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 a01a88d · on Feb 13, 2026 History
1,699 lines · 89.6 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.42
4
5- Fixed /resume showing interrupt messages as session titles
6- Fixed Opus 4.6 launch announcement showing for Bedrock/Vertex/Foundry users
7- Improved error message for many-image dimension limit errors with /compact suggestion
8
9## 2.1.41
10
11- Added guard against launching Claude Code inside another Claude Code session
12- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
13- Fixed a crash when MCP tools return image content during streaming
14- Fixed /resume session previews showing raw XML tags instead of readable command names
15- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
16- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
17- Fixed hook blocking errors (exit code 2) not showing stderr to the user
18- Added `speed` attribute to OTel events and trace spans for fast mode visibility
19- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
20- Added Windows ARM64 (win32-arm64) native binary support
21- Improved `/rename` to auto-generate session name from conversation context when called without arguments
22- Improved narrow terminal layout for prompt footer
23- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
24- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
25- Fixed background task notifications not being delivered in streaming Agent SDK mode
26- Fixed cursor jumping to end on each keystroke in classifier rule input
27- Fixed markdown link display text being dropped for raw URL
28- Fixed auto-compact failure error notifications being shown to users
29- Fixed permission wait time being included in subagent elapsed time display
30- Fixed proactive ticks firing while in plan mode
31- Fixed clear stale permission rules when settings change on disk
32- Fixed hook blocking errors showing stderr content in UI
33
34## 2.1.39
35
36- Improved terminal rendering performance
37- Fixed fatal errors being swallowed instead of displayed
38- Fixed process hanging after session close
39- Fixed character loss at terminal screen boundary
40- Fixed blank lines in verbose transcript view
41
42## 2.1.38
43
44- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
45- Fixed Tab key queueing slash commands instead of autocompleting
46- Fixed bash permission matching for commands using environment variable wrappers
47- Fixed text between tool uses disappearing when not using streaming
48- Fixed duplicate sessions when resuming in VS Code extension
49- Improved heredoc delimiter parsing to prevent command smuggling
50- Blocked writes to `.claude/skills` directory in sandbox mode
51
52## 2.1.37
53
54- Fixed an issue where /fast was not immediately available after enabling /extra-usage
55
56## 2.1.36
57
58- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
59
60## 2.1.34
61
62- Fixed a crash when agent teams setting changed between renders
63- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
64
65## 2.1.33
66
67- Fixed agent teammate sessions in tmux to send and receive messages
68- Fixed warnings about agent teams not being available on your current plan
69- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
70- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
71- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
72- Added plugin name to skill descriptions and `/skills` menu for better discoverability
73- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
74- Fixed an API error that could occur when aborting mid-stream, where whitespace text combined with a thinking block would bypass normalization and produce an invalid request
75- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
76- Fixed an issue where proxy settings configured via `settings.json` environment variables were not applied to WebFetch and other HTTP requests on the Node.js build
77- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
78- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
79- Errors from invalid managed settings are now surfaced
80- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
81- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
82- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
83
84## 2.1.32
85
86- Claude Opus 4.6 is now available!
87- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
88- Claude now automatically records and recalls memories as it works
89- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
90- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
91- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
92- Updated --resume to re-use --agent value specified in previous conversation by default.
93- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
94- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
95- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
96- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
97- VSCode: Added spinner when loading past conversations list
98
99## 2.1.31
100
101- Added session resume hint on exit, showing how to continue your conversation later
102- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
103- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
104- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
105- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
106- Fixed `temperatureOverride` being silently ignored in the streaming API path, causing all streaming requests to use the default temperature (1) regardless of the configured override
107- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
108- Improved system prompts to more clearly guide the model toward using dedicated tools (Read, Edit, Glob, Grep) instead of bash equivalents (`cat`, `sed`, `grep`, `find`), reducing unnecessary bash command usage
109- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
110- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
111- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
112
113## 2.1.30
114
115- Added `pages` parameter to the Read tool for PDFs, allowing specific page ranges to be read (e.g., `pages: "1-5"`). Large PDFs (>10 pages) now return a lightweight reference when `@` mentioned instead of being inlined into context.
116- Added pre-configured OAuth client credentials for MCP servers that don't support Dynamic Client Registration (e.g., Slack). Use `--client-id` and `--client-secret` with `claude mcp add`.
