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 923d727 · on Jan 17, 2026 History
1,420 lines · 70.1 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.11
4
5- Fixed excessive MCP connection requests for HTTP/SSE transports
6
7## 2.1.10
8
9- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
10- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
11- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
12- Improved startup to capture keystrokes typed before the REPL is fully ready
13- Improved file suggestions to show as removable attachments instead of inserting text when accepted
14- [VSCode] Added install count display to plugin listings
15- [VSCode] Added trust warning when installing plugins
16
17## 2.1.9
18
19- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
20- Added `plansDirectory` setting to customize where plan files are stored
21- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
22- Added session URL attribution to commits and PRs created from web sessions
23- Added support for `PreToolUse` hooks to return `additionalContext` to the model
24- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
25- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
26- Fixed MCP server reconnection hanging when cached connection promise never resolves
27- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
28
29## 2.1.7
30
31- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
32- Added ability to provide feedback when accepting permission prompts
33- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
34- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
35- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
36- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
37- 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)
38- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
39- Fixed terminal title animation jitter by using fixed-width braille characters
40- Fixed plugins with git submodules not being fully initialized when installed
41- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
42- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
43- 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.
44- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
45- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
46
47## 2.1.6
48
49- Added search functionality to `/config` command for quickly filtering settings
50- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
51- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
52- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
53- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
54- Added an error display when the editor fails during Ctrl+G
55- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
56- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
57- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
58- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
59- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
60- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
61- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
62- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
63- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
64- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
65- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
66- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
67- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
68- Improved @ autocomplete with icons for different suggestion types and single-line formatting
69- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
70- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
71- Changed terminal title to "Claude Code" on startup for better window identification
72- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
73- [VSCode] Fixed usage indicator not updating after manual compact
74
75## 2.1.5
76
77- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
78
79## 2.1.4
80
81- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
82- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
83
84## 2.1.3
85
86- Merged slash commands and skills, simplifying the mental model with no change in behavior
87- Added release channel (`stable` or `latest`) toggle to `/config`
88- 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
89- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
90- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
91- Fixed mismatch between background task count in status bar and items shown in tasks dialog
92- Fixed sub-agents using the wrong model during conversation compaction
93- Fixed web search in sub-agents using incorrect model
94- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
95- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
96- Improved slash command suggestion readability by truncating long descriptions to 2 lines
97- Changed tool hook execution timeout from 60 seconds to 10 minutes
98- [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)
99
100## 2.1.2
101
102- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
103- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
104- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
105- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
106- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
107- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
108- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
109- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
110- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
111- Fixed updates incorrectly claiming another installation is in progress
112- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
113- Fixed remote session URL and teleport being broken when using `/tasks` command
114- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
115- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
116- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
117- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
118- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
119- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
120- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
121- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
122- [SDK] Changed minimum zod peer dependency to ^4.0.0
123- [VSCode] Fixed usage display not updating after manual compact
124
125## 2.1.0
126
127- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
128- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
129- Added support for `agent` field in skills to specify agent type for execution
130- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
131- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
132- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
133- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
134- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
135- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
136- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
137- Fixed Esc key with queued prompts to only move them to input without canceling the running task
138- Reduced permission prompts for complex bash commands
139- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
140- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
141- Fixed plugin path resolution for file-based marketplace sources
142- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
143- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
144- Fixed Claude in Chrome support for WSL environments
145- Fixed Windows native installer silently failing when executable creation fails
146- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
147- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
148- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
149- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
150- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
151- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
152- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
153- Added hooks support for skill and slash command frontmatter
154- 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
155- Added `/plan` command shortcut to enable plan mode directly from the prompt
156- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
157- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
158- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
159- Added support for `once: true` config for hooks
160- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
161- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
162- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
163- Added left/right arrow key navigation for cycling through tabs in dialogs
164- Added real-time thinking block display in Ctrl+O transcript mode
165- Added filepath to full output in background bash task details dialog
166- Added Skills as a separate category in the context visualization
167- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
168- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
169- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
170- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
171- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
172- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
173- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
174- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
175- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
176- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
177- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
178- Fixed commands with `$()` command substitution failing with parse errors
179- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
180- 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)
181- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
182- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
183- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
184- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
185- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
186- Fixed subagents sometimes not inheriting the parent's model by default
187- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
188- Fixed duplicate Bash commands appearing in permission request option labels
189- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
