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 00553de · on Mar 10, 2026 History
2,189 lines · 138.9 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.72
4
5- Changed tool search to bypass the third-party proxy gate when the environment variable is set (replaces `CLAUDE_CODE_PROXY_SUPPORTS_TOOL_REFERENCE`, now removed)
6- Added `w` key in `/copy` to write the focused selection directly to a file, bypassing the clipboard (useful over SSH)
7- Added optional description argument to `/plan` (e.g., `/plan fix the auth bug`) that enters plan mode and immediately starts
8- Added `claude plugins` as an alias for `claude plugin`
9- Added `ExitWorktree` tool to leave an `EnterWorktree` session
10- Added `CLAUDE_CODE_DISABLE_CRON` environment variable to immediately stop scheduled cron jobs mid-session
11- Added `lsof`, `pgrep`, `tput`, `ss`, `fd`, and `fdfind` to the bash auto-approval allowlist, reducing permission prompts for common read-only operations
12- Added support for marketplace git URLs without `.git` suffix (Azure DevOps, AWS CodeCommit)
13- Restored the `model` parameter on the Agent tool for per-invocation model overrides
14- Simplified effort levels to low/medium/high (removed max) with new symbols (○ ◐ ●) and a brief notification instead of a persistent icon. Use `/effort auto` to reset to default
15- Improved `/config` — Escape now cancels changes, Enter saves and closes, Space toggles settings
16- Improved up-arrow history to show current session's messages first when running multiple concurrent sessions
17- Improved voice input transcription accuracy for repo names and common dev terms (regex, OAuth, JSON)
18- Improved marketplace clone failure messages to show diagnostic info even when git produces no stderr
19- Improved `claude plugin validate` to explain that marketplace.json source paths are relative to the repo root when rejecting `../` paths
20- Improved bash command parsing by switching to a native module — faster initialization and no memory leak
21- Reduced false-positive bash permission prompts — tree-sitter parsing now handles `find -exec`, variable assignments, command substitutions, and many other patterns that previously triggered unnecessary prompts. Also fixed tree-sitter not loading in npm-installed versions
22- Reduced bundle size by ~510 KB
23- Changed CLAUDE.md HTML comments (`<!-- ... -->`) to be hidden from Claude when auto-injected. Comments remain visible when read with the Read tool
24- Fixed slow exits when background tasks or hooks were slow to respond
25- Fixed agent task progress stuck on "Initializing…"
26- Fixed skill hooks firing twice per event when a hooks-enabled skill is invoked by the model
27- Fixed several voice mode issues: occasional input lag, false "No speech detected" errors after releasing push-to-talk, and stale transcripts re-filling the prompt after submission
28- Fixed `--continue` not resuming from the most recent point after `--compact`
29- Fixed bash security parsing edge cases
30- Fixed several plugin issues: installation failing on Windows with `EEXIST` error in OneDrive folders, marketplace blocking user-scope installs when a project-scope install exists, `CLAUDE_CODE_PLUGIN_CACHE_DIR` creating literal `~` directories, and `plugin.json` with marketplace-only fields failing to load
31- Fixed feedback survey appearing too frequently in long sessions
32- Fixed `--effort` CLI flag being reset by unrelated settings writes on startup
33- Fixed backgrounded Ctrl+B queries losing their transcript or corrupting the new conversation after `/clear`
34- Fixed `/clear` killing background agent/bash tasks — only foreground tasks are now cleared
35- Fixed worktree isolation issues: Task tool resume not restoring cwd, and background task notifications missing `worktreePath` and `worktreeBranch`
36- Fixed `/model` not displaying results when run while Claude is working
37- Fixed digit keys selecting menu options instead of typing in plan mode permission prompt's text input
38- Fixed sandbox permission issues: certain file write operations incorrectly allowed without prompting, and output redirections to allowlisted directories (like `/tmp/claude/`) prompting unnecessarily
39- Improved CPU utilization in long sessions
40- Fixed prompt cache invalidation in SDK `query()` calls, reducing input token costs up to 12x
41- Fixed Escape key becoming unresponsive after cancelling a query
42- Fixed double Ctrl+C not exiting when background agents or tasks are running
43- Fixed team agents to inherit the leader's model
44- Fixed "Always Allow" saving permission rules that never match again
45- Fixed several hooks issues: `transcript_path` pointing to the wrong directory for resumed/forked sessions, agent `prompt` being silently deleted from settings.json on every settings write, PostToolUse block reason displaying twice, async hooks not receiving stdin with bash `read -r`, and validation error message showing an example that fails validation
46- Fixed session crashes in Desktop/SDK when Read returned files containing U+2028/U+2029 characters
47- Fixed terminal title being cleared on exit even when `CLAUDE_CODE_DISABLE_TERMINAL_TITLE` was set
48- Fixed several permission rule matching issues: wildcard rules not matching commands with heredocs, embedded newlines, or no arguments; `sandbox.excludedCommands` failing with env var prefixes; "always allow" suggesting overly broad prefixes for nested CLI tools; and deny rules not applying to all command forms
49- Fixed oversized and truncated images from Bash data-URL output
50- Fixed a crash when resuming sessions that contained Bedrock API errors
51- Fixed intermittent "expected boolean, received string" validation errors on Edit, Bash, and Grep tool inputs
52- Fixed multi-line session titles when forking from a conversation whose first message contained newlines
53- Fixed queued messages not showing attached images, and images being lost when pressing ↑ to edit a queued message
54- Fixed parallel tool calls where a failed Read/WebFetch/Glob would cancel its siblings — only Bash errors now cascade
55- VSCode: Fixed scroll speed in integrated terminals not matching native terminals
56- VSCode: Fixed Shift+Enter submitting input instead of inserting a newline for users with older keybindings
57- VSCode: Added effort level indicator on the input border
58- VSCode: Added `vscode://anthropic.claude-code/open` URI handler to open a new Claude Code tab programmatically, with optional `prompt` and `session` query parameters
59
60## 2.1.71
61
62- Added `/loop` command to run a prompt or slash command on a recurring interval (e.g. `/loop 5m check the deploy`)