117- Added `/debug` for Claude to help troubleshoot the current session
118- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
119- Added token count, tool uses, and duration metrics to Task tool results
120- Added reduced motion mode to the config
121- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
122- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
123- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
124- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
125- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
126- Fixed permission dialogs stealing focus while actively typing
127- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
128- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
129- Improved memory usage for `--resume` (68% reduction for users with many sessions) by replacing the session index with lightweight stat-based loading and progressive enrichment
130- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
131- Changed `/model` to execute immediately instead of being queued
132- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
133- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
134
135## 2.1.29
136
137- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
138
139## 2.1.27
140
141- Added tool call failures and denials to debug logs
142- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
143- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
144- Sessions are now automatically linked to PRs when created via `gh pr create`
145- Fixed /context command not displaying colored output
146- Fixed status bar duplicating background task indicator when PR status was shown
147- Windows: Fixed bash command execution failing for users with `.bashrc` files
148- Windows: Fixed console windows flashing when spawning child processes
149- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
150
151## 2.1.25
152
153- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
154
155## 2.1.23
156
157- Added customizable spinner verbs setting (`spinnerVerbs`)
158- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
159- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
160- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
161- Fixed pending async hooks not being cancelled when headless streaming sessions ended
162- Fixed tab completion not updating the input field when accepting a suggestion
163- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
164- Improved terminal rendering performance with optimized screen data layout
165- Changed Bash commands to show timeout duration alongside elapsed time
166- Changed merged pull requests to show a purple status indicator in the prompt footer
167- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
168
169## 2.1.22
170
171- Fixed structured outputs for non-interactive (-p) mode
172
173## 2.1.21
174
175- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
176- Fixed shell completion cache files being truncated on exit
177- Fixed API errors when resuming sessions that were interrupted during tool execution
178- Fixed auto-compact triggering too early on models with large output token limits
179- Fixed task IDs potentially being reused after deletion
180- Fixed file search not working in VS Code extension on Windows
181- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
182- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
183- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
184- [VSCode] Fixed message action buttons having incorrect background colors
185
186## 2.1.20
187
188- Added arrow key history navigation in vim normal mode when cursor cannot move further
189- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
190- Added PR review status indicator to the prompt footer, showing the current branch's PR state (approved, changes requested, pending, or draft) as a colored dot with a clickable link
191- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
192- Added ability to delete tasks via the `TaskUpdate` tool
193- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
194- Fixed agents sometimes ignoring user messages sent while actively working on a task
195- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
196- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
197- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
198- Fixed draft prompt being lost when pressing UP arrow to navigate command history
199- Fixed ghost text flickering when typing slash commands mid-input
200- Fixed marketplace source removal not properly deleting settings
201- Fixed duplicate output in some commands like `/context`
202- Fixed task list sometimes showing outside the main conversation view
203- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
204- Fixed crashes when cancelling tool use
205- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
206- Improved thinking status text with a subtle shimmer animation
207- Improved task list to dynamically adjust visible items based on terminal height
208- Improved fork conversation hint to show how to resume the original session
209- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
210- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
211- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
212- Changed the `/copy` command to be available to all users
213- Changed background agents to prompt for tool permissions before launching
214- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
215- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
216
217## 2.1.19
218
219- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
220- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
221- Fixed crashes on processors without AVX instruction support
222- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
223- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
224- Fixed resuming sessions by custom title not working when run from a different directory
225- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
226- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
227- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
228- Fixed file write preview omitting empty lines
229- Changed skills without additional permissions or hooks to be allowed without requiring approval
230- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
231- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
232- [VSCode] Enabled session forking and rewind functionality for all users
233
234## 2.1.18
235
236- Added customizable keyboard shortcuts. Configure keybindings per context, create chord sequences, and personalize your workflow. Run `/keybindings` to get started. Learn more at https://code.claude.com/docs/en/keybindings
237
238## 2.1.17
239
240- Fixed crashes on processors without AVX instruction support
241
242## 2.1.16
243
244- Added new task management system, including new capabilities like dependency tracking
245- [VSCode] Added native plugin management support
246- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
247- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
248- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
249- Fixed session titles on the resume screen not respecting the user's language setting
250- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
251
252## 2.1.15
253
254- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
255- Improved UI rendering performance with React Compiler
256- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
257- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
258
259## 2.1.14
260
261- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
262- Added search to installed plugins list - type to filter by name or description
263- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
264- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
265- Fixed memory issues that could cause crashes when running parallel subagents
266- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
267- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
268- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
269- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
270- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
271- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
272- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
273- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
274- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
275- Improved backspace to delete pasted text as a single token instead of one character at a time
276- [VSCode] Added `/usage` command to display current plan usage
277
278## 2.1.12
279
280- Fixed message rendering bug
281
282## 2.1.11
283
284- Fixed excessive MCP connection requests for HTTP/SSE transports
285
286## 2.1.10
287
288- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
289- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
290- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
291- Improved startup to capture keystrokes typed before the REPL is fully ready
292- Improved file suggestions to show as removable attachments instead of inserting text when accepted
293- [VSCode] Added install count display to plugin listings
294- [VSCode] Added trust warning when installing plugins
295
296## 2.1.9
297
298- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
299- Added `plansDirectory` setting to customize where plan files are stored
300- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
301- Added session URL attribution to commits and PRs created from web sessions
302- Added support for `PreToolUse` hooks to return `additionalContext` to the model
303- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
304- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
305- Fixed MCP server reconnection hanging when cached connection promise never resolves
306- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
307
308## 2.1.7
309
310- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
311- Added ability to provide feedback when accepting permission prompts
312- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
313- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
314- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
315- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
316- 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)
317- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
318- Fixed terminal title animation jitter by using fixed-width braille characters
319- Fixed plugins with git submodules not being fully initialized when installed
320- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
321- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
322- 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.
323- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
324- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
325
326## 2.1.6
327
328- Added search functionality to `/config` command for quickly filtering settings
329- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
330- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
331- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
332- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
333- Added an error display when the editor fails during Ctrl+G
334- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
335- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
336- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
337- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
338- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
339- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
340- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
341- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
342- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
343- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
344- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
345- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
346- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
347- Improved @ autocomplete with icons for different suggestion types and single-line formatting
348- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
349- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
350- Changed terminal title to "Claude Code" on startup for better window identification
351- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
352- [VSCode] Fixed usage indicator not updating after manual compact
353
354## 2.1.5
355
356- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
357
358## 2.1.4
359
360- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
361- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
362
363## 2.1.3
364
365- Merged slash commands and skills, simplifying the mental model with no change in behavior
366- Added release channel (`stable` or `latest`) toggle to `/config`
367- 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
368- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
369- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
370- Fixed mismatch between background task count in status bar and items shown in tasks dialog
371- Fixed sub-agents using the wrong model during conversation compaction
372- Fixed web search in sub-agents using incorrect model
373- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
374- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
375- Improved slash command suggestion readability by truncating long descriptions to 2 lines
376- Changed tool hook execution timeout from 60 seconds to 10 minutes
377- [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)
378
379## 2.1.2
380
381- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
382- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
383- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
384- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
385- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
386- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
387- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
388- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
389- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
390- Fixed updates incorrectly claiming another installation is in progress
391- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
392- Fixed remote session URL and teleport being broken when using `/tasks` command
393- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
394- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
395- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
396- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
397- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
398- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
399- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
400- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
401- [SDK] Changed minimum zod peer dependency to ^4.0.0
402- [VSCode] Fixed usage display not updating after manual compact
403
404## 2.1.0
405
406- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
407- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
408- Added support for `agent` field in skills to specify agent type for execution
409- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
410- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
411- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
412- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
413- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
414- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
415- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
416- Fixed Esc key with queued prompts to only move them to input without canceling the running task
417- Reduced permission prompts for complex bash commands
418- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
419- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
420- Fixed plugin path resolution for file-based marketplace sources
421- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
422- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
423- Fixed Claude in Chrome support for WSL environments
424- Fixed Windows native installer silently failing when executable creation fails
425- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
426- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
427- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
428- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
429- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
430- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
431- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
432- Added hooks support for skill and slash command frontmatter
433- 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
434- Added `/plan` command shortcut to enable plan mode directly from the prompt
435- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
436- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
437- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
438- Added support for `once: true` config for hooks
439- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
440- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
441- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
442- Added left/right arrow key navigation for cycling through tabs in dialogs
443- Added real-time thinking block display in Ctrl+O transcript mode
444- Added filepath to full output in background bash task details dialog
445- Added Skills as a separate category in the context visualization
446- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
447- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
448- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
449- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
450- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
451- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
452- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
453- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
454- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
455- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
456- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
457- Fixed commands with `$()` command substitution failing with parse errors
458- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
459- 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)
460- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
461- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
462- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
463- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
464- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
465- Fixed subagents sometimes not inheriting the parent's model by default
466- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
467- Fixed duplicate Bash commands appearing in permission request option labels
468- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
469- Fixed background task completion notifications to appear proactively with bullet point
470- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
471- Fixed cursor disappearing after dismissing permission dialogs
472- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
473- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
474- Fixed images being silently dropped when queueing messages while backgrounding a task
475- Fixed large pasted images failing with "Image was too large" error
476- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
477- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
478- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
479- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
480- Fixed spinner token counter to properly accumulate tokens from subagents during execution
481- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
482- Fixed race condition where LSP tool could return "no server available" during startup
483- Fixed feedback submission hanging indefinitely when network requests timeout
484- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
485- Fixed hook success message showing trailing colon when hook has no output
486- Multiple optimizations to improve startup performance
487- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
488- Improved performance when reading Jupyter notebooks with many cells
489- Improved reliability for piped input like `cat refactor.md | claude`
490- Improved reliability for AskQuestion tool
491- Improved sed in-place edit commands to render as file edits with diff preview
492- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
493- Improved compaction reliability
494- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
495- Improved skills to show progress while executing, displaying tool uses as they happen
496- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
497- Improved skill suggestions to prioritize recently and frequently used skills
498- Improved spinner feedback when waiting for the first response token
499- Improved token count display in spinner to include tokens from background agents
500- Improved incremental output for async agents to give the main thread more control and visibility
501- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
502- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
503- Improved macOS screenshot paste reliability with TIFF format support
504- Improved `/stats` output
505- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
506- Changed "Interrupted" message color from red to grey for a less alarming appearance
507- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
508- Removed underline styling from image reference links
509- [SDK] Changed minimum zod peer dependency to ^4.0.0
510- [VSCode] Added currently selected model name to the context menu
511- [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")
512- [VSCode] Fixed paragraph breaks not rendering in markdown content
513- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
514- [Windows] Fixed issue with improper rendering
515
516## 2.0.76
517
518- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
519
520## 2.0.75
521
522- Minor bugfixes
523
524## 2.0.74
525
526- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
527- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
528- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
529- Added syntax highlighting info to theme picker
530- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
531- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
532- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
533- Fixed a potential crash when syntax highlighting isn't initialized correctly
534- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
535- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
536- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
537- [Windows] Fixed issue with improper rendering
538- [VSCode] Added gift tag pictogram for year-end promotion message
539
540## 2.0.73
541
542- Added clickable `[Image #N]` links that open attached images in the default viewer
543- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
544- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
545- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
546- Fixed slow input history cycling and race condition that could overwrite text after message submission
547- Improved `/theme` command to open theme picker directly
548- Improved theme picker UI
549- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
550- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
551
552## 2.0.72
553
554- 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
555- Reduced terminal flickering
556- Added scannable QR code to mobile app tip for quick app downloads
557- Added loading indicator when resuming conversations for better feedback
558- Fixed `/context` command not respecting custom system prompts in non-interactive mode
559- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
560- Improved @ mention file suggestion speed (~3× faster in git repositories)
561- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
562- Improved settings validation errors to be more prominent
563- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
564
565## 2.0.71
566
567- Added /config toggle to enable/disable prompt suggestions
568- Added `/settings` as an alias for the `/config` command
569- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
570- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
571- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
572- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
573- New syntax highlighting engine for native build
574
575## 2.0.70
576
577- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
578- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
579- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
580- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
581- Fixed input being cleared when processing queued commands while the user was typing
582- Fixed prompt suggestions replacing typed input when pressing Tab
583- Fixed diff view not updating when terminal is resized
584- Improved memory usage by 3x for large conversations
585- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
586- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
587- Fix thinking mode toggle in /config not persisting correctly
588- Improve UI for file creation permission dialog
589
590## 2.0.69
591
592- Minor bugfixes
593
594## 2.0.68
595
596- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
597- Fixed a bug where disallowed MCP tools were visible to the model
598- Fixed an issue where steering messages could be lost while a subagent is working
599- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
600- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
601- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
602
603## 2.0.67
604
605- Thinking mode is now enabled by default for Opus 4.5
606- Thinking mode configuration has moved to /config
607- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
608- Show reason why autoupdater is disabled in `/doctor`
609- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
610- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
611- Fixed scroll position resetting after deleting a permission rule in `/permissions`
612- 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
613- Fixed `claude install --force` not bypassing stale lock files
614- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
615- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
616
617## 2.0.65
618
619- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
620- Added context window information to status line input
621- Added `fileSuggestion` setting for custom `@` file search commands
622- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
623- Fixed prompt not being saved to history when aborting a query with Escape
624- Fixed Read tool image handling to identify format from bytes instead of file extension
625
626## 2.0.64
627
628- Made auto-compacting instant
629- Agents and bash commands can run asynchronously and send messages to wake up the main agent
630- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
631- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
632- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
633- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
634- Fixed auto-loading .env when using native installer
635- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
636- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
637- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
638- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
639- Bedrock: Improve efficiency of token counting
640- Bedrock: Add support for `aws login` AWS Management Console credentials
641- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
642
643## 2.0.62
644
645- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
646- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
647- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
648- Fixed slash command selection not working when multiple commands share the same name
649- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
650- Fixed running versions getting removed because lock file incorrectly going stale
651- Fixed IDE diff tab not closing when rejecting file changes
652
653## 2.0.61
654
655- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
656
657## 2.0.60
658
659- Added background agent support. Agents run in the background while you work
660- Added --disable-slash-commands CLI flag to disable all slash commands
661- Added model name to "Co-Authored-By" commit messages
662- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
663- Updated Fetch to skip summarization for pre-approved websites
664- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
665
666## 2.0.59
667
668- Added --agent CLI flag to override the agent setting for the current session
669- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
670- VS Code: Fixed .claude.json config file being read from incorrect location
671
672## 2.0.58
673
674- Pro users now have access to Opus 4.5 as part of their subscription!