190- Fixed background task completion notifications to appear proactively with bullet point
191- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
192- Fixed cursor disappearing after dismissing permission dialogs
193- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
194- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
195- Fixed images being silently dropped when queueing messages while backgrounding a task
196- Fixed large pasted images failing with "Image was too large" error
197- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
198- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
199- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
200- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
201- Fixed spinner token counter to properly accumulate tokens from subagents during execution
202- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
203- Fixed race condition where LSP tool could return "no server available" during startup
204- Fixed feedback submission hanging indefinitely when network requests timeout
205- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
206- Fixed hook success message showing trailing colon when hook has no output
207- Multiple optimizations to improve startup performance
208- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
209- Improved performance when reading Jupyter notebooks with many cells
210- Improved reliability for piped input like `cat refactor.md | claude`
211- Improved reliability for AskQuestion tool
212- Improved sed in-place edit commands to render as file edits with diff preview
213- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
214- Improved compaction reliability
215- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
216- Improved skills to show progress while executing, displaying tool uses as they happen
217- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
218- Improved skill suggestions to prioritize recently and frequently used skills
219- Improved spinner feedback when waiting for the first response token
220- Improved token count display in spinner to include tokens from background agents
221- Improved incremental output for async agents to give the main thread more control and visibility
222- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
223- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
224- Improved macOS screenshot paste reliability with TIFF format support
225- Improved `/stats` output
226- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
227- Changed "Interrupted" message color from red to grey for a less alarming appearance
228- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
229- Removed underline styling from image reference links
230- [SDK] Changed minimum zod peer dependency to ^4.0.0
231- [VSCode] Added currently selected model name to the context menu
232- [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")
233- [VSCode] Fixed paragraph breaks not rendering in markdown content
234- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
235- [Windows] Fixed issue with improper rendering
236
237## 2.0.76
238
239- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
240
241## 2.0.75
242
243- Minor bugfixes
244
245## 2.0.74
246
247- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
248- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
249- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
250- Added syntax highlighting info to theme picker
251- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
252- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
253- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
254- Fixed a potential crash when syntax highlighting isn't initialized correctly
255- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
256- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
257- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
258- [Windows] Fixed issue with improper rendering
259- [VSCode] Added gift tag pictogram for year-end promotion message
260
261## 2.0.73
262
263- Added clickable `[Image #N]` links that open attached images in the default viewer
264- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
265- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
266- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
267- Fixed slow input history cycling and race condition that could overwrite text after message submission
268- Improved `/theme` command to open theme picker directly
269- Improved theme picker UI
270- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
271- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
272
273## 2.0.72
274
275- 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
276- Reduced terminal flickering
277- Added scannable QR code to mobile app tip for quick app downloads
278- Added loading indicator when resuming conversations for better feedback
279- Fixed `/context` command not respecting custom system prompts in non-interactive mode
280- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
281- Improved @ mention file suggestion speed (~3x faster in git repositories)
282- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
283- Improved settings validation errors to be more prominent
284- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
285
286## 2.0.71
287
288- Added /config toggle to enable/disable prompt suggestions
289- Added `/settings` as an alias for the `/config` command
290- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
291- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
292- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
293- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
294- New syntax highlighting engine for native build
295
296## 2.0.70
297
298- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
299- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
300- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
301- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
302- Fixed input being cleared when processing queued commands while the user was typing
303- Fixed prompt suggestions replacing typed input when pressing Tab
304- Fixed diff view not updating when terminal is resized
305- Improved memory usage by 3x for large conversations
306- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
307- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
308- Fix thinking mode toggle in /config not persisting correctly
309- Improve UI for file creation permission dialog
310
311## 2.0.69
312
313- Minor bugfixes
314
315## 2.0.68
316
317- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
318- Fixed a bug where disallowed MCP tools were visible to the model
319- Fixed an issue where steering messages could be lost while a subagent is working
320- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
321- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
322- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
323
324## 2.0.67
325
326- Thinking mode is now enabled by default for Opus 4.5
327- Thinking mode configuration has moved to /config
328- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
329- Show reason why autoupdater is disabled in `/doctor`
330- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
331- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
332- Fixed scroll position resetting after deleting a permission rule in `/permissions`
333- 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
334- Fixed `claude install --force` not bypassing stale lock files
335- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
336- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
337
338## 2.0.65
339
340- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
341- Added context window information to status line input
342- Added `fileSuggestion` setting for custom `@` file search commands
343- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
344- Fixed prompt not being saved to history when aborting a query with Escape
345- Fixed Read tool image handling to identify format from bytes instead of file extension
346
347## 2.0.64
348
349- Made auto-compacting instant
350- Agents and bash commands can run asynchronously and send messages to wake up the main agent
351- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
352- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
353- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
354- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
355- Fixed auto-loading .env when using native installer
356- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
357- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
358- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
359- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
360- Bedrock: Improve efficiency of token counting
361- Bedrock: Add support for `aws login` AWS Management Console credentials
362- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
363
364## 2.0.62
365
366- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
367- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
368- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
369- Fixed slash command selection not working when multiple commands share the same name
370- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
371- Fixed running versions getting removed because lock file incorrectly going stale
372- Fixed IDE diff tab not closing when rejecting file changes
373
374## 2.0.61
375
376- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
377
378## 2.0.60
379
380- Added background agent support. Agents run in the background while you work
381- Added --disable-slash-commands CLI flag to disable all slash commands
382- Added model name to "Co-Authored-By" commit messages
383- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
384- Updated Fetch to skip summarization for pre-approved websites
385- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
386
387## 2.0.59
388
389- Added --agent CLI flag to override the agent setting for the current session
390- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
391- VS Code: Fixed .claude.json config file being read from incorrect location
392
393## 2.0.58
394
395- Pro users now have access to Opus 4.5 as part of their subscription!
