14- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
15- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
16- Fixed memory issues that could cause crashes when running parallel subagents
17- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
18- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
19- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
20- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
21- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
22- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
23- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
24- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
25- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
26- Improved backspace to delete pasted text as a single token instead of one character at a time
27- [VSCode] Added `/usage` command to display current plan usage
28
29## 2.1.12
30
31- Fixed message rendering bug
32
33## 2.1.11
34
35- Fixed excessive MCP connection requests for HTTP/SSE transports
36
37## 2.1.10
38
39- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
40- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
41- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
42- Improved startup to capture keystrokes typed before the REPL is fully ready
43- Improved file suggestions to show as removable attachments instead of inserting text when accepted
44- [VSCode] Added install count display to plugin listings
45- [VSCode] Added trust warning when installing plugins
46
47## 2.1.9
48
49- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
50- Added `plansDirectory` setting to customize where plan files are stored
51- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
52- Added session URL attribution to commits and PRs created from web sessions
53- Added support for `PreToolUse` hooks to return `additionalContext` to the model
54- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
55- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
56- Fixed MCP server reconnection hanging when cached connection promise never resolves
57- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
58
59## 2.1.7
60
61- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
62- Added ability to provide feedback when accepting permission prompts
63- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
64- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
65- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
66- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
67- Fixed context window blocking limit being calculated using the full context window instead of the effective context window (which reserves space for max output tokens)
68- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
69- Fixed terminal title animation jitter by using fixed-width braille characters
70- Fixed plugins with git submodules not being fully initialized when installed
71- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
72- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
73- Enabled MCP tool search auto mode by default for all users. When MCP tool descriptions exceed 10% of the context window, they are automatically deferred and discovered via the MCPSearch tool instead of being loaded upfront. This reduces context usage for users with many MCP tools configured. Users can disable this by adding `MCPSearch` to `disallowedTools` in their settings.
74- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
75- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
76
77## 2.1.6
78
79- Added search functionality to `/config` command for quickly filtering settings
80- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
81- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
82- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
83- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
84- Added an error display when the editor fails during Ctrl+G
85- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
86- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
87- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
88- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
89- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
90- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
91- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
92- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
93- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
94- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
95- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
96- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
97- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
98- Improved @ autocomplete with icons for different suggestion types and single-line formatting
99- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
100- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
101- Changed terminal title to "Claude Code" on startup for better window identification
102- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
103- [VSCode] Fixed usage indicator not updating after manual compact
104
105## 2.1.5
106
107- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
108
109## 2.1.4
110
111- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
112- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
113
114## 2.1.3
115
116- Merged slash commands and skills, simplifying the mental model with no change in behavior
117- Added release channel (`stable` or `latest`) toggle to `/config`
118- 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
119- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
120- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
121- Fixed mismatch between background task count in status bar and items shown in tasks dialog
122- Fixed sub-agents using the wrong model during conversation compaction
123- Fixed web search in sub-agents using incorrect model
124- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
125- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
126- Improved slash command suggestion readability by truncating long descriptions to 2 lines
127- Changed tool hook execution timeout from 60 seconds to 10 minutes
128- [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)
129
130## 2.1.2
131
132- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
133- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
134- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
135- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
136- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
137- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
138- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
139- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
140- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
141- Fixed updates incorrectly claiming another installation is in progress
142- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
143- Fixed remote session URL and teleport being broken when using `/tasks` command
144- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
145- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
146- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
147- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
148- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
149- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
150- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
151- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
152- [SDK] Changed minimum zod peer dependency to ^4.0.0
153- [VSCode] Fixed usage display not updating after manual compact
154
155## 2.1.0
156
157- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
158- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
159- Added support for `agent` field in skills to specify agent type for execution
160- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
161- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
162- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
163- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
164- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
165- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
166- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
167- Fixed Esc key with queued prompts to only move them to input without canceling the running task
168- Reduced permission prompts for complex bash commands
169- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
170- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
171- Fixed plugin path resolution for file-based marketplace sources
172- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
173- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
174- Fixed Claude in Chrome support for WSL environments
175- Fixed Windows native installer silently failing when executable creation fails
176- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
177- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
178- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
179- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
180- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
181- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
182- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
183- Added hooks support for skill and slash command frontmatter
184- 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
185- Added `/plan` command shortcut to enable plan mode directly from the prompt
186- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
187- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
188- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
189- Added support for `once: true` config for hooks
190- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
191- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
192- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
193- Added left/right arrow key navigation for cycling through tabs in dialogs
194- Added real-time thinking block display in Ctrl+O transcript mode
195- Added filepath to full output in background bash task details dialog
196- Added Skills as a separate category in the context visualization
197- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
198- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
199- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
200- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
201- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
202- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
203- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
204- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
205- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
206- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
207- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
208- Fixed commands with `$()` command substitution failing with parse errors
209- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