675- Fixed timer duration showing "11m 60s" instead of "12m 0s"
676- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
677
678## 2.0.57
679
680- Added feedback input when rejecting plans, allowing users to tell Claude what to change
681- VSCode: Added streaming message support for real-time response display
682
683## 2.0.56
684
685- Added setting to enable/disable terminal progress bar (OSC 9;4)
686- 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.
687
688## 2.0.55
689
690- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
691- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
692- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
693- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
694
695## 2.0.54
696
697- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
698- Fix issue with excessive iTerm notifications
699
700## 2.0.52
701
702- Fixed duplicate message display when starting Claude with a command line argument
703- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
704- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
705- Permit some uses of `$!` in bash commands
706
707## 2.0.51
708
709- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
710- Introducing Claude Code for Desktop: https://claude.com/download
711- 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
712- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
713- Plan Mode now builds more precise plans and executes more thoroughly
714- Usage limit notifications now easier to understand
715- Switched `/usage` back to "% used"
716- Fixed handling of thinking errors
717- Fixed performance regression
718
719## 2.0.50
720
721- Fixed bug preventing calling MCP tools that have nested references in their input schemas
722- Silenced a noisy but harmless error during upgrades
723- Improved ultrathink text display
724- Improved clarity of 5-hour session limit warning message
725
726## 2.0.49
727
728- Added readline-style ctrl-y for pasting deleted text
729- Improved clarity of usage limit warning message
730- Fixed handling of subagent permissions
731
732## 2.0.47
733
734- Improved error messages and validation for `claude --teleport`
735- Improved error handling in `/usage`
736- Fixed race condition with history entry not getting logged at exit
737- Fixed Vertex AI configuration not being applied from `settings.json`
738
739## 2.0.46
740
741- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
742
743## 2.0.45
744
745- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
746- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
747- Send background tasks to Claude Code on the web by starting a message with `&`
748
749## 2.0.43
750
751- Added `permissionMode` field for custom agents
752- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
753- Added skills frontmatter field to declare skills to auto-load for subagents
754- Added the `SubagentStart` hook event
755- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
756- Fixed duplicate rendering of some messages in the UI
757- Fixed some visual flickers
758- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
759
760## 2.0.42
761
762- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
763
764## 2.0.41
765
766- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
767- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
768- Fixed incorrect labeling of user settings vs project settings in command descriptions
769- Fixed crash when plugin command hooks timeout during execution
770- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
771- Fixed broken security documentation links in trust dialogs and onboarding
772- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
773- ctrl-r history search landing on a slash command no longer cancels the search
774- SDK: Support custom timeouts for hooks
775- Allow more safe git commands to run without approval
776- Plugins: Added support for sharing and installing output styles
777- Teleporting a session from web will automatically set the upstream branch
778
779## 2.0.37
780
781- Fixed how idleness is computed for notifications
782- Hooks: Added matcher values for Notification hook events
783- Output Styles: Added `keep-coding-instructions` option to frontmatter
784
785## 2.0.36
786
787- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
788- Fixed queued messages being incorrectly executed as bash commands
789- Fixed input being lost when typing while a queued message is processed
790
791## 2.0.35
792
793- Improve fuzzy search results when searching commands
794- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
795- 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
796- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
797- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
798
799## 2.0.34
800
801- VSCode Extension: Added setting to configure the initial permission mode for new conversations
802- Improved file path suggestion performance with native Rust-based fuzzy finder
803- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
804- Fixed memory crash when reading or writing large files (especially base64-encoded images)
805
806## 2.0.33
807
808- Native binary installs now launch quicker.
809- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
810- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
811
812## 2.0.32
813
814- Un-deprecate output styles based on community feedback
815- Added `companyAnnouncements` setting for displaying announcements on startup
816- Fixed hook progress messages not updating correctly during PostToolUse hook execution
817
818## 2.0.31
819
820- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
821- Vertex: add support for Web Search on supported models
822- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
823- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
824- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
825- Fixed plugin uninstall not removing plugins
826
827## 2.0.30
828
829- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
830- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
831- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
832- Added prompt-based stop hooks
833- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
834- Enabled SSE MCP servers on native build
835- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
836- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
837- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
838- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
839- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
840- Fixed bug that saved session permissions to local settings
841- Fixed MCP tools not being available to sub-agents
842- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
843- Fixed delay when navigating through typeahead suggestions with arrow keys
844- VSCode: Restored selection indicator in input footer showing current file or code selection status
845
846## 2.0.28
847
848- Plan mode: introduced new Plan subagent
849- Subagents: claude can now choose to resume subagents
850- Subagents: claude can dynamically choose the model used by its subagents
851- SDK: added --max-budget-usd flag
852- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
853- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
854- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
855- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
856- Various other bug fixes
857
858## 2.0.27
859
860- New UI for permission prompts
861- Added current branch filtering and search to session resume screen for easier navigation
862- Fixed directory @-mention causing "No assistant message found" error
863- VSCode Extension: Add config setting to include .gitignored files in file searches
864- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
865
866## 2.0.25
867
868- 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
869
870## 2.0.24
871
872- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
873- Claude Code Web: Support for Web -> CLI teleport
874- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
875- Bedrock: Display awsAuthRefresh output when auth is required
876
877## 2.0.22
878
879- Fixed content layout shift when scrolling through slash commands
880- IDE: Add toggle to enable/disable thinking.