396- Fixed timer duration showing "11m 60s" instead of "12m 0s"
397- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
398
399## 2.0.57
400
401- Added feedback input when rejecting plans, allowing users to tell Claude what to change
402- VSCode: Added streaming message support for real-time response display
403
404## 2.0.56
405
406- Added setting to enable/disable terminal progress bar (OSC 9;4)
407- 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.
408
409## 2.0.55
410
411- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
412- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
413- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
414- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
415
416## 2.0.54
417
418- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
419- Fix issue with excessive iTerm notifications
420
421## 2.0.52
422
423- Fixed duplicate message display when starting Claude with a command line argument
424- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
425- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
426- Permit some uses of `$!` in bash commands
427
428## 2.0.51
429
430- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
431- Introducing Claude Code for Desktop: https://claude.com/download
432- 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
433- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
434- Plan Mode now builds more precise plans and executes more thoroughly
435- Usage limit notifications now easier to understand
436- Switched `/usage` back to "% used"
437- Fixed handling of thinking errors
438- Fixed performance regression
439
440## 2.0.50
441
442- Fixed bug preventing calling MCP tools that have nested references in their input schemas
443- Silenced a noisy but harmless error during upgrades
444- Improved ultrathink text display
445- Improved clarity of 5-hour session limit warning message
446
447## 2.0.49
448
449- Added readline-style ctrl-y for pasting deleted text
450- Improved clarity of usage limit warning message
451- Fixed handling of subagent permissions
452
453## 2.0.47
454
455- Improved error messages and validation for `claude --teleport`
456- Improved error handling in `/usage`
457- Fixed race condition with history entry not getting logged at exit
458- Fixed Vertex AI configuration not being applied from `settings.json`
459
460## 2.0.46
461
462- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
463
464## 2.0.45
465
466- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
467- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
468- Send background tasks to Claude Code on the web by starting a message with `&`
469
470## 2.0.43
471
472- Added `permissionMode` field for custom agents
473- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
474- Added skills frontmatter field to declare skills to auto-load for subagents
475- Added the `SubagentStart` hook event
476- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
477- Fixed duplicate rendering of some messages in the UI
478- Fixed some visual flickers
479- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
480
481## 2.0.42
482
483- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
484
485## 2.0.41
486
487- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
488- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
489- Fixed incorrect labeling of user settings vs project settings in command descriptions
490- Fixed crash when plugin command hooks timeout during execution
491- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
492- Fixed broken security documentation links in trust dialogs and onboarding
493- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
494- ctrl-r history search landing on a slash command no longer cancels the search
495- SDK: Support custom timeouts for hooks
496- Allow more safe git commands to run without approval
497- Plugins: Added support for sharing and installing output styles
498- Teleporting a session from web will automatically set the upstream branch
499
500## 2.0.37
501
502- Fixed how idleness is computed for notifications
503- Hooks: Added matcher values for Notification hook events
504- Output Styles: Added `keep-coding-instructions` option to frontmatter
505
506## 2.0.36
507
508- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
509- Fixed queued messages being incorrectly executed as bash commands
510- Fixed input being lost when typing while a queued message is processed
511
512## 2.0.35
513
514- Improve fuzzy search results when searching commands
515- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
516- 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
517- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
518- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
519
520## 2.0.34
521
522- VSCode Extension: Added setting to configure the initial permission mode for new conversations
523- Improved file path suggestion performance with native Rust-based fuzzy finder
524- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
525- Fixed memory crash when reading or writing large files (especially base64-encoded images)
526
527## 2.0.33
528
529- Native binary installs now launch quicker.
530- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
531- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
532
533## 2.0.32
534
535- Un-deprecate output styles based on community feedback
536- Added `companyAnnouncements` setting for displaying announcements on startup
537- Fixed hook progress messages not updating correctly during PostToolUse hook execution
538
539## 2.0.31
540
541- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
542- Vertex: add support for Web Search on supported models
543- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
544- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
545- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
546- Fixed plugin uninstall not removing plugins
547
548## 2.0.30
549
550- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
551- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
552- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
553- Added prompt-based stop hooks
554- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
555- Enabled SSE MCP servers on native build
556- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
557- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
558- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
559- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
560- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
561- Fixed bug that saved session permissions to local settings
562- Fixed MCP tools not being available to sub-agents
563- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
564- Fixed delay when navigating through typeahead suggestions with arrow keys
565- VSCode: Restored selection indicator in input footer showing current file or code selection status
566
567## 2.0.28
568
569- Plan mode: introduced new Plan subagent
570- Subagents: claude can now choose to resume subagents
571- Subagents: claude can dynamically choose the model used by its subagents
572- SDK: added --max-budget-usd flag
573- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
574- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
575- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
576- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
577- Various other bug fixes
578
579## 2.0.27
580
581- New UI for permission prompts
582- Added current branch filtering and search to session resume screen for easier navigation
583- Fixed directory @-mention causing "No assistant message found" error
584- VSCode Extension: Add config setting to include .gitignored files in file searches
585- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
586
587## 2.0.25
588
589- 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
590
591## 2.0.24
592
593- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
594- Claude Code Web: Support for Web -> CLI teleport
595- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
596- Bedrock: Display awsAuthRefresh output when auth is required
597
598## 2.0.22
599
600- Fixed content layout shift when scrolling through slash commands
601- IDE: Add toggle to enable/disable thinking.
602- Fix bug causing duplicate permission prompts with parallel tool calls
603- Add support for enterprise managed MCP allowlist and denylist
604
605## 2.0.21
606
607- Support MCP `structuredContent` field in tool responses
608- Added an interactive question tool
609- Claude will now ask you questions more often in plan mode
610- Added Haiku 4.5 as a model option for Pro users
611- Fixed an issue where queued commands don't have access to previous messages' output
612
613## 2.0.20
614
615- Added support for Claude Skills
616
617## 2.0.19
618
619- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
620- Fixed a bug where Haiku was unnecessarily called in print mode
621
622## 2.0.17
623
624- Added Haiku 4.5 to model selector!
625- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
626- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
627- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
628- OTEL: support HTTP_PROXY and HTTPS_PROXY
629- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
630
631## 2.0.15
632
633- Fixed bug with resuming where previously created files needed to be read again before writing
634- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
635
636## 2.0.14
637
638- Fix @-mentioning MCP servers to toggle them on/off
639- Improve permission checks for bash with inline env vars
640- Fix ultrathink + thinking toggle
641- Reduce unnecessary logins
642- Document --system-prompt
643- Several improvements to rendering
644- Plugins UI polish
645
646## 2.0.13
647
648- Fixed `/plugin` not working on native build
649
650## 2.0.12
651
652- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
653- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
654- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
655- `/plugin validate` command for validating plugin structure and configuration
656- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
657- Plugin documentation available at https://code.claude.com/docs/en/plugins
658- Comprehensive error messages and diagnostics via `/doctor` command
659- Avoid flickering in `/model` selector
660- Improvements to `/help`
661- Avoid mentioning hooks in `/resume` summaries
662- Changes to the "verbose" setting in `/config` now persist across sessions
663
664## 2.0.11
665
666- Reduced system prompt size by 1.4k tokens
667- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
668- Fixed Opus fallback rate limit errors appearing incorrectly
669- Fixed /add-dir command selecting wrong default tab
670
671## 2.0.10
672
673- Rewrote terminal renderer for buttery smooth UI
674- Enable/disable MCP servers by @mentioning, or in /mcp
675- Added tab completion for shell commands in bash mode
676- PreToolUse hooks can now modify tool inputs
677- Press Ctrl-G to edit your prompt in your system's configured text editor
678- Fixes for bash permission checks with environment variables in the command
679
680## 2.0.9
681
682- Fix regression where bash backgrounding stopped working
683
684## 2.0.8
685
686- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
687- IDE: Add drag-and-drop support for files and folders in chat
688- /context: Fix counting for thinking blocks
689- Improve message rendering for users with light themes on dark terminals
690- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
691
692## 2.0.5
693
694- IDE: Fix IME unintended message submission with Enter and Tab
695- IDE: Add "Open in Terminal" link in login screen
696- Fix unhandled OAuth expiration 401 API errors
697- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
698
699## 2.0.1
700
701- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
702- Various bug fixes and presentation improvements
703
704## 2.0.0
705
706- New native VS Code extension
707- Fresh coat of paint throughout the whole app
708- /rewind a conversation to undo code changes
709- /usage command to see plan limits
710- Tab to toggle thinking (sticky across sessions)
711- Ctrl-R to search history