63- Added cron scheduling tools for recurring prompts within a session
64- Added `voice:pushToTalk` keybinding to make the voice activation key rebindable in `keybindings.json` (default: space) — modifier+letter combos like `meta+k` have zero typing interference
65- Added `fmt`, `comm`, `cmp`, `numfmt`, `expr`, `test`, `printf`, `getconf`, `seq`, `tsort`, and `pr` to the bash auto-approval allowlist
66- Fixed stdin freeze in long-running sessions where keystrokes stop being processed but the process stays alive
67- Fixed a 5–8 second startup freeze for users with voice mode enabled, caused by CoreAudio initialization blocking the main thread after system wake
68- Fixed startup UI freeze when many claude.ai proxy connectors refresh an expired OAuth token simultaneously
69- Fixed forked conversations (`/fork`) sharing the same plan file, which caused plan edits in one fork to overwrite the other
70- Fixed the Read tool putting oversized images into context when image processing failed, breaking subsequent turns in long image-heavy sessions
71- Fixed false-positive permission prompts for compound bash commands containing heredoc commit messages
72- Fixed plugin installations being lost when running multiple Claude Code instances
73- Fixed claude.ai connectors failing to reconnect after OAuth token refresh
74- Fixed claude.ai MCP connector startup notifications appearing for every org-configured connector instead of only previously connected ones
75- Fixed background agent completion notifications missing the output file path, which made it difficult for parent agents to recover agent results after context compaction
76- Fixed duplicate output in Bash tool error messages when commands exit with non-zero status
77- Fixed Chrome extension auto-detection getting permanently stuck on "not installed" after running on a machine without local Chrome
78- Fixed `/plugin marketplace update` failing with merge conflicts when the marketplace is pinned to a branch/tag ref
79- Fixed `/plugin marketplace add owner/repo@ref` incorrectly parsing `@` — previously only `#` worked as a ref separator, causing undiagnosable errors with `strictKnownMarketplaces`
80- Fixed duplicate entries in `/permissions` Workspace tab when the same directory is added with and without a trailing slash
81- Fixed `--print` hanging forever when team agents are configured — the exit loop no longer waits on long-lived `in_process_teammate` tasks
82- Fixed "❯ Tool loaded." appearing in the REPL after every `ToolSearch` call
83- Fixed prompting for `cd <cwd> && git ...` on Windows when the model uses a mingw-style path
84- Improved startup time by deferring native image processor loading to first use
85- Improved bridge session reconnection to complete within seconds after laptop wake from sleep, instead of waiting up to 10 minutes
86- Improved `/plugin uninstall` to disable project-scoped plugins in `.claude/settings.local.json` instead of modifying `.claude/settings.json`, so changes don't affect teammates
87- Improved plugin-provided MCP server deduplication — servers that duplicate a manually-configured server (same command/URL) are now skipped, preventing duplicate connections and tool sets. Suppressions are shown in the `/plugin` menu.
88- Updated `/debug` to toggle debug logging on mid-session, since debug logs are no longer written by default
89- Removed startup notification noise for unauthenticated org-registered claude.ai connectors
90
91## 2.1.70
92
93- Fixed API 400 errors when using `ANTHROPIC_BASE_URL` with a third-party gateway — tool search now correctly detects proxy endpoints and disables `tool_reference` blocks
94- Fixed `API Error: 400 This model does not support the effort parameter` when using custom Bedrock inference profiles or other model identifiers not matching standard Claude naming patterns
95- Fixed empty model responses immediately after `ToolSearch` — the server renders tool schemas with system-prompt-style tags at the prompt tail, which could confuse models into stopping early
96- Fixed prompt-cache bust when an MCP server with `instructions` connects after the first turn
97- Fixed Enter inserting a newline instead of submitting when typing over a slow SSH connection
98- Fixed clipboard corrupting non-ASCII text (CJK, emoji) on Windows/WSL by using PowerShell `Set-Clipboard`
99- Fixed extra VS Code windows opening at startup on Windows when running from the VS Code integrated terminal
100- Fixed voice mode failing on Windows native binary with "native audio module could not be loaded"
101- Fixed push-to-talk not activating on session start when `voiceEnabled: true` was set in settings
102- Fixed markdown links containing `#NNN` references incorrectly pointing to the current repository instead of the linked URL
103- Fixed repeated "Model updated to Opus 4.6" notification when a project's `.claude/settings.json` has a legacy Opus model string pinned
104- Fixed plugins showing as inaccurately installed in `/plugin`
105- Fixed plugins showing "not found in marketplace" errors on fresh startup by auto-refreshing after marketplace installation
106- Fixed `/security-review` command failing with `unknown option merge-base` on older git versions
107- Fixed `/color` command having no way to reset back to the default color — `/color default`, `/color gray`, `/color reset`, and `/color none` now restore the default
108- Fixed a performance regression in the `AskUserQuestion` preview dialog that re-ran markdown rendering on every keystroke in the notes input
109- Fixed feature flags read during early startup never refreshing their disk cache, causing stale values to persist across sessions
110- Fixed `permissions.defaultMode` settings values other than `acceptEdits` or `plan` being applied in Claude Code Remote environments — they are now ignored
111- Fixed skill listing being re-injected on every `--resume` (~600 tokens saved per resume)
112- Fixed teleport marker not rendering in VS Code teleported sessions
113- Improved error message when microphone captures silence to distinguish from "no speech detected"
114- Improved compaction to preserve images in the summarizer request, allowing prompt cache reuse for faster and cheaper compaction
115- Improved `/rename` to work while Claude is processing, instead of being silently queued
116- Reduced prompt input re-renders during turns by ~74%
117- Reduced startup memory by ~426KB for users without custom CA certificates
118- Reduced Remote Control `/poll` rate to once per 10 minutes while connected (was 1–2s), cutting server load ~300×. Reconnection is unaffected — transport loss immediately wakes fast polling.