210- 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)
211- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
212- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
213- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
214- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
215- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
216- Fixed subagents sometimes not inheriting the parent's model by default
217- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
218- Fixed duplicate Bash commands appearing in permission request option labels
219- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
220- Fixed background task completion notifications to appear proactively with bullet point
221- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
222- Fixed cursor disappearing after dismissing permission dialogs
223- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
224- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
225- Fixed images being silently dropped when queueing messages while backgrounding a task
226- Fixed large pasted images failing with "Image was too large" error
227- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
228- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
229- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
230- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
231- Fixed spinner token counter to properly accumulate tokens from subagents during execution
232- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
233- Fixed race condition where LSP tool could return "no server available" during startup
234- Fixed feedback submission hanging indefinitely when network requests timeout
235- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
236- Fixed hook success message showing trailing colon when hook has no output
237- Multiple optimizations to improve startup performance
238- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
239- Improved performance when reading Jupyter notebooks with many cells
240- Improved reliability for piped input like `cat refactor.md | claude`
241- Improved reliability for AskQuestion tool
242- Improved sed in-place edit commands to render as file edits with diff preview
243- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
244- Improved compaction reliability
245- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
246- Improved skills to show progress while executing, displaying tool uses as they happen
247- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
248- Improved skill suggestions to prioritize recently and frequently used skills
249- Improved spinner feedback when waiting for the first response token
250- Improved token count display in spinner to include tokens from background agents
251- Improved incremental output for async agents to give the main thread more control and visibility
252- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
253- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
254- Improved macOS screenshot paste reliability with TIFF format support
255- Improved `/stats` output
256- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
257- Changed "Interrupted" message color from red to grey for a less alarming appearance
258- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
259- Removed underline styling from image reference links
260- [SDK] Changed minimum zod peer dependency to ^4.0.0
261- [VSCode] Added currently selected model name to the context menu
262- [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")
263- [VSCode] Fixed paragraph breaks not rendering in markdown content
264- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
265- [Windows] Fixed issue with improper rendering
266
267## 2.0.76
268
269- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
270
271## 2.0.75
272
273- Minor bugfixes
274
275## 2.0.74
276
277- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
278- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
279- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
280- Added syntax highlighting info to theme picker
281- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
282- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
283- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
284- Fixed a potential crash when syntax highlighting isn't initialized correctly
285- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
286- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
287- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
288- [Windows] Fixed issue with improper rendering
289- [VSCode] Added gift tag pictogram for year-end promotion message
290
291## 2.0.73
292
293- Added clickable `[Image #N]` links that open attached images in the default viewer
294- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
295- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
296- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
297- Fixed slow input history cycling and race condition that could overwrite text after message submission
298- Improved `/theme` command to open theme picker directly
299- Improved theme picker UI
300- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
301- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
302
303## 2.0.72
304
305- 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
306- Reduced terminal flickering
307- Added scannable QR code to mobile app tip for quick app downloads
308- Added loading indicator when resuming conversations for better feedback
309- Fixed `/context` command not respecting custom system prompts in non-interactive mode
310- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
311- Improved @ mention file suggestion speed (~3x faster in git repositories)
312- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
313- Improved settings validation errors to be more prominent
314- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
315
316## 2.0.71
317
318- Added /config toggle to enable/disable prompt suggestions
319- Added `/settings` as an alias for the `/config` command
320- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
321- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
322- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
323- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
324- New syntax highlighting engine for native build
325
326## 2.0.70
327
328- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
329- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
330- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
331- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
332- Fixed input being cleared when processing queued commands while the user was typing
333- Fixed prompt suggestions replacing typed input when pressing Tab
334- Fixed diff view not updating when terminal is resized
335- Improved memory usage by 3x for large conversations
336- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
337- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
338- Fix thinking mode toggle in /config not persisting correctly
339- Improve UI for file creation permission dialog
340
341## 2.0.69
342
343- Minor bugfixes
344
345## 2.0.68
346
347- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
348- Fixed a bug where disallowed MCP tools were visible to the model
349- Fixed an issue where steering messages could be lost while a subagent is working
350- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
351- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
352- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
353
354## 2.0.67
355
356- Thinking mode is now enabled by default for Opus 4.5
357- Thinking mode configuration has moved to /config
358- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
359- Show reason why autoupdater is disabled in `/doctor`
360- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
361- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
362- Fixed scroll position resetting after deleting a permission rule in `/permissions`
363- 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
364- Fixed `claude install --force` not bypassing stale lock files
365- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
366- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
367
368## 2.0.65
369
370- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
371- Added context window information to status line input
372- Added `fileSuggestion` setting for custom `@` file search commands
373- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
374- Fixed prompt not being saved to history when aborting a query with Escape
375- Fixed Read tool image handling to identify format from bytes instead of file extension
376
377## 2.0.64
378
379- Made auto-compacting instant
380- Agents and bash commands can run asynchronously and send messages to wake up the main agent
381- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
382- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
383- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
384- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
385- Fixed auto-loading .env when using native installer
386- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
387- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
388- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
389- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
390- Bedrock: Improve efficiency of token counting
391- Bedrock: Add support for `aws login` AWS Management Console credentials
392- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
393
394## 2.0.62
395
396- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
397- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
398- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
399- Fixed slash command selection not working when multiple commands share the same name
400- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
401- Fixed running versions getting removed because lock file incorrectly going stale
402- Fixed IDE diff tab not closing when rejecting file changes
403
404## 2.0.61
405
406- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
407
408## 2.0.60
409
410- Added background agent support. Agents run in the background while you work
411- Added --disable-slash-commands CLI flag to disable all slash commands
412- Added model name to "Co-Authored-By" commit messages
413- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
414- Updated Fetch to skip summarization for pre-approved websites
415- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
416
417## 2.0.59
418
419- Added --agent CLI flag to override the agent setting for the current session
420- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
421- VS Code: Fixed .claude.json config file being read from incorrect location
422
423## 2.0.58
424
425- Pro users now have access to Opus 4.5 as part of their subscription!