881- Fix bug causing duplicate permission prompts with parallel tool calls
882- Add support for enterprise managed MCP allowlist and denylist
883
884## 2.0.21
885
886- Support MCP `structuredContent` field in tool responses
887- Added an interactive question tool
888- Claude will now ask you questions more often in plan mode
889- Added Haiku 4.5 as a model option for Pro users
890- Fixed an issue where queued commands don't have access to previous messages' output
891
892## 2.0.20
893
894- Added support for Claude Skills
895
896## 2.0.19
897
898- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
899- Fixed a bug where Haiku was unnecessarily called in print mode
900
901## 2.0.17
902
903- Added Haiku 4.5 to model selector!
904- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
905- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
906- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
907- OTEL: support HTTP_PROXY and HTTPS_PROXY
908- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
909
910## 2.0.15
911
912- Fixed bug with resuming where previously created files needed to be read again before writing
913- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
914
915## 2.0.14
916
917- Fix @-mentioning MCP servers to toggle them on/off
918- Improve permission checks for bash with inline env vars
919- Fix ultrathink + thinking toggle
920- Reduce unnecessary logins
921- Document --system-prompt
922- Several improvements to rendering
923- Plugins UI polish
924
925## 2.0.13
926
927- Fixed `/plugin` not working on native build
928
929## 2.0.12
930
931- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
932- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
933- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
934- `/plugin validate` command for validating plugin structure and configuration
935- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
936- Plugin documentation available at https://code.claude.com/docs/en/plugins
937- Comprehensive error messages and diagnostics via `/doctor` command
938- Avoid flickering in `/model` selector
939- Improvements to `/help`
940- Avoid mentioning hooks in `/resume` summaries
941- Changes to the "verbose" setting in `/config` now persist across sessions
942
943## 2.0.11
944
945- Reduced system prompt size by 1.4k tokens
946- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
947- Fixed Opus fallback rate limit errors appearing incorrectly
948- Fixed /add-dir command selecting wrong default tab
949
950## 2.0.10
951
952- Rewrote terminal renderer for buttery smooth UI
953- Enable/disable MCP servers by @mentioning, or in /mcp
954- Added tab completion for shell commands in bash mode
955- PreToolUse hooks can now modify tool inputs
956- Press Ctrl-G to edit your prompt in your system's configured text editor
957- Fixes for bash permission checks with environment variables in the command
958
959## 2.0.9
960
961- Fix regression where bash backgrounding stopped working
962
963## 2.0.8
964
965- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
966- IDE: Add drag-and-drop support for files and folders in chat
967- /context: Fix counting for thinking blocks
968- Improve message rendering for users with light themes on dark terminals
969- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
970
971## 2.0.5
972
973- IDE: Fix IME unintended message submission with Enter and Tab
974- IDE: Add "Open in Terminal" link in login screen
975- Fix unhandled OAuth expiration 401 API errors
976- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
977
978## 2.0.1
979
980- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
981- Various bug fixes and presentation improvements
982
983## 2.0.0
984
985- New native VS Code extension
986- Fresh coat of paint throughout the whole app
987- /rewind a conversation to undo code changes
988- /usage command to see plan limits
989- Tab to toggle thinking (sticky across sessions)
990- Ctrl-R to search history
991- Unshipped claude config command
992- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
993- SDK: The Claude Code SDK is now the Claude Agent SDK
994- Add subagents dynamically with `--agents` flag
995
996## 1.0.126
997
998- Enable /context command for Bedrock and Vertex
999- Add mTLS support for HTTP-based OpenTelemetry exporters
1000
1001## 1.0.124
1002
1003- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1004- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1005- No longer inform Claude of the list of allowed tools when permission is denied
1006- Fixed security vulnerability in Bash tool permission checks
1007- Improved VSCode extension performance for large files
1008
1009## 1.0.123
1010
1011- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1012- Fixed thinking mode triggering on negation phrases like "don't think"
1013- Fixed rendering performance degradation during token streaming
1014- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1015- Enhanced BashTool environment snapshot logging
1016- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1017- Migrated --debug logging to a file, to enable easy tailing & filtering
1018
1019## 1.0.120
1020
1021- Fix input lag during typing, especially noticeable with large prompts
1022- Improved VSCode extension command registry and sessions dialog user experience
1023- Enhanced sessions dialog responsiveness and visual feedback
1024- Fixed IDE compatibility issue by removing worktree support check
1025- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1026
1027## 1.0.119
1028
1029- Fix Windows issue where process visually freezes on entering interactive mode
1030- Support dynamic headers for MCP servers via headersHelper configuration
1031- Fix thinking mode not working in headless sessions
1032- Fix slash commands now properly update allowed tools instead of replacing them
1033
1034## 1.0.117
1035
1036- Add Ctrl-R history search to recall previous commands like bash/zsh
1037- Fix input lag while typing, especially on Windows
1038- Add sed command to auto-allowed commands in acceptEdits mode
1039- Fix Windows PATH comparison to be case-insensitive for drive letters
1040- Add permissions management hint to /add-dir output
1041
1042## 1.0.115
1043
1044- Improve thinking mode display with enhanced visual effects
1045- Type /t to temporarily disable thinking mode in your prompt
1046- Improve path validation for glob and grep tools
1047- Show condensed output for post-tool hooks to reduce visual clutter
1048- Fix visual feedback when loading state completes
1049- Improve UI consistency for permission request dialogs
1050
1051## 1.0.113
1052
1053- Deprecated piped input in interactive mode
1054- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1055
1056## 1.0.112
1057
1058- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1059- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1060- Hooks: Added systemMessage support for SessionEnd hooks
1061- Added `spinnerTipsEnabled` setting to disable spinner tips
1062- IDE: Various improvements and bug fixes
1063
1064## 1.0.111
1065
1066- /model now validates provided model names
1067- Fixed Bash tool crashes caused by malformed shell syntax parsing
1068
1069## 1.0.110
1070
1071- /terminal-setup command now supports WezTerm
1072- MCP: OAuth tokens now proactively refresh before expiration
1073- Fixed reliability issues with background Bash processes
1074
1075## 1.0.109
1076
1077- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1078
1079## 1.0.106
1080
1081- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1082
1083## 1.0.97
1084
1085- Settings: /doctor now validates permission rule syntax and suggests corrections
1086
1087## 1.0.94
1088
1089- Vertex: add support for global endpoints for supported models
1090- /memory command now allows direct editing of all imported memory files
1091- SDK: Add custom tools as callbacks
1092- Added /todos command to list current todo items
1093
1094## 1.0.93
1095
1096- Windows: Add alt + v shortcut for pasting images from clipboard
1097- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1098
1099## 1.0.90
1100
1101- Settings file changes take effect immediately - no restart required
1102
1103## 1.0.88
1104
1105- Fixed issue causing "OAuth authentication is currently not supported"
1106- Status line input now includes `exceeds_200k_tokens`
1107- Fixed incorrect usage tracking in /cost.