119- [VSCode] Added spark icon in VS Code activity bar that lists all Claude Code sessions, with sessions opening as full editors
120- [VSCode] Added full markdown document view for plans in VS Code, with support for adding comments to provide feedback
121- [VSCode] Added native MCP server management dialog — use `/mcp` in the chat panel to enable/disable servers, reconnect, and manage OAuth authentication without switching to the terminal
122
123## 2.1.69
124
125- Added the `/claude-api` skill for building applications with the Claude API and Anthropic SDK
126- Added Ctrl+U on an empty bash prompt (`!`) to exit bash mode, matching `escape` and `backspace`
127- Added numeric keypad support for selecting options in Claude's interview questions (previously only the number row above QWERTY worked)
128- Added optional name argument to `/remote-control` and `claude remote-control` (`/remote-control My Project` or `--name "My Project"`) to set a custom session title visible in claude.ai/code
129- Added Voice STT support for 10 new languages (20 total) — Russian, Polish, Turkish, Dutch, Ukrainian, Greek, Czech, Danish, Swedish, Norwegian
130- Added effort level display (e.g., "with low effort") to the logo and spinner, making it easier to see which effort setting is active
131- Added agent name display in terminal title when using `claude --agent`
132- Added `sandbox.enableWeakerNetworkIsolation` setting (macOS only) to allow Go programs like `gh`, `gcloud`, and `terraform` to verify TLS certificates when using a custom MITM proxy with `httpProxyPort`
133- Added `includeGitInstructions` setting (and `CLAUDE_CODE_DISABLE_GIT_INSTRUCTIONS` env var) to remove built-in commit and PR workflow instructions from Claude's system prompt
134- Added `/reload-plugins` command to activate pending plugin changes without restarting
135- Added a one-time startup prompt suggesting Claude Code Desktop on macOS and Windows (max 3 showings, dismissible)
136- Added `${CLAUDE_SKILL_DIR}` variable for skills to reference their own directory in SKILL.md content
137- Added `InstructionsLoaded` hook event that fires when CLAUDE.md or `.claude/rules/*.md` files are loaded into context
138- Added `agent_id` (for subagents) and `agent_type` (for subagents and `--agent`) to hook events
139- Added `worktree` field to status line hook commands with name, path, branch, and original repo directory when running in a `--worktree` session
140- Added `pluginTrustMessage` in managed settings to append organization-specific context to the plugin trust warning shown before installation
141- Added policy limit fetching (e.g., remote control restrictions) for Team plan OAuth users, not just Enterprise
142- Added `pathPattern` to `strictKnownMarketplaces` for regex-matching file/directory marketplace sources alongside `hostPattern` restrictions
143- Added plugin source type `git-subdir` to point to a subdirectory within a git repo
144- Added `oauth.authServerMetadataUrl` config option for MCP servers to specify a custom OAuth metadata discovery URL when standard discovery fails
145- Fixed a security issue where nested skill discovery could load skills from gitignored directories like `node_modules`
146- Fixed trust dialog silently enabling all `.mcp.json` servers on first run. You'll now see the per-server approval dialog as expected
147- Fixed `claude remote-control` crashing immediately on npm installs with "bad option: --sdk-url" (anthropics/claude-code#28334)
148- Fixed `--model claude-opus-4-0` and `--model claude-opus-4-1` resolving to deprecated Opus versions instead of current
149- Fixed macOS keychain corruption when using multiple OAuth MCP servers. Large OAuth metadata blobs could overflow the `security -i` stdin buffer, silently leaving stale credentials behind and causing repeated `/login` prompts.
150- Fixed `.credentials.json` losing `subscriptionType` (showing "Claude API" instead of "Claude Pro"/"Claude Max") when the profile endpoint transiently fails during token refresh (anthropics/claude-code#30185)
151- Fixed ghost dotfiles (`.bashrc`, `HEAD`, etc.) appearing as untracked files in the working directory after sandboxed Bash commands on Linux
152- Fixed Shift+Enter printing `[27;2;13~` instead of inserting a newline in Ghostty over SSH
153- Fixed stash (Ctrl+S) being cleared when submitting a message while Claude is working
154- Fixed ctrl+o (transcript toggle) freezing for many seconds in long sessions with lots of file edits
155- Fixed plan mode feedback input not supporting multi-line text entry (backslash+Enter and Shift+Enter now insert newlines)
156- Fixed cursor not moving down into blank lines at the top of the input box
157- Fixed `/stats` crash when transcript files contain entries with missing or malformed timestamps
158- Fixed a brief hang after a streaming error on long sessions (the transcript was being fully rewritten to drop one line; it is now truncated in place)
159- Fixed `--setting-sources user` not blocking dynamically discovered project skills
160- Fixed duplicate CLAUDE.md, slash commands, agents, and rules when running from a worktree nested inside its main repo (e.g. `claude -w`)
161- Fixed plugin Stop/SessionEnd/etc hooks not firing after any `/plugin` operation
162- Fixed plugin hooks being silently dropped when two plugins use the same `${CLAUDE_PLUGIN_ROOT}/...` command template
163- Fixed memory leak in long-running SDK/CCR sessions where conversation messages were retained unnecessarily
164- Fixed API 400 errors in forked agents (autocompact, summarization) when resuming sessions that were interrupted mid-tool-batch
165- Fixed "unexpected tool_use_id found in tool_result blocks" error when resuming conversations that start with an orphaned tool result
166- Fixed teammates accidentally spawning nested teammates via the Agent tool's `name` parameter
167- Fixed `CLAUDE_CODE_MAX_OUTPUT_TOKENS` being ignored during conversation compaction
168- Fixed `/compact` summary rendering as a user bubble in SDK consumers (Claude Code Remote web UI, VSCode extension)
169- Fixed voice space bar getting stuck after a failed voice activation (module loading race, cold GrowthBook)
170- Fixed worktree file copy on Windows
171- Fixed global `.claude` folder detection on Windows
172- Fixed symlink bypass where writing new files through a symlinked parent directory could escape the working directory in `acceptEdits` mode
173- Fixed sandbox prompting users to approve non-allowed domains when `allowManagedDomainsOnly` is enabled in managed settings — non-allowed domains are now blocked automatically with no bypass
174- Fixed interactive tools (e.g., `AskUserQuestion`) being silently auto-allowed when listed in a skill's allowed-tools, bypassing the permission prompt and running with empty answers
175- Fixed multi-GB memory spike when committing with large untracked binary files in the working tree
176- Fixed Escape not interrupting a running turn when the input box has draft text. Use Up arrow to pull queued messages back for editing, or Ctrl+U to clear the input line.