426- Fixed timer duration showing "11m 60s" instead of "12m 0s"
427- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
428
429## 2.0.57
430
431- Added feedback input when rejecting plans, allowing users to tell Claude what to change
432- VSCode: Added streaming message support for real-time response display
433
434## 2.0.56
435
436- Added setting to enable/disable terminal progress bar (OSC 9;4)
437- 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.
438
439## 2.0.55
440
441- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
442- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
443- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
444- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
445
446## 2.0.54
447
448- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
449- Fix issue with excessive iTerm notifications
450
451## 2.0.52
452
453- Fixed duplicate message display when starting Claude with a command line argument
454- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
455- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
456- Permit some uses of `$!` in bash commands
457
458## 2.0.51
459
460- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
461- Introducing Claude Code for Desktop: https://claude.com/download
462- 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
463- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
464- Plan Mode now builds more precise plans and executes more thoroughly
465- Usage limit notifications now easier to understand
466- Switched `/usage` back to "% used"
467- Fixed handling of thinking errors
468- Fixed performance regression
469
470## 2.0.50
471
472- Fixed bug preventing calling MCP tools that have nested references in their input schemas
473- Silenced a noisy but harmless error during upgrades
474- Improved ultrathink text display
475- Improved clarity of 5-hour session limit warning message
476
477## 2.0.49
478
479- Added readline-style ctrl-y for pasting deleted text
480- Improved clarity of usage limit warning message
481- Fixed handling of subagent permissions
482
483## 2.0.47
484
485- Improved error messages and validation for `claude --teleport`
486- Improved error handling in `/usage`
487- Fixed race condition with history entry not getting logged at exit
488- Fixed Vertex AI configuration not being applied from `settings.json`
489
490## 2.0.46
491
492- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
493
494## 2.0.45
495
496- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
497- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
498- Send background tasks to Claude Code on the web by starting a message with `&`
499
500## 2.0.43
501
502- Added `permissionMode` field for custom agents
503- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
504- Added skills frontmatter field to declare skills to auto-load for subagents
505- Added the `SubagentStart` hook event
506- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
507- Fixed duplicate rendering of some messages in the UI
508- Fixed some visual flickers
509- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
510
511## 2.0.42
512
513- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
514
515## 2.0.41
516
517- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
518- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
519- Fixed incorrect labeling of user settings vs project settings in command descriptions
520- Fixed crash when plugin command hooks timeout during execution
521- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
522- Fixed broken security documentation links in trust dialogs and onboarding
523- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
524- ctrl-r history search landing on a slash command no longer cancels the search
525- SDK: Support custom timeouts for hooks
526- Allow more safe git commands to run without approval
527- Plugins: Added support for sharing and installing output styles
528- Teleporting a session from web will automatically set the upstream branch
529
530## 2.0.37
531
532- Fixed how idleness is computed for notifications
533- Hooks: Added matcher values for Notification hook events
534- Output Styles: Added `keep-coding-instructions` option to frontmatter
535
536## 2.0.36
537
538- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
539- Fixed queued messages being incorrectly executed as bash commands
540- Fixed input being lost when typing while a queued message is processed
541
542## 2.0.35
543
544- Improve fuzzy search results when searching commands
545- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
546- 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
547- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
548- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
549
550## 2.0.34
551
552- VSCode Extension: Added setting to configure the initial permission mode for new conversations
553- Improved file path suggestion performance with native Rust-based fuzzy finder
554- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
555- Fixed memory crash when reading or writing large files (especially base64-encoded images)
556
557## 2.0.33
558
559- Native binary installs now launch quicker.
560- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
561- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
562
563## 2.0.32
564
565- Un-deprecate output styles based on community feedback
566- Added `companyAnnouncements` setting for displaying announcements on startup
567- Fixed hook progress messages not updating correctly during PostToolUse hook execution
568
569## 2.0.31
570
571- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
572- Vertex: add support for Web Search on supported models
573- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
574- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
575- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
576- Fixed plugin uninstall not removing plugins
577
578## 2.0.30
579
580- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
581- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
582- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
583- Added prompt-based stop hooks
584- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
585- Enabled SSE MCP servers on native build
586- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
587- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
588- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
589- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
590- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
591- Fixed bug that saved session permissions to local settings
592- Fixed MCP tools not being available to sub-agents
593- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
594- Fixed delay when navigating through typeahead suggestions with arrow keys
595- VSCode: Restored selection indicator in input footer showing current file or code selection status
596
597## 2.0.28
598
599- Plan mode: introduced new Plan subagent
600- Subagents: claude can now choose to resume subagents
601- Subagents: claude can dynamically choose the model used by its subagents
602- SDK: added --max-budget-usd flag
603- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
604- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
605- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
606- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
607- Various other bug fixes
608
609## 2.0.27
610
611- New UI for permission prompts
612- Added current branch filtering and search to session resume screen for easier navigation
613- Fixed directory @-mention causing "No assistant message found" error
614- VSCode Extension: Add config setting to include .gitignored files in file searches
615- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
616
617## 2.0.25
618
619- 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
620
621## 2.0.24
622
623- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
624- Claude Code Web: Support for Web -> CLI teleport
625- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
626- Bedrock: Display awsAuthRefresh output when auth is required
627
628## 2.0.22
629
630- Fixed content layout shift when scrolling through slash commands
631- IDE: Add toggle to enable/disable thinking.
632- Fix bug causing duplicate permission prompts with parallel tool calls
633- Add support for enterprise managed MCP allowlist and denylist
634
635## 2.0.21
636
637- Support MCP `structuredContent` field in tool responses
638- Added an interactive question tool
639- Claude will now ask you questions more often in plan mode
640- Added Haiku 4.5 as a model option for Pro users
641- Fixed an issue where queued commands don't have access to previous messages' output
642
643## 2.0.20
644
645- Added support for Claude Skills
646
647## 2.0.19
648
649- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
650- Fixed a bug where Haiku was unnecessarily called in print mode
651
652## 2.0.17
653
654- Added Haiku 4.5 to model selector!
655- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
656- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
657- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
658- OTEL: support HTTP_PROXY and HTTPS_PROXY
659- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
660
661## 2.0.15
662
663- Fixed bug with resuming where previously created files needed to be read again before writing
664- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
665
666## 2.0.14
667
668- Fix @-mentioning MCP servers to toggle them on/off
669- Improve permission checks for bash with inline env vars
670- Fix ultrathink + thinking toggle
671- Reduce unnecessary logins
672- Document --system-prompt
673- Several improvements to rendering
674- Plugins UI polish
675
676## 2.0.13
677
678- Fixed `/plugin` not working on native build
679
680## 2.0.12
681
682- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
683- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
684- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
685- `/plugin validate` command for validating plugin structure and configuration
686- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
687- Plugin documentation available at https://code.claude.com/docs/en/plugins
688- Comprehensive error messages and diagnostics via `/doctor` command
689- Avoid flickering in `/model` selector
690- Improvements to `/help`
691- Avoid mentioning hooks in `/resume` summaries
692- Changes to the "verbose" setting in `/config` now persist across sessions
693
694## 2.0.11
695
696- Reduced system prompt size by 1.4k tokens
697- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
698- Fixed Opus fallback rate limit errors appearing incorrectly
699- Fixed /add-dir command selecting wrong default tab
700
701## 2.0.10
702
703- Rewrote terminal renderer for buttery smooth UI
704- Enable/disable MCP servers by @mentioning, or in /mcp
705- Added tab completion for shell commands in bash mode
706- PreToolUse hooks can now modify tool inputs
707- Press Ctrl-G to edit your prompt in your system's configured text editor
708- Fixes for bash permission checks with environment variables in the command
709
710## 2.0.9
711
712- Fix regression where bash backgrounding stopped working
713
714## 2.0.8
715
716- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
717- IDE: Add drag-and-drop support for files and folders in chat
718- /context: Fix counting for thinking blocks
719- Improve message rendering for users with light themes on dark terminals
720- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
721
722## 2.0.5
723
724- IDE: Fix IME unintended message submission with Enter and Tab
725- IDE: Add "Open in Terminal" link in login screen
726- Fix unhandled OAuth expiration 401 API errors
727- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
728
729## 2.0.1
730
731- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
732- Various bug fixes and presentation improvements
733
734## 2.0.0
735
736- New native VS Code extension
737- Fresh coat of paint throughout the whole app
738- /rewind a conversation to undo code changes
739- /usage command to see plan limits
740- Tab to toggle thinking (sticky across sessions)
741- Ctrl-R to search history