712- Unshipped claude config command
713- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
714- SDK: The Claude Code SDK is now the Claude Agent SDK
715- Add subagents dynamically with `--agents` flag
716
717## 1.0.126
718
719- Enable /context command for Bedrock and Vertex
720- Add mTLS support for HTTP-based OpenTelemetry exporters
721
722## 1.0.124
723
724- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
725- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
726- No longer inform Claude of the list of allowed tools when permission is denied
727- Fixed security vulnerability in Bash tool permission checks
728- Improved VSCode extension performance for large files
729
730## 1.0.123
731
732- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
733- Fixed thinking mode triggering on negation phrases like "don't think"
734- Fixed rendering performance degradation during token streaming
735- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
736- Enhanced BashTool environment snapshot logging
737- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
738- Migrated --debug logging to a file, to enable easy tailing & filtering
739
740## 1.0.120
741
742- Fix input lag during typing, especially noticeable with large prompts
743- Improved VSCode extension command registry and sessions dialog user experience
744- Enhanced sessions dialog responsiveness and visual feedback
745- Fixed IDE compatibility issue by removing worktree support check
746- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
747
748## 1.0.119
749
750- Fix Windows issue where process visually freezes on entering interactive mode
751- Support dynamic headers for MCP servers via headersHelper configuration
752- Fix thinking mode not working in headless sessions
753- Fix slash commands now properly update allowed tools instead of replacing them
754
755## 1.0.117
756
757- Add Ctrl-R history search to recall previous commands like bash/zsh
758- Fix input lag while typing, especially on Windows
759- Add sed command to auto-allowed commands in acceptEdits mode
760- Fix Windows PATH comparison to be case-insensitive for drive letters
761- Add permissions management hint to /add-dir output
762
763## 1.0.115
764
765- Improve thinking mode display with enhanced visual effects
766- Type /t to temporarily disable thinking mode in your prompt
767- Improve path validation for glob and grep tools
768- Show condensed output for post-tool hooks to reduce visual clutter
769- Fix visual feedback when loading state completes
770- Improve UI consistency for permission request dialogs
771
772## 1.0.113
773
774- Deprecated piped input in interactive mode
775- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
776
777## 1.0.112
778
779- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
780- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
781- Hooks: Added systemMessage support for SessionEnd hooks
782- Added `spinnerTipsEnabled` setting to disable spinner tips
783- IDE: Various improvements and bug fixes
784
785## 1.0.111
786
787- /model now validates provided model names
788- Fixed Bash tool crashes caused by malformed shell syntax parsing
789
790## 1.0.110
791
792- /terminal-setup command now supports WezTerm
793- MCP: OAuth tokens now proactively refresh before expiration
794- Fixed reliability issues with background Bash processes
795
796## 1.0.109
797
798- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
799
800## 1.0.106
801
802- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
803
804## 1.0.97
805
806- Settings: /doctor now validates permission rule syntax and suggests corrections
807
808## 1.0.94
809
810- Vertex: add support for global endpoints for supported models
811- /memory command now allows direct editing of all imported memory files
812- SDK: Add custom tools as callbacks
813- Added /todos command to list current todo items
814
815## 1.0.93
816
817- Windows: Add alt + v shortcut for pasting images from clipboard
818- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
819
820## 1.0.90
821
822- Settings file changes take effect immediately - no restart required
823
824## 1.0.88
825
826- Fixed issue causing "OAuth authentication is currently not supported"
827- Status line input now includes `exceeds_200k_tokens`
828- Fixed incorrect usage tracking in /cost.
829- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
830- Bedrock: Updated default Sonnet model to Sonnet 4
831
832## 1.0.86
833
834- Added /context to help users self-serve debug context issues
835- SDK: Added UUID support for all SDK messages
836- SDK: Added `--replay-user-messages` to replay user messages back to stdout
837
838## 1.0.85
839
840- Status line input now includes session cost info
841- Hooks: Introduced SessionEnd hook
842
843## 1.0.84
844
845- Fix tool_use/tool_result id mismatch error when network is unstable
846- Fix Claude sometimes ignoring real-time steering when wrapping up a task
847- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
848- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
849
850## 1.0.83
851
852- @-mention: Support files with spaces in path
853- New shimmering spinner
854
855## 1.0.82
856
857- SDK: Add request cancellation support
858- SDK: New additionalDirectories option to search custom paths, improved slash command processing
859- Settings: Validation prevents invalid fields in .claude/settings.json files
860- MCP: Improve tool name consistency
861- Bash: Fix crash when Claude tries to automatically read large files
862
863## 1.0.81
864
865- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
866- Agents: Fix custom agent loading when agent files are unparsable
867
868## 1.0.80
869
870- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
871
872## 1.0.77
873
874- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
875- SDK: Add session support and permission denial tracking
876- Fix token limit errors in conversation summarization
877- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
878
879## 1.0.73
880
881- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
882- MCP: Press Esc to cancel OAuth authentication flows
883- Bash: Improved command validation and reduced false security warnings
884- UI: Enhanced spinner animations and status line visual hierarchy
885- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
886
887## 1.0.72
888
889- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
890
891## 1.0.71
892
893- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
894- Customizable status line: add your terminal prompt to Claude Code with /statusline
895
896## 1.0.70
897
898- Performance: Optimized message rendering for better performance with large contexts
899- Windows: Fixed native file search, ripgrep, and subagent functionality
900- Added support for @-mentions in slash command arguments
901
902## 1.0.69
903
904- Upgraded Opus to version 4.1
905
906## 1.0.68
907
908- Fix incorrect model names being used for certain commands like `/pr-comments`
909- 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.
910- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
911- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
912- SDK: Added canUseTool callback support for tool confirmation
913- Added `disableAllHooks` setting
914- Improved file suggestions performance in large repos
915
916## 1.0.65
917
918- IDE: Fixed connection stability issues and error handling for diagnostics
919- Windows: Fixed shell environment setup for users without .bashrc files
920
921## 1.0.64
922
923- Agents: Added model customization support - you can now specify which model an agent should use
924- Agents: Fixed unintended access to the recursive agent tool
925- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
926- SDK: Fixed user input tracking across multi-turn conversations
927- Added hidden files to file search and @-mention suggestions
928
929## 1.0.63
930
931- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
932
933## 1.0.62
934
935- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
936- Hooks: Added SessionStart hook for new session initialization
937- /add-dir command now supports typeahead for directory paths
938- Improved network connectivity check reliability
939
940## 1.0.61
941
942- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
943- Settings: Added `--settings` flag to load settings from a JSON file
944- Settings: Fixed resolution of settings files paths that are symlinks
945- OTEL: Fixed reporting of wrong organization after authentication changes
946- Slash commands: Fixed permissions checking for allowed-tools with Bash
947- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
948- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
949- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
950
951## 1.0.60
952
953- You can now create custom subagents for specialized tasks! Run /agents to get started
954
955## 1.0.59
956
957- SDK: Added tool confirmation support with canUseTool callback
958- SDK: Allow specifying env for spawned process
959- Hooks: Exposed PermissionDecision to hooks (including "ask")
960- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
961- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
962
963## 1.0.58
964
965- Added support for reading PDFs
966- MCP: Improved server health status display in 'claude mcp list'
967- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
968
969## 1.0.57
970
971- Added support for specifying a model in slash commands
972- Improved permission messages to help Claude understand allowed tools
973- Fix: Remove trailing newlines from bash output in terminal wrapping
974
975## 1.0.56
976
977- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
978- Fixes for WSL IDE detection
979- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
980
981## 1.0.55
982
983- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
984- Windows: fixed Ctrl+Z crash
985- SDK: Added ability to capture error logging
986- Add --system-prompt-file option to override system prompt in print mode
987
988## 1.0.54
989
990- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
991- Custom slash commands: Added argument-hint to frontmatter
992- Windows: OAuth uses port 45454 and properly constructs browser URL
993- Windows: mode switching now uses alt + m, and plan mode renders properly
994- Shell: Switch to in-memory shell snapshot to fix file-related errors
995
996## 1.0.53
997
998- Updated @-mention file truncation from 100 lines to 2000 lines
999- 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).
1000
1001## 1.0.52
1002
1003- Added support for MCP server instructions
1004
1005## 1.0.51
1006
1007- Added support for native Windows (requires Git for Windows)
1008- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1009- Settings: /doctor can now help you identify and fix invalid setting files
1010- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1011- Increased auto-compact warning threshold from 60% to 80%
1012- Fixed an issue with handling user directories with spaces for shell snapshots
1013- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1014- Custom slash commands: Fixed user-level commands in subdirectories
1015- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1016
1017## 1.0.48
1018
1019- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1020- Added progress messages to Bash tool based on the last 5 lines of command output
1021- Added expanding variables support for MCP server configuration
1022- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1023- Improved IDE extension path handling when Claude Code runs in WSL
1024- Hooks: Added a PreCompact hook
1025- Vim mode: Added c, f/F, t/T
1026
1027## 1.0.45
1028
1029- Redesigned Search (Grep) tool with new tool input parameters and features
1030- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1031- Fixed config file corruption issue by enforcing atomic writes
1032- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1033- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1034- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1035
1036## 1.0.44
1037
1038- New /export command lets you quickly export a conversation for sharing
1039- MCP: resource_link tool results are now supported
1040- MCP: tool annotations and tool titles now display in /mcp view
1041- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1042
1043## 1.0.43
1044
1045- Fixed a bug where the theme selector was saving excessively
1046- Hooks: Added EPIPE system error handling
1047
1048## 1.0.42
1049
1050- Added tilde (`~`) expansion support to `/add-dir` command
1051
1052## 1.0.41
1053
1054- Hooks: Split Stop hook triggering into Stop and SubagentStop
1055- Hooks: Enabled optional timeout configuration for each command
1056- Hooks: Added "hook_event_name" to hook input