177- Fixed Android app crash when running local slash commands (`/voice`, `/cost`) in Remote Control sessions
178- Fixed a memory leak where old message array versions accumulated in React Compiler `memoCache` over long sessions
179- Fixed a memory leak where REPL render scopes accumulated over long sessions (~35MB over 1000 turns)
180- Fixed memory retention in in-process teammates where the parent's full conversation history was pinned for the teammate's lifetime, preventing GC after `/clear` or auto-compact
181- Fixed a memory leak in interactive mode where hook events could accumulate unboundedly during long sessions
182- Fixed hang when `--mcp-config` points to a corrupted file
183- Fixed slow startup when many skills/plugins are installed
184- Fixed `cd <outside-dir> && <cmd>` permission prompt to surface the chained command instead of only showing "Yes, allow reading from <dir>/"
185- Fixed conditional `.claude/rules/*.md` files (with `paths:` frontmatter) and nested CLAUDE.md files not loading in print mode (`claude -p`)
186- Fixed `/clear` not fully clearing all session caches, reducing memory retention in long sessions
187- Fixed terminal flicker caused by animated elements at the scrollback boundary
188- Fixed UI frame drops on macOS when using MCP servers with OAuth (regression from 2.1.x)
189- Fixed occasional frame stalls during typing caused by synchronous debug log flushes
190- Fixed `TeammateIdle` and `TaskCompleted` hooks to support `{"continue": false, "stopReason": "..."}` to stop the teammate, matching `Stop` hook behavior
191- Fixed `WorktreeCreate` and `WorktreeRemove` plugin hooks being silently ignored
192- Fixed skill descriptions with colons (e.g., "Triggers include: X, Y, Z") failing to load from SKILL.md frontmatter
193- Fixed project skills without a `description:` frontmatter field not appearing in Claude's available skills list
194- Fixed `/context` showing identical token counts for all MCP tools from a server
195- Fixed literal `nul` file creation on Windows when the model uses CMD-style `2>nul` redirection in Git Bash
196- Fixed extra blank lines appearing below each tool call in the expanded subagent transcript view (Ctrl+O)
197- Fixed Tab/arrow keys not cycling Settings tabs when `/config` search box is focused but empty
198- Fixed service key OAuth sessions (CCR containers) spamming `[ERROR]` logs with 403s from profile-scoped endpoints
199- Fixed inconsistent color for "Remote Control active" status indicator
200- Fixed Voice waveform cursor covering the first suffix letter when dictating mid-input
201- Fixed Voice input showing all 5 spaces during warmup instead of capping at ~2 (aligning with the "keep holding…" hint)
202- Improved spinner performance by isolating the 50ms animation loop from the surrounding shell, reducing render and CPU overhead during turns
203- Improved UI rendering performance in native binaries with React Compiler
204- Improved `--worktree` startup by eliminating a git subprocess on the startup path
205- Improved macOS startup by eliminating redundant settings-file reloads when managed settings resolve
206- Improved macOS startup for Claude.ai enterprise/team users by skipping an unnecessary keychain lookup
207- Improved MCP `-p` startup by pipelining claude.ai config fetch with local connections and using a concurrency pool instead of sequential batching
208- Improved voice startup by removing imperceptible warmup pulse animations that were causing re-render stutter
209- Improved MCP binary content handling: tools returning PDFs, Office documents, or audio now save decoded bytes to disk with the correct file extension instead of dumping raw base64 into the conversation context. WebFetch also saves binary responses alongside its summary.
210- Improved memory usage in long sessions by stabilizing `onSubmit` across message updates
211- Improved LSP tool rendering and memory context building to no longer read entire files
212- Improved session upload and memory sync to avoid reading large files into memory before size/binary checks
213- Improved file operation performance by avoiding reading file contents for existence checks (6 sites)
214- Improved documentation to clarify that `--append-system-prompt-file` and `--system-prompt-file` work in interactive mode (the docs previously said print mode only)
215- Reduced baseline memory by ~16MB by deferring Yoga WASM preloading
216- Reduced memory footprint for SDK and CCR sessions using stream-json output
217- Reduced memory usage when resuming large sessions (including compacted history)
218- Reduced token usage on multi-agent tasks with more concise subagent final reports
219- Changed Sonnet 4.5 users on Pro/Max/Team Premium to be automatically migrated to Sonnet 4.6
220- Changed the `/resume` picker to show your most recent prompt instead of the first one. This also resolves some titles appearing as `(session)`.
221- Changed claude.ai MCP connector failures to show a notification instead of silently disappearing from the tool list
222- Changed example command suggestions to be generated deterministically instead of calling Haiku
223- Changed resuming after compaction to no longer produce a preamble recap before continuing
224- [SDK] Changed task creation to no longer require the `activeForm` field — the spinner falls back to the task subject
225- [VSCode] Added compaction display as a collapsible "Compacted chat" card with the summary inside
226- [VSCode] The permission mode picker now respects `permissions.disableBypassPermissionsMode` from your effective Claude Code settings (including managed/policy settings) — when set to `disable`, bypass permissions mode is hidden from the picker
227- [VSCode] Fixed RTL text (Arabic, Hebrew, Persian) rendering reversed in the chat panel (regression in v2.1.63)
228
229## 2.1.68
230
231- Opus 4.6 now defaults to medium effort for Max and Team subscribers. Medium effort works well for most tasks — it's the sweet spot between speed and thoroughness. You can change this anytime with `/model`
232- Re-introduced the "ultrathink" keyword to enable high effort for the next turn
233- Removed Opus 4 and 4.1 from Claude Code on the first-party API — users with these models pinned are automatically moved to Opus 4.6
234
235## 2.1.66
236
237- Reduced spurious error logging
238
239## 2.1.63
240
241- Added `/simplify` and `/batch` bundled slash commands
242- Fixed local slash command output like /cost appearing as user-sent messages instead of system messages in the UI
243- Project configs & auto memory now shared across git worktrees of the same repository
244- Added `ENABLE_CLAUDEAI_MCP_SERVERS=false` env var to opt out from making claude.ai MCP servers available
245- Improved `/model` command to show the currently active model in the slash command menu
246- Added HTTP hooks, which can POST JSON to a URL and receive JSON instead of running a shell command
247- Fixed listener leak in bridge polling loop
248- Fixed listener leak in MCP OAuth flow cleanup
249- Added manual URL paste fallback during MCP OAuth authentication. If the automatic localhost redirect doesn't work, you can paste the callback URL to complete authentication.
250- Fixed memory leak when navigating hooks configuration menu
251- Fixed listener leak in interactive permission handler during auto-approvals
252- Fixed file count cache ignoring glob ignore patterns
253- Fixed memory leak in bash command prefix cache
254- Fixed MCP tool/resource cache leak on server reconnect
255- Fixed IDE host IP detection cache incorrectly sharing results across ports
256- Fixed WebSocket listener leak on transport reconnect
257- Fixed memory leak in git root detection cache that could cause unbounded growth in long-running sessions
258- Fixed memory leak in JSON parsing cache that grew unbounded over long sessions
259- VSCode: Fixed remote sessions not appearing in conversation history
260- Fixed a race condition in the REPL bridge where new messages could arrive at the server interleaved with historical messages during the initial connection flush, causing message ordering issues.