742- Unshipped claude config command
743- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
744- SDK: The Claude Code SDK is now the Claude Agent SDK
745- Add subagents dynamically with `--agents` flag
746
747## 1.0.126
748
749- Enable /context command for Bedrock and Vertex
750- Add mTLS support for HTTP-based OpenTelemetry exporters
751
752## 1.0.124
753
754- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
755- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
756- No longer inform Claude of the list of allowed tools when permission is denied
757- Fixed security vulnerability in Bash tool permission checks
758- Improved VSCode extension performance for large files
759
760## 1.0.123
761
762- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
763- Fixed thinking mode triggering on negation phrases like "don't think"
764- Fixed rendering performance degradation during token streaming
765- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
766- Enhanced BashTool environment snapshot logging
767- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
768- Migrated --debug logging to a file, to enable easy tailing & filtering
769
770## 1.0.120
771
772- Fix input lag during typing, especially noticeable with large prompts
773- Improved VSCode extension command registry and sessions dialog user experience
774- Enhanced sessions dialog responsiveness and visual feedback
775- Fixed IDE compatibility issue by removing worktree support check
776- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
777
778## 1.0.119
779
780- Fix Windows issue where process visually freezes on entering interactive mode
781- Support dynamic headers for MCP servers via headersHelper configuration
782- Fix thinking mode not working in headless sessions
783- Fix slash commands now properly update allowed tools instead of replacing them
784
785## 1.0.117
786
787- Add Ctrl-R history search to recall previous commands like bash/zsh
788- Fix input lag while typing, especially on Windows
789- Add sed command to auto-allowed commands in acceptEdits mode
790- Fix Windows PATH comparison to be case-insensitive for drive letters
791- Add permissions management hint to /add-dir output
792
793## 1.0.115
794
795- Improve thinking mode display with enhanced visual effects
796- Type /t to temporarily disable thinking mode in your prompt
797- Improve path validation for glob and grep tools
798- Show condensed output for post-tool hooks to reduce visual clutter
799- Fix visual feedback when loading state completes
800- Improve UI consistency for permission request dialogs
801
802## 1.0.113
803
804- Deprecated piped input in interactive mode
805- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
806
807## 1.0.112
808
809- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
810- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
811- Hooks: Added systemMessage support for SessionEnd hooks
812- Added `spinnerTipsEnabled` setting to disable spinner tips
813- IDE: Various improvements and bug fixes
814
815## 1.0.111
816
817- /model now validates provided model names
818- Fixed Bash tool crashes caused by malformed shell syntax parsing
819
820## 1.0.110
821
822- /terminal-setup command now supports WezTerm
823- MCP: OAuth tokens now proactively refresh before expiration
824- Fixed reliability issues with background Bash processes
825
826## 1.0.109
827
828- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
829
830## 1.0.106
831
832- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
833
834## 1.0.97
835
836- Settings: /doctor now validates permission rule syntax and suggests corrections
837
838## 1.0.94
839
840- Vertex: add support for global endpoints for supported models
841- /memory command now allows direct editing of all imported memory files
842- SDK: Add custom tools as callbacks
843- Added /todos command to list current todo items
844
845## 1.0.93
846
847- Windows: Add alt + v shortcut for pasting images from clipboard
848- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
849
850## 1.0.90
851
852- Settings file changes take effect immediately - no restart required
853
854## 1.0.88
855
856- Fixed issue causing "OAuth authentication is currently not supported"
857- Status line input now includes `exceeds_200k_tokens`
858- Fixed incorrect usage tracking in /cost.
859- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
860- Bedrock: Updated default Sonnet model to Sonnet 4
861
862## 1.0.86
863
864- Added /context to help users self-serve debug context issues
865- SDK: Added UUID support for all SDK messages
866- SDK: Added `--replay-user-messages` to replay user messages back to stdout
867
868## 1.0.85
869
870- Status line input now includes session cost info
871- Hooks: Introduced SessionEnd hook
872
873## 1.0.84
874
875- Fix tool_use/tool_result id mismatch error when network is unstable
876- Fix Claude sometimes ignoring real-time steering when wrapping up a task
877- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
878- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
879
880## 1.0.83
881
882- @-mention: Support files with spaces in path
883- New shimmering spinner
884
885## 1.0.82
886
887- SDK: Add request cancellation support
888- SDK: New additionalDirectories option to search custom paths, improved slash command processing
889- Settings: Validation prevents invalid fields in .claude/settings.json files
890- MCP: Improve tool name consistency
891- Bash: Fix crash when Claude tries to automatically read large files
892
893## 1.0.81
894
895- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
896- Agents: Fix custom agent loading when agent files are unparsable
897
898## 1.0.80
899
900- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
901
902## 1.0.77
903
904- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
905- SDK: Add session support and permission denial tracking
906- Fix token limit errors in conversation summarization
907- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
908
909## 1.0.73
910
911- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
912- MCP: Press Esc to cancel OAuth authentication flows
913- Bash: Improved command validation and reduced false security warnings
914- UI: Enhanced spinner animations and status line visual hierarchy
915- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
916
917## 1.0.72
918
919- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
920
921## 1.0.71
922
923- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
924- Customizable status line: add your terminal prompt to Claude Code with /statusline
925
926## 1.0.70
927
928- Performance: Optimized message rendering for better performance with large contexts
929- Windows: Fixed native file search, ripgrep, and subagent functionality
930- Added support for @-mentions in slash command arguments
931
932## 1.0.69
933
934- Upgraded Opus to version 4.1
935
936## 1.0.68
937
938- Fix incorrect model names being used for certain commands like `/pr-comments`
939- 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.
940- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
941- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
942- SDK: Added canUseTool callback support for tool confirmation
943- Added `disableAllHooks` setting
944- Improved file suggestions performance in large repos
945
946## 1.0.65
947
948- IDE: Fixed connection stability issues and error handling for diagnostics
949- Windows: Fixed shell environment setup for users without .bashrc files
950
951## 1.0.64
952
953- Agents: Added model customization support - you can now specify which model an agent should use
954- Agents: Fixed unintended access to the recursive agent tool
955- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
956- SDK: Fixed user input tracking across multi-turn conversations
957- Added hidden files to file search and @-mention suggestions
958
959## 1.0.63
960
961- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
962
963## 1.0.62
964
965- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
966- Hooks: Added SessionStart hook for new session initialization
967- /add-dir command now supports typeahead for directory paths
968- Improved network connectivity check reliability
969
970## 1.0.61
971
972- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
973- Settings: Added `--settings` flag to load settings from a JSON file
974- Settings: Fixed resolution of settings files paths that are symlinks
975- OTEL: Fixed reporting of wrong organization after authentication changes
976- Slash commands: Fixed permissions checking for allowed-tools with Bash
977- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
978- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
979- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
980
981## 1.0.60
982
983- You can now create custom subagents for specialized tasks! Run /agents to get started
984
985## 1.0.59
986
987- SDK: Added tool confirmation support with canUseTool callback
988- SDK: Allow specifying env for spawned process
989- Hooks: Exposed PermissionDecision to hooks (including "ask")
990- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
991- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
992
993## 1.0.58
994
995- Added support for reading PDFs
996- MCP: Improved server health status display in 'claude mcp list'
997- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
998
999## 1.0.57
1000
1001- Added support for specifying a model in slash commands
1002- Improved permission messages to help Claude understand allowed tools
1003- Fix: Remove trailing newlines from bash output in terminal wrapping
1004
1005## 1.0.56
1006
1007- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1008- Fixes for WSL IDE detection
1009- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1010
1011## 1.0.55
1012
1013- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1014- Windows: fixed Ctrl+Z crash
1015- SDK: Added ability to capture error logging
1016- Add --system-prompt-file option to override system prompt in print mode
1017
1018## 1.0.54
1019
1020- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1021- Custom slash commands: Added argument-hint to frontmatter
1022- Windows: OAuth uses port 45454 and properly constructs browser URL
1023- Windows: mode switching now uses alt + m, and plan mode renders properly
1024- Shell: Switch to in-memory shell snapshot to fix file-related errors
1025
1026## 1.0.53
1027
1028- Updated @-mention file truncation from 100 lines to 2000 lines
1029- 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).
1030
1031## 1.0.52
1032
1033- Added support for MCP server instructions
1034
1035## 1.0.51
1036
1037- Added support for native Windows (requires Git for Windows)
1038- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1039- Settings: /doctor can now help you identify and fix invalid setting files
1040- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1041- Increased auto-compact warning threshold from 60% to 80%
1042- Fixed an issue with handling user directories with spaces for shell snapshots
1043- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1044- Custom slash commands: Fixed user-level commands in subdirectories
1045- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1046
1047## 1.0.48
1048
1049- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1050- Added progress messages to Bash tool based on the last 5 lines of command output
1051- Added expanding variables support for MCP server configuration
1052- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1053- Improved IDE extension path handling when Claude Code runs in WSL
1054- Hooks: Added a PreCompact hook
1055- Vim mode: Added c, f/F, t/T
1056
1057## 1.0.45
1058
1059- Redesigned Search (Grep) tool with new tool input parameters and features
1060- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1061- Fixed config file corruption issue by enforcing atomic writes
1062- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1063- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1064- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1065
1066## 1.0.44
1067
1068- New /export command lets you quickly export a conversation for sharing
1069- MCP: resource_link tool results are now supported
1070- MCP: tool annotations and tool titles now display in /mcp view
1071- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1072
1073## 1.0.43
1074
1075- Fixed a bug where the theme selector was saving excessively
1076- Hooks: Added EPIPE system error handling
1077
1078## 1.0.42
1079
1080- Added tilde (`~`) expansion support to `/add-dir` command
1081
1082## 1.0.41
1083
1084- Hooks: Split Stop hook triggering into Stop and SubagentStop
1085- Hooks: Enabled optional timeout configuration for each command