1057- Fixed a bug where MCP tools would display twice in tool list
1058- New tool parameters JSON for Bash tool in `tool_decision` event
1059
1060## 1.0.40
1061
1062- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1063
1064## 1.0.39
1065
1066- New Active Time metric in OpenTelemetry logging
1067
1068## 1.0.38
1069
1070- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1071
1072## 1.0.37
1073
1074- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1075
1076## 1.0.36
1077
1078- Web search now takes today's date into context
1079- Fixed a bug where stdio MCP servers were not terminating properly on exit
1080
1081## 1.0.35
1082
1083- Added support for MCP OAuth Authorization Server discovery
1084
1085## 1.0.34
1086
1087- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1088
1089## 1.0.33
1090
1091- Improved logging functionality with session ID support
1092- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1093- Improvements to plan mode
1094
1095## 1.0.32
1096
1097- Updated loopback config for litellm
1098- Added forceLoginMethod setting to bypass login selection screen
1099
1100## 1.0.31
1101
1102- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1103
1104## 1.0.30
1105
1106- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1107- Improved file path autocomplete with filename matching
1108- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1109- Enhanced jq regex support for complex filters with pipes and select
1110
1111## 1.0.29
1112
1113- Improved CJK character support in cursor navigation and rendering
1114
1115## 1.0.28
1116
1117- Slash commands: Fix selector display during history navigation
1118- Resizes images before upload to prevent API size limit errors
1119- Added XDG_CONFIG_HOME support to configuration directory
1120- Performance optimizations for memory usage
1121- New attributes (terminal.type, language) in OpenTelemetry logging
1122
1123## 1.0.27
1124
1125- Streamable HTTP MCP servers are now supported
1126- Remote MCP servers (SSE and HTTP) now support OAuth
1127- MCP resources can now be @-mentioned
1128- /resume slash command to switch conversations within Claude Code
1129
1130## 1.0.25
1131
1132- Slash commands: moved "project" and "user" prefixes to descriptions
1133- Slash commands: improved reliability for command discovery
1134- Improved support for Ghostty
1135- Improved web search reliability
1136
1137## 1.0.24
1138
1139- Improved /mcp output
1140- Fixed a bug where settings arrays got overwritten instead of merged
1141
1142## 1.0.23
1143
1144- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1145- Released Python SDK: pip install claude-code-sdk to get started
1146
1147## 1.0.22
1148
1149- SDK: Renamed `total_cost` to `total_cost_usd`
1150
1151## 1.0.21
1152
1153- Improved editing of files with tab-based indentation
1154- Fix for tool_use without matching tool_result errors
1155- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1156
1157## 1.0.18
1158
1159- Added --add-dir CLI argument for specifying additional working directories
1160- Added streaming input support without require -p flag
1161- Improved startup performance and session storage performance
1162- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1163- Added detailed MCP server tools display (/mcp)
1164- MCP authentication and permission improvements
1165- Added auto-reconnection for MCP SSE connections on disconnect
1166- Fixed issue where pasted content was lost when dialogs appeared
1167
1168## 1.0.17
1169
1170- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1171- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1172- MCP server list UI improvements
1173- Update Claude Code process title to display "claude" instead of "node"
1174
1175## 1.0.11
1176
1177- Claude Code can now also be used with a Claude Pro subscription
1178- Added /upgrade for smoother switching to Claude Max plans
1179- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1180- Improved shell configuration error handling
1181- Improved todo list handling during compaction
1182
1183## 1.0.10
1184
1185- Added markdown table support
1186- Improved streaming performance
1187
1188## 1.0.8
1189
1190- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1191- Increased default otel interval from 1s -> 5s
1192- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1193- Fixed a regression where search tools unnecessarily asked for permissions
1194- Added support for triggering thinking non-English languages
1195- Improved compacting UI
1196
1197## 1.0.7
1198
1199- Renamed /allowed-tools -> /permissions
1200- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1201- Deprecated claude config commands in favor of editing settings.json
1202- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1203- Improved error handling for /install-github-app
1204- Bugfixes, UI polish, and tool reliability improvements
1205
1206## 1.0.6
1207
1208- Improved edit reliability for tab-indented files
1209- Respect CLAUDE_CONFIG_DIR everywhere
1210- Reduced unnecessary tool permission prompts
1211- Added support for symlinks in @file typeahead
1212- Bugfixes, UI polish, and tool reliability improvements
1213
1214## 1.0.4
1215
1216- Fixed a bug where MCP tool errors weren't being parsed correctly
1217
1218## 1.0.1
1219
1220- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1221- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1222- Updated documentation links and OAuth process descriptions
1223
1224## 1.0.0
1225
1226- Claude Code is now generally available
1227- Introducing Sonnet 4 and Opus 4 models
1228
1229## 0.2.125
1230
1231- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
1232- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