261- Fixed memory leak where long-running teammates retained all messages in AppState even after conversation compaction
262- Fixed a memory leak where MCP server fetch caches were not cleared on disconnect, causing growing memory usage with servers that reconnect frequently
263- Improved memory usage in long sessions with subagents by stripping heavy progress message payloads during context compaction
264- Added "Always copy full response" option to the `/copy` picker. When selected, future `/copy` commands will skip the code block picker and copy the full response directly.
265- VSCode: Added session rename and remove actions to the sessions list
266- Fixed `/clear` not resetting cached skills, which could cause stale skill content to persist in the new conversation
267
268## 2.1.62
269
270- Fixed prompt suggestion cache regression that reduced cache hit rates
271
272## 2.1.61
273
274- Fixed concurrent writes corrupting config file on Windows
275
276## 2.1.59
277
278- Claude automatically saves useful context to auto-memory. Manage with /memory
279- Added `/copy` command to show an interactive picker when code blocks are present, allowing selection of individual code blocks or the full response.
280- Improved "always allow" prefix suggestions for compound bash commands (e.g. `cd /tmp && git fetch && git push`) to compute smarter per-subcommand prefixes instead of treating the whole command as one
281- Improved ordering of short task lists
282- Improved memory usage in multi-agent sessions by releasing completed subagent task state
283- Fixed MCP OAuth token refresh race condition when running multiple Claude Code instances simultaneously
284- Fixed shell commands not showing a clear error message when the working directory has been deleted
285- Fixed config file corruption that could wipe authentication when multiple Claude Code instances ran simultaneously
286
287## 2.1.58
288
289- Expand Remote Control to more users
290
291## 2.1.56
292
293- VS Code: Fixed another cause of "command 'claude-vscode.editor.openLast' not found" crashes
294
295## 2.1.55
296
297- Fixed BashTool failing on Windows with EINVAL error
298
299## 2.1.53
300
301- Fixed a UI flicker where user input would briefly disappear after submission before the message rendered
302- Fixed bulk agent kill (ctrl+f) to send a single aggregate notification instead of one per agent, and to properly clear the command queue
303- Fixed graceful shutdown sometimes leaving stale sessions when using Remote Control by parallelizing teardown network calls
304- Fixed `--worktree` sometimes being ignored on first launch
305- Fixed a panic ("switch on corrupted value") on Windows
306- Fixed a crash that could occur when spawning many processes on Windows
307- Fixed a crash in the WebAssembly interpreter on Linux x64 & Windows x64
308- Fixed a crash that sometimes occurred after 2 minutes on Windows ARM64
309
310## 2.1.52
311
312- VS Code: Fixed extension crash on Windows ("command 'claude-vscode.editor.openLast' not found")
313
314## 2.1.51
315
316- Added `claude remote-control` subcommand for external builds, enabling local environment serving for all users.
317- Updated plugin marketplace default git timeout from 30s to 120s and added `CLAUDE_CODE_PLUGIN_GIT_TIMEOUT_MS` to configure.
318- Added support for custom npm registries and specific version pinning when installing plugins from npm sources
319- BashTool now skips login shell (`-l` flag) by default when a shell snapshot is available, improving command execution performance. Previously this required setting `CLAUDE_BASH_NO_LOGIN=true`.
320- Fixed a security issue where `statusLine` and `fileSuggestion` hook commands could execute without workspace trust acceptance in interactive mode.
321- Tool results larger than 50K characters are now persisted to disk (previously 100K). This reduces context window usage and improves conversation longevity.
322- Fixed a bug where duplicate `control_response` messages (e.g. from WebSocket reconnects) could cause API 400 errors by pushing duplicate assistant messages into the conversation.
323- Added `CLAUDE_CODE_ACCOUNT_UUID`, `CLAUDE_CODE_USER_EMAIL`, and `CLAUDE_CODE_ORGANIZATION_UUID` environment variables for SDK callers to provide account info synchronously, eliminating a race condition where early telemetry events lacked account metadata.
324- Fixed slash command autocomplete crashing when a plugin's SKILL.md description is a YAML array or other non-string type
325- The `/model` picker now shows human-readable labels (e.g., "Sonnet 4.5") instead of raw model IDs for pinned model versions, with an upgrade hint when a newer version is available.
326- Managed settings can now be set via macOS plist or Windows Registry. Learn more at https://code.claude.com/docs/en/settings#settings-files
327
328## 2.1.50
329
330- Added support for `startupTimeout` configuration for LSP servers
331- Added `WorktreeCreate` and `WorktreeRemove` hook events, enabling custom VCS setup and teardown when agent worktree isolation creates or removes worktrees.
332- Fixed a bug where resumed sessions could be invisible when the working directory involved symlinks, because the session storage path was resolved at different times during startup. Also fixed session data loss on SSH disconnect by flushing session data before hooks and analytics in the graceful shutdown sequence.
333- Linux: Fixed native modules not loading on systems with glibc older than 2.30 (e.g., RHEL 8)
334- Fixed memory leak in agent teams where completed teammate tasks were never garbage collected from session state
335- Fixed `CLAUDE_CODE_SIMPLE` to fully strip down skills, session memory, custom agents, and CLAUDE.md token counting
336- Fixed `/mcp reconnect` freezing the CLI when given a server name that doesn't exist
337- Fixed memory leak where completed task state objects were never removed from AppState
338- Added support for `isolation: worktree` in agent definitions, allowing agents to declaratively run in isolated git worktrees.
339- `CLAUDE_CODE_SIMPLE` mode now also disables MCP tools, attachments, hooks, and CLAUDE.md file loading for a fully minimal experience.