1086- Hooks: Added "hook_event_name" to hook input
1087- Fixed a bug where MCP tools would display twice in tool list
1088- New tool parameters JSON for Bash tool in `tool_decision` event
1089
1090## 1.0.40
1091
1092- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1093
1094## 1.0.39
1095
1096- New Active Time metric in OpenTelemetry logging
1097
1098## 1.0.38
1099
1100- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1101
1102## 1.0.37
1103
1104- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1105
1106## 1.0.36
1107
1108- Web search now takes today's date into context
1109- Fixed a bug where stdio MCP servers were not terminating properly on exit
1110
1111## 1.0.35
1112
1113- Added support for MCP OAuth Authorization Server discovery
1114
1115## 1.0.34
1116
1117- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1118
1119## 1.0.33
1120
1121- Improved logging functionality with session ID support
1122- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1123- Improvements to plan mode
1124
1125## 1.0.32
1126
1127- Updated loopback config for litellm
1128- Added forceLoginMethod setting to bypass login selection screen
1129
1130## 1.0.31
1131
1132- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1133
1134## 1.0.30
1135
1136- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1137- Improved file path autocomplete with filename matching
1138- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1139- Enhanced jq regex support for complex filters with pipes and select
1140
1141## 1.0.29
1142
1143- Improved CJK character support in cursor navigation and rendering
1144
1145## 1.0.28
1146
1147- Slash commands: Fix selector display during history navigation
1148- Resizes images before upload to prevent API size limit errors
1149- Added XDG_CONFIG_HOME support to configuration directory
1150- Performance optimizations for memory usage
1151- New attributes (terminal.type, language) in OpenTelemetry logging
1152
1153## 1.0.27
1154
1155- Streamable HTTP MCP servers are now supported
1156- Remote MCP servers (SSE and HTTP) now support OAuth
1157- MCP resources can now be @-mentioned
1158- /resume slash command to switch conversations within Claude Code
1159
1160## 1.0.25
1161
1162- Slash commands: moved "project" and "user" prefixes to descriptions
1163- Slash commands: improved reliability for command discovery
1164- Improved support for Ghostty
1165- Improved web search reliability
1166
1167## 1.0.24
1168
1169- Improved /mcp output
1170- Fixed a bug where settings arrays got overwritten instead of merged
1171
1172## 1.0.23
1173
1174- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1175- Released Python SDK: pip install claude-code-sdk to get started
1176
1177## 1.0.22
1178
1179- SDK: Renamed `total_cost` to `total_cost_usd`
1180
1181## 1.0.21
1182
1183- Improved editing of files with tab-based indentation
1184- Fix for tool_use without matching tool_result errors
1185- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1186
1187## 1.0.18
1188
1189- Added --add-dir CLI argument for specifying additional working directories
1190- Added streaming input support without require -p flag
1191- Improved startup performance and session storage performance
1192- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1193- Added detailed MCP server tools display (/mcp)
1194- MCP authentication and permission improvements
1195- Added auto-reconnection for MCP SSE connections on disconnect
1196- Fixed issue where pasted content was lost when dialogs appeared
1197
1198## 1.0.17
1199
1200- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1201- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1202- MCP server list UI improvements
1203- Update Claude Code process title to display "claude" instead of "node"
1204
1205## 1.0.11
1206
1207- Claude Code can now also be used with a Claude Pro subscription
1208- Added /upgrade for smoother switching to Claude Max plans
1209- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1210- Improved shell configuration error handling
1211- Improved todo list handling during compaction
1212
1213## 1.0.10
1214
1215- Added markdown table support
1216- Improved streaming performance
1217
1218## 1.0.8
1219
1220- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1221- Increased default otel interval from 1s -> 5s
1222- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1223- Fixed a regression where search tools unnecessarily asked for permissions
1224- Added support for triggering thinking non-English languages
1225- Improved compacting UI
1226
1227## 1.0.7
1228
1229- Renamed /allowed-tools -> /permissions
1230- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1231- Deprecated claude config commands in favor of editing settings.json
1232- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1233- Improved error handling for /install-github-app
1234- Bugfixes, UI polish, and tool reliability improvements
1235
1236## 1.0.6
1237
1238- Improved edit reliability for tab-indented files
1239- Respect CLAUDE_CONFIG_DIR everywhere
1240- Reduced unnecessary tool permission prompts
1241- Added support for symlinks in @file typeahead
1242- Bugfixes, UI polish, and tool reliability improvements
1243
1244## 1.0.4
1245
1246- Fixed a bug where MCP tool errors weren't being parsed correctly
1247
1248## 1.0.1
1249
1250- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1251- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1252- Updated documentation links and OAuth process descriptions
1253
1254## 1.0.0
1255
1256- Claude Code is now generally available
1257- Introducing Sonnet 4 and Opus 4 models
1258
1259## 0.2.125
1260