1108- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1109- Bedrock: Updated default Sonnet model to Sonnet 4
1110
1111## 1.0.86
1112
1113- Added /context to help users self-serve debug context issues
1114- SDK: Added UUID support for all SDK messages
1115- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1116
1117## 1.0.85
1118
1119- Status line input now includes session cost info
1120- Hooks: Introduced SessionEnd hook
1121
1122## 1.0.84
1123
1124- Fix tool_use/tool_result id mismatch error when network is unstable
1125- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1126- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1127- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1128
1129## 1.0.83
1130
1131- @-mention: Support files with spaces in path
1132- New shimmering spinner
1133
1134## 1.0.82
1135
1136- SDK: Add request cancellation support
1137- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1138- Settings: Validation prevents invalid fields in .claude/settings.json files
1139- MCP: Improve tool name consistency
1140- Bash: Fix crash when Claude tries to automatically read large files
1141
1142## 1.0.81
1143
1144- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1145- Agents: Fix custom agent loading when agent files are unparsable
1146
1147## 1.0.80
1148
1149- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1150
1151## 1.0.77
1152
1153- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1154- SDK: Add session support and permission denial tracking
1155- Fix token limit errors in conversation summarization
1156- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1157
1158## 1.0.73
1159
1160- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1161- MCP: Press Esc to cancel OAuth authentication flows
1162- Bash: Improved command validation and reduced false security warnings
1163- UI: Enhanced spinner animations and status line visual hierarchy
1164- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1165
1166## 1.0.72
1167
1168- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1169
1170## 1.0.71
1171
1172- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1173- Customizable status line: add your terminal prompt to Claude Code with /statusline
1174
1175## 1.0.70
1176
1177- Performance: Optimized message rendering for better performance with large contexts
1178- Windows: Fixed native file search, ripgrep, and subagent functionality
1179- Added support for @-mentions in slash command arguments
1180
1181## 1.0.69
1182
1183- Upgraded Opus to version 4.1
1184
1185## 1.0.68
1186
1187- Fix incorrect model names being used for certain commands like `/pr-comments`
1188- 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.
1189- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1190- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1191- SDK: Added canUseTool callback support for tool confirmation
1192- Added `disableAllHooks` setting
1193- Improved file suggestions performance in large repos
1194
1195## 1.0.65
1196
1197- IDE: Fixed connection stability issues and error handling for diagnostics
1198- Windows: Fixed shell environment setup for users without .bashrc files
1199
1200## 1.0.64
1201
1202- Agents: Added model customization support - you can now specify which model an agent should use
1203- Agents: Fixed unintended access to the recursive agent tool
1204- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1205- SDK: Fixed user input tracking across multi-turn conversations
1206- Added hidden files to file search and @-mention suggestions
1207
1208## 1.0.63
1209
1210- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1211
1212## 1.0.62
1213
1214- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1215- Hooks: Added SessionStart hook for new session initialization
1216- /add-dir command now supports typeahead for directory paths
1217- Improved network connectivity check reliability
1218
1219## 1.0.61
1220
1221- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1222- Settings: Added `--settings` flag to load settings from a JSON file
1223- Settings: Fixed resolution of settings files paths that are symlinks
1224- OTEL: Fixed reporting of wrong organization after authentication changes
1225- Slash commands: Fixed permissions checking for allowed-tools with Bash
1226- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1227- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1228- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1229
1230## 1.0.60
1231
1232- You can now create custom subagents for specialized tasks! Run /agents to get started
1233
1234## 1.0.59
1235
1236- SDK: Added tool confirmation support with canUseTool callback
1237- SDK: Allow specifying env for spawned process
1238- Hooks: Exposed PermissionDecision to hooks (including "ask")
1239- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1240- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1241
1242## 1.0.58
1243
1244- Added support for reading PDFs
1245- MCP: Improved server health status display in 'claude mcp list'
1246- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1247
1248## 1.0.57
1249
1250- Added support for specifying a model in slash commands
1251- Improved permission messages to help Claude understand allowed tools
1252- Fix: Remove trailing newlines from bash output in terminal wrapping
1253
1254## 1.0.56
1255
1256- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1257- Fixes for WSL IDE detection
1258- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1259
1260## 1.0.55
1261
1262- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1263- Windows: fixed Ctrl+Z crash
1264- SDK: Added ability to capture error logging
1265- Add --system-prompt-file option to override system prompt in print mode
1266
1267## 1.0.54
1268
1269- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1270- Custom slash commands: Added argument-hint to frontmatter
1271- Windows: OAuth uses port 45454 and properly constructs browser URL
1272- Windows: mode switching now uses alt + m, and plan mode renders properly
1273- Shell: Switch to in-memory shell snapshot to fix file-related errors
1274
1275## 1.0.53
1276
1277- Updated @-mention file truncation from 100 lines to 2000 lines
1278- 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).