340- Fixed bug where MCP tools were not discovered when tool search is enabled and a prompt is passed in as a launch argument
341- Improved memory usage during long sessions by clearing internal caches after compaction
342- Added `claude agents` CLI command to list all configured agents
343- Improved memory usage during long sessions by clearing large tool results after they have been processed
344- Fixed a memory leak where LSP diagnostic data was never cleaned up after delivery, causing unbounded memory growth in long sessions
345- Fixed a memory leak where completed task output was not freed from memory, reducing memory usage in long sessions with many tasks
346- Improved startup performance for headless mode (`-p` flag) by deferring Yoga WASM and UI component imports
347- Fixed prompt suggestion cache regression that reduced cache hit rates
348- Fixed unbounded memory growth in long sessions by capping file history snapshots
349- Added `CLAUDE_CODE_DISABLE_1M_CONTEXT` environment variable to disable 1M context window support
350- Opus 4.6 (fast mode) now includes the full 1M context window
351- VSCode: Added `/extra-usage` command support in VS Code sessions
352- Fixed memory leak where TaskOutput retained recent lines after cleanup
353- Fixed memory leak in CircularBuffer where cleared items were retained in the backing array
354- Fixed memory leak in shell command execution where ChildProcess and AbortController references were retained after cleanup
355
356## 2.1.49
357
358- Improved MCP OAuth authentication with step-up auth support and discovery caching, reducing redundant network requests during server connections
359- Added `--worktree` (`-w`) flag to start Claude in an isolated git worktree
360- Subagents support `isolation: "worktree"` for working in a temporary git worktree
361- Added Ctrl+F keybinding to kill background agents (two-press confirmation)
362- Agent definitions support `background: true` to always run as a background task
363- Plugins can ship `settings.json` for default configuration
364- Fixed file-not-found errors to suggest corrected paths when the model drops the repo folder
365- Fixed Ctrl+C and ESC being silently ignored when background agents are running and the main thread is idle. Pressing twice within 3 seconds now kills all background agents.
366- Fixed prompt suggestion cache regression that reduced cache hit rates.
367- Fixed `plugin enable` and `plugin disable` to auto-detect the correct scope when `--scope` is not specified, instead of always defaulting to user scope
368- Simple mode (`CLAUDE_CODE_SIMPLE`) now includes the file edit tool in addition to the Bash tool, allowing direct file editing in simple mode.
369- Permission suggestions are now populated when safety checks trigger an ask response, enabling SDK consumers to display permission options
370- Sonnet 4.5 with 1M context is being removed from the Max plan in favor of our frontier Sonnet 4.6 model, which now has 1M context. Please switch in /model.
371- Fixed verbose mode not updating thinking block display when toggled via `/config` — memo comparators now correctly detect verbose changes
372- Fixed unbounded WASM memory growth during long sessions by periodically resetting the tree-sitter parser
373- Fixed potential rendering issues caused by stale yoga layout references
374- Improved performance in non-interactive mode (`-p`) by skipping unnecessary API calls during startup
375- Improved performance by caching authentication failures for HTTP and SSE MCP servers, avoiding repeated connection attempts to servers requiring auth
376- Fixed unbounded memory growth during long-running sessions caused by Yoga WASM linear memory never shrinking
377- SDK model info now includes `supportsEffort`, `supportedEffortLevels`, and `supportsAdaptiveThinking` fields so consumers can discover model capabilities.
378- Added `ConfigChange` hook event that fires when configuration files change during a session, enabling enterprise security auditing and optional blocking of settings changes.
379- Improved startup performance by caching MCP auth failures to avoid redundant connection attempts
380- Improved startup performance by reducing HTTP calls for analytics token counting
381- Improved startup performance by batching MCP tool token counting into a single API call
382- Fixed `disableAllHooks` setting to respect managed settings hierarchy — non-managed settings can no longer disable managed hooks set by policy (#26637)
383- Fixed `--resume` session picker showing raw XML tags for sessions that start with commands like `/clear`. Now correctly falls through to the session ID fallback.
384- Improved permission prompts for path safety and working directory blocks to show the reason for the restriction instead of a bare prompt with no context
385
386## 2.1.47
387
388- Fixed FileWriteTool line counting to preserve intentional trailing blank lines instead of stripping them with `trimEnd()`.
389- Fixed Windows terminal rendering bugs caused by `os.EOL` (`\r\n`) in display code — line counts now show correct values instead of always showing 1 on Windows.
390- Improved VS Code plan preview: auto-updates as Claude iterates, enables commenting only when the plan is ready for review, and keeps the preview open when rejecting so Claude can revise.
391- Fixed a bug where bold and colored text in markdown output could shift to the wrong characters on Windows due to `\r\n` line endings.
392- Fixed compaction failing when conversation contains many PDF documents by stripping document blocks alongside images before sending to the compaction API (anthropics/claude-code#26188)
393- Improved memory usage in long-running sessions by releasing API stream buffers, agent context, and skill state after use
394- Improved startup performance by deferring SessionStart hook execution, reducing time-to-interactive by ~500ms.
395- Fixed an issue where bash tool output was silently discarded on Windows when using MSYS2 or Cygwin shells.
396- Improved performance of `@` file mentions - file suggestions now appear faster by pre-warming the index on startup and using session-based caching with background refresh.
397- Improved memory usage by trimming agent task message history after tasks complete
398- Improved memory usage during long agent sessions by eliminating O(n²) message accumulation in progress updates
399- Fixed the bash permission classifier to validate that returned match descriptions correspond to actual input rules, preventing hallucinated descriptions from incorrectly granting permissions
400- Fixed user-defined agents only loading one file on NFS/FUSE filesystems that report zero inodes (anthropics/claude-code#26044)
401- Fixed plugin agent skills silently failing to load when referenced by bare name instead of fully-qualified plugin name (anthropics/claude-code#25834)
402- Search patterns in collapsed tool results are now displayed in quotes for clarity
403- Windows: Fixed CWD tracking temp files never being cleaned up, causing them to accumulate indefinitely (anthropics/claude-code#17600)
404- Use `ctrl+f` to kill all background agents instead of double-pressing ESC. Background agents now continue running when you press ESC to cancel the main thread, giving you more control over agent lifecycle.
405- Fixed API 400 errors ("thinking blocks cannot be modified") that occurred in sessions with concurrent agents, caused by interleaved streaming content blocks preventing proper message merging.
406- Simplified teammate navigation to use only Shift+Down (with wrapping) instead of both Shift+Up and Shift+Down.
407- Fixed an issue where a single file write/edit error would abort all other parallel file write/edit operations. Independent file mutations now complete even when a sibling fails.
408- Added `last_assistant_message` field to Stop and SubagentStop hook inputs, providing the final assistant response text so hooks can access it without parsing transcript files.