1233
1234## 0.2.117
1235
1236- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
1237- Introduced settings.cleanupPeriodDays
1238- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
1239- Introduced --debug mode
1240
1241## 0.2.108
1242
1243- You can now send messages to Claude while it works to steer Claude in real-time
1244- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
1245- Fixed a bug where thinking was not working in -p mode
1246- Fixed a regression in /cost reporting
1247- Deprecated MCP wizard interface in favor of other MCP commands
1248- Lots of other bugfixes and improvements
1249
1250## 0.2.107
1251
1252- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
1253
1254## 0.2.106
1255
1256- MCP SSE server configs can now specify custom headers
1257- Fixed a bug where MCP permission prompt didn't always show correctly
1258
1259## 0.2.105
1260
1261- Claude can now search the web
1262- Moved system & account status to /status
1263- Added word movement keybindings for Vim
1264- Improved latency for startup, todo tool, and file edits
1265
1266## 0.2.102
1267
1268- Improved thinking triggering reliability
1269- Improved @mention reliability for images and folders
1270- You can now paste multiple large chunks into one prompt
1271
1272## 0.2.100
1273
1274- Fixed a crash caused by a stack overflow error
1275- Made db storage optional; missing db support disables --continue and --resume
1276
1277## 0.2.98
1278
1279- Fixed an issue where auto-compact was running twice
1280
1281## 0.2.96
1282
1283- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
1284
1285## 0.2.93
1286
1287- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
1288- Claude now has access to a Todo list that helps it stay on track and be more organized
1289
1290## 0.2.82
1291
1292- Added support for --disallowedTools
1293- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
1294
1295## 0.2.75
1296
1297- Hit Enter to queue up additional messages while Claude is working
1298- Drag in or copy/paste image files directly into the prompt
1299- @-mention files to directly add them to context
1300- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
1301- Improved performance for filename auto-complete
1302
1303## 0.2.74
1304
1305- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
1306- Task tool can now perform writes and run bash commands
1307
1308## 0.2.72
1309
1310- Updated spinner to indicate tokens loaded and tool usage
1311
1312## 0.2.70
1313
1314- Network commands like curl are now available for Claude to use
1315- Claude can now run multiple web queries in parallel
1316- Pressing ESC once immediately interrupts Claude in Auto-accept mode
1317
1318## 0.2.69
1319
1320- Fixed UI glitches with improved Select component behavior
1321- Enhanced terminal output display with better text truncation logic
1322
1323## 0.2.67
1324
1325- Shared project permission rules can be saved in .claude/settings.json
1326
1327## 0.2.66
1328
1329- Print mode (-p) now supports streaming output via --output-format=stream-json
1330- Fixed issue where pasting could trigger memory or bash mode unexpectedly
1331
1332## 0.2.63
1333
1334- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
1335
1336## 0.2.61
1337
1338- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
1339- Enhanced image detection for more reliable clipboard paste functionality
1340- Fixed an issue where ESC key could crash the conversation history selector
1341
1342## 0.2.59
1343
1344- Copy+paste images directly into your prompt
1345- Improved progress indicators for bash and fetch tools
1346- Bugfixes for non-interactive mode (-p)
1347
1348## 0.2.54
1349
1350- Quickly add to Memory by starting your message with '#'
1351- Press ctrl+r to see full output for long tool results
1352- Added support for MCP SSE transport
1353
1354## 0.2.53
1355
1356- New web fetch tool lets Claude view URLs that you paste in
1357- Fixed a bug with JPEG detection
1358
1359## 0.2.50
1360
1361- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
1362
1363## 0.2.49
1364
1365- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
1366
1367## 0.2.47
1368
1369- Press Tab to auto-complete file and folder names
1370- Press Shift + Tab to toggle auto-accept for file edits
1371- Automatic conversation compaction for infinite conversation length (toggle with /config)
1372
1373## 0.2.44
1374
1375- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
1376
1377## 0.2.41
1378
1379- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
1380- MCP server startup no longer blocks the app from starting up
1381
1382## 0.2.37
1383
1384- New /release-notes command lets you view release notes at any time
1385- `claude config add/remove` commands now accept multiple values separated by commas or spaces
1386
1387## 0.2.36
1388
1389- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
1390- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
1391
1392## 0.2.34
1393
1394- Vim bindings for text input - enable with /vim or /config
1395
1396## 0.2.32
1397
1398- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
1399- Fix for some PersistentShell issues
1400
1401## 0.2.31
1402
1403- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
1404- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
1405
1406## 0.2.30
1407
1408- Added ANSI color theme for better terminal compatibility
1409- Fixed issue where slash command arguments weren't being sent properly
1410- (Mac-only) API keys are now stored in macOS Keychain
1411
1412## 0.2.26
1413
1414- New /approved-tools command for managing tool permissions
1415- Word-level diff display for improved code readability
1416- Fuzzy matching for slash commands
1417
1418## 0.2.21
1419
1420- Fuzzy matching for /commands