1279
1280## 1.0.52
1281
1282- Added support for MCP server instructions
1283
1284## 1.0.51
1285
1286- Added support for native Windows (requires Git for Windows)
1287- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1288- Settings: /doctor can now help you identify and fix invalid setting files
1289- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1290- Increased auto-compact warning threshold from 60% to 80%
1291- Fixed an issue with handling user directories with spaces for shell snapshots
1292- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1293- Custom slash commands: Fixed user-level commands in subdirectories
1294- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1295
1296## 1.0.48
1297
1298- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1299- Added progress messages to Bash tool based on the last 5 lines of command output
1300- Added expanding variables support for MCP server configuration
1301- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1302- Improved IDE extension path handling when Claude Code runs in WSL
1303- Hooks: Added a PreCompact hook
1304- Vim mode: Added c, f/F, t/T
1305
1306## 1.0.45
1307
1308- Redesigned Search (Grep) tool with new tool input parameters and features
1309- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1310- Fixed config file corruption issue by enforcing atomic writes
1311- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1312- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1313- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1314
1315## 1.0.44
1316
1317- New /export command lets you quickly export a conversation for sharing
1318- MCP: resource_link tool results are now supported
1319- MCP: tool annotations and tool titles now display in /mcp view
1320- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1321
1322## 1.0.43
1323
1324- Fixed a bug where the theme selector was saving excessively
1325- Hooks: Added EPIPE system error handling
1326
1327## 1.0.42
1328
1329- Added tilde (`~`) expansion support to `/add-dir` command
1330
1331## 1.0.41
1332
1333- Hooks: Split Stop hook triggering into Stop and SubagentStop
1334- Hooks: Enabled optional timeout configuration for each command
1335- Hooks: Added "hook_event_name" to hook input
1336- Fixed a bug where MCP tools would display twice in tool list
1337- New tool parameters JSON for Bash tool in `tool_decision` event
1338
1339## 1.0.40
1340
1341- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1342
1343## 1.0.39
1344
1345- New Active Time metric in OpenTelemetry logging
1346
1347## 1.0.38
1348
1349- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1350
1351## 1.0.37
1352
1353- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1354
1355## 1.0.36
1356
1357- Web search now takes today's date into context
1358- Fixed a bug where stdio MCP servers were not terminating properly on exit
1359
1360## 1.0.35
1361
1362- Added support for MCP OAuth Authorization Server discovery
1363
1364## 1.0.34
1365
1366- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1367
1368## 1.0.33
1369
1370- Improved logging functionality with session ID support
1371- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1372- Improvements to plan mode
1373
1374## 1.0.32
1375
1376- Updated loopback config for litellm
1377- Added forceLoginMethod setting to bypass login selection screen
1378
1379## 1.0.31
1380
1381- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1382
1383## 1.0.30
1384
1385- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1386- Improved file path autocomplete with filename matching
1387- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1388- Enhanced jq regex support for complex filters with pipes and select
1389
1390## 1.0.29
1391
1392- Improved CJK character support in cursor navigation and rendering
1393
1394## 1.0.28
1395
1396- Slash commands: Fix selector display during history navigation
1397- Resizes images before upload to prevent API size limit errors
1398- Added XDG_CONFIG_HOME support to configuration directory
1399- Performance optimizations for memory usage
1400- New attributes (terminal.type, language) in OpenTelemetry logging
1401
1402## 1.0.27
1403
1404- Streamable HTTP MCP servers are now supported
1405- Remote MCP servers (SSE and HTTP) now support OAuth
1406- MCP resources can now be @-mentioned
1407- /resume slash command to switch conversations within Claude Code
1408
1409## 1.0.25
1410
1411- Slash commands: moved "project" and "user" prefixes to descriptions
1412- Slash commands: improved reliability for command discovery
1413- Improved support for Ghostty
1414- Improved web search reliability
1415
1416## 1.0.24
1417
1418- Improved /mcp output
1419- Fixed a bug where settings arrays got overwritten instead of merged
1420
1421## 1.0.23
1422
1423- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1424- Released Python SDK: pip install claude-code-sdk to get started
1425
1426## 1.0.22
1427
1428- SDK: Renamed `total_cost` to `total_cost_usd`
1429
1430## 1.0.21
1431
1432- Improved editing of files with tab-based indentation
1433- Fix for tool_use without matching tool_result errors
1434- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1435
1436## 1.0.18
1437
1438- Added --add-dir CLI argument for specifying additional working directories
1439- Added streaming input support without require -p flag
1440- Improved startup performance and session storage performance
1441- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1442- Added detailed MCP server tools display (/mcp)
1443- MCP authentication and permission improvements
1444- Added auto-reconnection for MCP SSE connections on disconnect
1445- Fixed issue where pasted content was lost when dialogs appeared
1446
1447## 1.0.17
1448
1449- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1450- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1451- MCP server list UI improvements
1452- Update Claude Code process title to display "claude" instead of "node"
1453
1454## 1.0.11
1455
1456- Claude Code can now also be used with a Claude Pro subscription
1457- Added /upgrade for smoother switching to Claude Max plans
1458- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1459- Improved shell configuration error handling
1460- Improved todo list handling during compaction
1461
1462## 1.0.10
1463
1464- Added markdown table support
1465- Improved streaming performance
1466
1467## 1.0.8
1468
1469- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1470- Increased default otel interval from 1s -> 5s
1471- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1472- Fixed a regression where search tools unnecessarily asked for permissions
1473- Added support for triggering thinking non-English languages
1474- Improved compacting UI
1475
1476## 1.0.7
1477