409- Fixed custom session titles set via `/rename` being lost after resuming a conversation (anthropics/claude-code#23610)
410- Fixed collapsed read/search hint text overflowing on narrow terminals by truncating from the start.
411- Fixed an issue where bash commands with backslash-newline continuation lines (e.g., long commands split across multiple lines with `\`) would produce spurious empty arguments, potentially breaking command execution.
412- Fixed built-in slash commands (`/help`, `/model`, `/compact`, etc.) being hidden from the autocomplete dropdown when many user skills are installed (anthropics/claude-code#22020)
413- Fixed MCP servers not appearing in the MCP Management Dialog after deferred loading
414- Fixed session name persisting in status bar after `/clear` command (anthropics/claude-code#26082)
415- Fixed crash when a skill's `name` or `description` in SKILL.md frontmatter is a bare number (e.g., `name: 3000`) — the value is now properly coerced to a string (anthropics/claude-code#25837)
416- Fixed /resume silently dropping sessions when the first message exceeds 16KB or uses array-format content (anthropics/claude-code#25721)
417- Added `chat:newline` keybinding action for configurable multi-line input (anthropics/claude-code#26075)
418- Added `added_dirs` to the statusline JSON `workspace` section, exposing directories added via `/add-dir` to external scripts (anthropics/claude-code#26096)
419- Fixed `claude doctor` misclassifying mise and asdf-managed installations as native installs (anthropics/claude-code#26033)
420- Fixed zsh heredoc failing with "read-only file system" error in sandboxed commands (anthropics/claude-code#25990)
421- Fixed agent progress indicator showing inflated tool use count (anthropics/claude-code#26023)
422- Fixed image pasting not working on WSL2 systems where Windows copies images as BMP format (anthropics/claude-code#25935)
423- Fixed background agent results returning raw transcript data instead of the agent's final answer (anthropics/claude-code#26012)
424- Fixed Warp terminal incorrectly prompting for Shift+Enter setup when it supports it natively (anthropics/claude-code#25957)
425- Fixed CJK wide characters causing misaligned timestamps and layout elements in the TUI (anthropics/claude-code#26084)
426- Fixed custom agent `model` field in `.claude/agents/*.md` being ignored when spawning team teammates (anthropics/claude-code#26064)
427- Fixed plan mode being lost after context compaction, causing the model to switch from planning to implementation mode (anthropics/claude-code#26061)
428- Fixed `alwaysThinkingEnabled: true` in settings.json not enabling thinking mode on Bedrock and Vertex providers (anthropics/claude-code#26074)
429- Fixed `tool_decision` OTel telemetry event not being emitted in headless/SDK mode (anthropics/claude-code#26059)
430- Fixed session name being lost after context compaction — renamed sessions now preserve their custom title through compaction (anthropics/claude-code#26121)
431- Increased initial session count in resume picker from 10 to 50 for faster session discovery (anthropics/claude-code#26123)
432- Windows: fixed worktree session matching when drive letter casing differs (anthropics/claude-code#26123)
433- Fixed `/resume <session-id>` failing to find sessions whose first message exceeds 16KB (anthropics/claude-code#25920)
434- Fixed "Always allow" on multiline bash commands creating invalid permission patterns that corrupt settings (anthropics/claude-code#25909)
435- Fixed React crash (error #31) when a skill's `argument-hint` in SKILL.md frontmatter uses YAML sequence syntax (e.g., `[topic: foo | bar]`) — the value is now properly coerced to a string (anthropics/claude-code#25826)
436- Fixed crash when using `/fork` on sessions that used web search — null entries in search results from transcript deserialization are now handled gracefully (anthropics/claude-code#25811)
437- Fixed read-only git commands triggering FSEvents file watcher loops on macOS by adding --no-optional-locks flag (anthropics/claude-code#25750)
438- Fixed custom agents and skills not being discovered when running from a git worktree — project-level `.claude/agents/` and `.claude/skills/` from the main repository are now included (anthropics/claude-code#25816)
439- Fixed non-interactive subcommands like `claude doctor` and `claude plugin validate` being blocked inside nested Claude sessions (anthropics/claude-code#25803)
440- Windows: Fixed the same CLAUDE.md file being loaded twice when drive letter casing differs between paths (anthropics/claude-code#25756)
441- Fixed inline code spans in markdown being incorrectly parsed as bash commands (anthropics/claude-code#25792)
442- Fixed teammate spinners not respecting custom spinnerVerbs from settings (anthropics/claude-code#25748)
443- Fixed shell commands permanently failing after a command deletes its own working directory (anthropics/claude-code#26136)
444- Fixed hooks (PreToolUse, PostToolUse) silently failing to execute on Windows by using Git Bash instead of cmd.exe (anthropics/claude-code#25981)
445- Fixed LSP `findReferences` and other location-based operations returning results from gitignored files (e.g., `node_modules/`, `venv/`) (anthropics/claude-code#26051)
446- Moved config backup files from home directory root to `~/.claude/backups/` to reduce home directory clutter (anthropics/claude-code#26130)
447- Fixed sessions with large first prompts (>16KB) disappearing from the /resume list (anthropics/claude-code#26140)
448- Fixed shell functions with double-underscore prefixes (e.g., `__git_ps1`) not being preserved across shell sessions (anthropics/claude-code#25824)
449- Fixed spinner showing "0 tokens" counter before any tokens have been received (anthropics/claude-code#26105)
450- VSCode: Fixed conversation messages appearing dimmed while the AskUserQuestion dialog is open (anthropics/claude-code#26078)
451- Fixed background tasks failing in git worktrees due to remote URL resolution reading from worktree-specific gitdir instead of the main repository config (anthropics/claude-code#26065)
452- Fixed Right Alt key leaving visible `[25~` escape sequence residue in the input field on Windows/Git Bash terminals (anthropics/claude-code#25943)
453- The `/rename` command now updates the terminal tab title by default (anthropics/claude-code#25789)
454- Fixed Edit tool silently corrupting Unicode curly quotes (\u201c\u201d \u2018\u2019) by replacing them with straight quotes when making edits (anthropics/claude-code#26141)
455- Fixed OSC 8 hyperlinks only being clickable on the first line when link text wraps across multiple terminal lines.