1261- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1262- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1263
1264## 0.2.117
1265
1266- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1267- Introduced settings.cleanupPeriodDays
1268- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1269- Introduced --debug mode
1270
1271## 0.2.108
1272
1273- You can now send messages to Claude while it works to steer Claude in real-time
1274- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1275- Fixed a bug where thinking was not working in -p mode
1276- Fixed a regression in /cost reporting
1277- Deprecated MCP wizard interface in favor of other MCP commands
1278- Lots of other bugfixes and improvements
1279
1280## 0.2.107
1281
1282- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1283
1284## 0.2.106
1285
1286- MCP SSE server configs can now specify custom headers
1287- Fixed a bug where MCP permission prompt didn't always show correctly
1288
1289## 0.2.105
1290
1291- Claude can now search the web
1292- Moved system & account status to /status
1293- Added word movement keybindings for Vim
1294- Improved latency for startup, todo tool, and file edits
1295
1296## 0.2.102
1297
1298- Improved thinking triggering reliability
1299- Improved @mention reliability for images and folders
1300- You can now paste multiple large chunks into one prompt
1301
1302## 0.2.100
1303
1304- Fixed a crash caused by a stack overflow error
1305- Made db storage optional; missing db support disables --continue and --resume
1306
1307## 0.2.98
1308
1309- Fixed an issue where auto-compact was running twice
1310
1311## 0.2.96
1312
1313- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1314
1315## 0.2.93
1316
1317- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1318- Claude now has access to a Todo list that helps it stay on track and be more organized
1319
1320## 0.2.82
1321
1322- Added support for --disallowedTools
1323- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1324
1325## 0.2.75
1326
1327- Hit Enter to queue up additional messages while Claude is working
1328- Drag in or copy/paste image files directly into the prompt
1329- @-mention files to directly add them to context
1330- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1331- Improved performance for filename auto-complete
1332
1333## 0.2.74
1334
1335- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1336- Task tool can now perform writes and run bash commands
1337
1338## 0.2.72
1339
1340- Updated spinner to indicate tokens loaded and tool usage
1341
1342## 0.2.70
1343
1344- Network commands like curl are now available for Claude to use
1345- Claude can now run multiple web queries in parallel
1346- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1347
1348## 0.2.69
1349
1350- Fixed UI glitches with improved Select component behavior
1351- Enhanced terminal output display with better text truncation logic
1352
1353## 0.2.67
1354
1355- Shared project permission rules can be saved in .claude/settings.json
1356
1357## 0.2.66
1358
1359- Print mode (-p) now supports streaming output via --output-format=stream-json
1360- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1361
1362## 0.2.63
1363
1364- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1365
1366## 0.2.61
1367
1368- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1369- Enhanced image detection for more reliable clipboard paste functionality
1370- Fixed an issue where ESC key could crash the conversation history selector
1371
1372## 0.2.59
1373
1374- Copy+paste images directly into your prompt
1375- Improved progress indicators for bash and fetch tools
1376- Bugfixes for non-interactive mode (-p)
1377
1378## 0.2.54
1379
1380- Quickly add to Memory by starting your message with '#'
1381- Press ctrl+r to see full output for long tool results
1382- Added support for MCP SSE transport
1383
1384## 0.2.53
1385
1386- New web fetch tool lets Claude view URLs that you paste in
1387- Fixed a bug with JPEG detection
1388
1389## 0.2.50
1390
1391- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1392
1393## 0.2.49
1394
1395- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1396
1397## 0.2.47
1398
1399- Press Tab to auto-complete file and folder names
1400- Press Shift + Tab to toggle auto-accept for file edits
1401- Automatic conversation compaction for infinite conversation length (toggle with /config)
1402
1403## 0.2.44
1404
1405- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1406
1407## 0.2.41
1408
1409- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1410- MCP server startup no longer blocks the app from starting up
1411
1412## 0.2.37
1413
1414- New /release-notes command lets you view release notes at any time
1415- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1416
1417## 0.2.36
1418
1419- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1420- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1421
1422## 0.2.34
1423
1424- Vim bindings for text input - enable with /vim or /config
1425
1426## 0.2.32
1427
1428- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1429- Fix for some PersistentShell issues
1430
1431## 0.2.31
1432
1433- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1434- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1435
1436## 0.2.30
1437
1438- Added ANSI color theme for better terminal compatibility
1439- Fixed issue where slash command arguments weren't being sent properly
1440- (Mac-only) API keys are now stored in macOS Keychain
1441
1442## 0.2.26
1443
1444- New /approved-tools command for managing tool permissions
1445- Word-level diff display for improved code readability
1446- Fuzzy matching for slash commands
1447
1448## 0.2.21
1449
1450- Fuzzy matching for /commands