1478- Renamed /allowed-tools -> /permissions
1479- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1480- Deprecated claude config commands in favor of editing settings.json
1481- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1482- Improved error handling for /install-github-app
1483- Bugfixes, UI polish, and tool reliability improvements
1484
1485## 1.0.6
1486
1487- Improved edit reliability for tab-indented files
1488- Respect CLAUDE_CONFIG_DIR everywhere
1489- Reduced unnecessary tool permission prompts
1490- Added support for symlinks in @file typeahead
1491- Bugfixes, UI polish, and tool reliability improvements
1492
1493## 1.0.4
1494
1495- Fixed a bug where MCP tool errors weren't being parsed correctly
1496
1497## 1.0.1
1498
1499- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1500- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1501- Updated documentation links and OAuth process descriptions
1502
1503## 1.0.0
1504
1505- Claude Code is now generally available
1506- Introducing Sonnet 4 and Opus 4 models
1507
1508## 0.2.125
1509
1510- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1511- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1512
1513## 0.2.117
1514
1515- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1516- Introduced settings.cleanupPeriodDays
1517- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1518- Introduced --debug mode
1519
1520## 0.2.108
1521
1522- You can now send messages to Claude while it works to steer Claude in real-time
1523- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1524- Fixed a bug where thinking was not working in -p mode
1525- Fixed a regression in /cost reporting
1526- Deprecated MCP wizard interface in favor of other MCP commands
1527- Lots of other bugfixes and improvements
1528
1529## 0.2.107
1530
1531- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1532
1533## 0.2.106
1534
1535- MCP SSE server configs can now specify custom headers
1536- Fixed a bug where MCP permission prompt didn't always show correctly
1537
1538## 0.2.105
1539
1540- Claude can now search the web
1541- Moved system & account status to /status
1542- Added word movement keybindings for Vim
1543- Improved latency for startup, todo tool, and file edits
1544
1545## 0.2.102
1546
1547- Improved thinking triggering reliability
1548- Improved @mention reliability for images and folders
1549- You can now paste multiple large chunks into one prompt
1550
1551## 0.2.100
1552
1553- Fixed a crash caused by a stack overflow error
1554- Made db storage optional; missing db support disables --continue and --resume
1555
1556## 0.2.98
1557
1558- Fixed an issue where auto-compact was running twice
1559
1560## 0.2.96
1561
1562- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1563
1564## 0.2.93
1565
1566- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1567- Claude now has access to a Todo list that helps it stay on track and be more organized
1568
1569## 0.2.82
1570
1571- Added support for --disallowedTools
1572- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1573
1574## 0.2.75
1575
1576- Hit Enter to queue up additional messages while Claude is working
1577- Drag in or copy/paste image files directly into the prompt
1578- @-mention files to directly add them to context
1579- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1580- Improved performance for filename auto-complete
1581
1582## 0.2.74
1583
1584- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1585- Task tool can now perform writes and run bash commands
1586
1587## 0.2.72
1588
1589- Updated spinner to indicate tokens loaded and tool usage
1590
1591## 0.2.70
1592
1593- Network commands like curl are now available for Claude to use
1594- Claude can now run multiple web queries in parallel
1595- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1596
1597## 0.2.69
1598
1599- Fixed UI glitches with improved Select component behavior
1600- Enhanced terminal output display with better text truncation logic
1601
1602## 0.2.67
1603
1604- Shared project permission rules can be saved in .claude/settings.json
1605
1606## 0.2.66
1607
1608- Print mode (-p) now supports streaming output via --output-format=stream-json
1609- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1610
1611## 0.2.63
1612
1613- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1614
1615## 0.2.61
1616
1617- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1618- Enhanced image detection for more reliable clipboard paste functionality
1619- Fixed an issue where ESC key could crash the conversation history selector
1620
1621## 0.2.59
1622
1623- Copy+paste images directly into your prompt
1624- Improved progress indicators for bash and fetch tools
1625- Bugfixes for non-interactive mode (-p)
1626
1627## 0.2.54
1628
1629- Quickly add to Memory by starting your message with '#'
1630- Press ctrl+r to see full output for long tool results
1631- Added support for MCP SSE transport
1632
1633## 0.2.53
1634
1635- New web fetch tool lets Claude view URLs that you paste in
1636- Fixed a bug with JPEG detection
1637
1638## 0.2.50
1639
1640- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1641
1642## 0.2.49
1643
1644- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1645
1646## 0.2.47
1647
1648- Press Tab to auto-complete file and folder names
1649- Press Shift + Tab to toggle auto-accept for file edits
1650- Automatic conversation compaction for infinite conversation length (toggle with /config)
1651
1652## 0.2.44
1653
1654- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1655
1656## 0.2.41
1657
1658- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1659- MCP server startup no longer blocks the app from starting up
1660
1661## 0.2.37
1662
1663- New /release-notes command lets you view release notes at any time
1664- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1665
1666## 0.2.36
1667
1668- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1669- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1670
1671## 0.2.34
1672
1673- Vim bindings for text input - enable with /vim or /config
1674
1675## 0.2.32
1676
1677- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1678- Fix for some PersistentShell issues
1679
1680## 0.2.31
1681
1682- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1683- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1684
1685## 0.2.30
1686
1687- Added ANSI color theme for better terminal compatibility
1688- Fixed issue where slash command arguments weren't being sent properly
1689- (Mac-only) API keys are now stored in macOS Keychain
1690
1691## 0.2.26
1692
1693- New /approved-tools command for managing tool permissions
1694- Word-level diff display for improved code readability
1695- Fuzzy matching for slash commands
1696
1697## 0.2.21
1698
1699- Fuzzy matching for /commands