456
457## 2.1.46
458
459- Fixed orphaned CC processes after terminal disconnect on macOS
460- Added support for using claude.ai MCP connectors in Claude Code
461
462## 2.1.45
463
464- Added support for Claude Sonnet 4.6
465- Added support for reading `enabledPlugins` and `extraKnownMarketplaces` from `--add-dir` directories
466- Added `spinnerTipsOverride` setting to customize spinner tips — configure `tips` with an array of custom tip strings, and optionally set `excludeDefault: true` to show only your custom tips instead of the built-in ones
467- Added `SDKRateLimitInfo` and `SDKRateLimitEvent` types to the SDK, enabling consumers to receive rate limit status updates including utilization, reset times, and overage information
468- Fixed Agent Teams teammates failing on Bedrock, Vertex, and Foundry by propagating API provider environment variables to tmux-spawned processes (anthropics/claude-code#23561)
469- Fixed sandbox "operation not permitted" errors when writing temporary files on macOS by using the correct per-user temp directory (anthropics/claude-code#21654)
470- Fixed Task tool (backgrounded agents) crashing with a `ReferenceError` on completion (anthropics/claude-code#22087)
471- Fixed autocomplete suggestions not being accepted on Enter when images are pasted in the input
472- Fixed skills invoked by subagents incorrectly appearing in main session context after compaction
473- Fixed excessive `.claude.json.backup` files accumulating on every startup
474- Fixed plugin-provided commands, agents, and hooks not being available immediately after installation without requiring a restart
475- Improved startup performance by removing eager loading of session history for stats caching
476- Improved memory usage for shell commands that produce large output — RSS no longer grows unboundedly with command output size
477- Improved collapsed read/search groups to show the current file or search pattern being processed beneath the summary line while active
478- [VSCode] Improved permission destination choice (project/user/session) to persist across sessions
479
480## 2.1.44
481
482- Fixed ENAMETOOLONG errors for deeply-nested directory paths
483- Fixed auth refresh errors
484
485## 2.1.43
486
487- Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
488- Fixed spurious warnings for non-agent markdown files in `.claude/agents/` directory
489- Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock
490
491## 2.1.42
492
493- Improved startup performance by deferring Zod schema construction
494- Improved prompt cache hit rates by moving date out of system prompt
495- Added one-time Opus 4.6 effort callout for eligible users
496- Fixed /resume showing interrupt messages as session titles
497- Fixed image dimension limit errors to suggest /compact
498
499## 2.1.41
500
501- Added guard against launching Claude Code inside another Claude Code session
502- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
503- Fixed a crash when MCP tools return image content during streaming
504- Fixed /resume session previews showing raw XML tags instead of readable command names
505- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
506- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
507- Fixed hook blocking errors (exit code 2) not showing stderr to the user
508- Added `speed` attribute to OTel events and trace spans for fast mode visibility
509- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
510- Added Windows ARM64 (win32-arm64) native binary support
511- Improved `/rename` to auto-generate session name from conversation context when called without arguments
512- Improved narrow terminal layout for prompt footer
513- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
514- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
515- Fixed background task notifications not being delivered in streaming Agent SDK mode
516- Fixed cursor jumping to end on each keystroke in classifier rule input
517- Fixed markdown link display text being dropped for raw URL
518- Fixed auto-compact failure error notifications being shown to users
519- Fixed permission wait time being included in subagent elapsed time display
520- Fixed proactive ticks firing while in plan mode
521- Fixed clear stale permission rules when settings change on disk
522- Fixed hook blocking errors showing stderr content in UI
523
524## 2.1.39
525
526- Improved terminal rendering performance
527- Fixed fatal errors being swallowed instead of displayed
528- Fixed process hanging after session close
529- Fixed character loss at terminal screen boundary
530- Fixed blank lines in verbose transcript view
531
532## 2.1.38
533
534- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
535- Fixed Tab key queueing slash commands instead of autocompleting
536- Fixed bash permission matching for commands using environment variable wrappers
537- Fixed text between tool uses disappearing when not using streaming
538- Fixed duplicate sessions when resuming in VS Code extension
539- Improved heredoc delimiter parsing to prevent command smuggling
540- Blocked writes to `.claude/skills` directory in sandbox mode
541
542## 2.1.37
543
544- Fixed an issue where /fast was not immediately available after enabling /extra-usage
545
546## 2.1.36
547
548- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
549
550## 2.1.34
551
552- Fixed a crash when agent teams setting changed between renders
553- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
554
555## 2.1.33
556
557- Fixed agent teammate sessions in tmux to send and receive messages
558- Fixed warnings about agent teams not being available on your current plan
559- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
560- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
561- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
562- Added plugin name to skill descriptions and `/skills` menu for better discoverability
563- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
564- 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
565- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
566- 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
567- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
568- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
569- Errors from invalid managed settings are now surfaced
570- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
571- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
572- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
573
574## 2.1.32
575
576- Claude Opus 4.6 is now available!
577- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
578- Claude now automatically records and recalls memories as it works
579- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
580- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
581- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
582- Updated --resume to re-use --agent value specified in previous conversation by default.
583- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
584- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
585- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
586- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
587- VSCode: Added spinner when loading past conversations list
588
589## 2.1.31
590
591- Added session resume hint on exit, showing how to continue your conversation later
592- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
593- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
594- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
595- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
596- 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
597- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
598- 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
599- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
600- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
601- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
602
603## 2.1.30
604
605- 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.
606- 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`.
607- Added `/debug` for Claude to help troubleshoot the current session
608- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
609- Added token count, tool uses, and duration metrics to Task tool results
610- Added reduced motion mode to the config
611- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
612- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
613- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
614- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
615- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
616- Fixed permission dialogs stealing focus while actively typing
617- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
618- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
619- 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
620- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
621- Changed `/model` to execute immediately instead of being queued
622- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
623- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
624
625## 2.1.29
626
627- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
628
629## 2.1.27
630
631- Added tool call failures and denials to debug logs
632- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
633- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
634- Sessions are now automatically linked to PRs when created via `gh pr create`
635- Fixed /context command not displaying colored output
636- Fixed status bar duplicating background task indicator when PR status was shown
637- Windows: Fixed bash command execution failing for users with `.bashrc` files
638- Windows: Fixed console windows flashing when spawning child processes
639- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
640
641## 2.1.25
642
643- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
644
645## 2.1.23
646
647- Added customizable spinner verbs setting (`spinnerVerbs`)
648- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
649- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
650- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
651- Fixed pending async hooks not being cancelled when headless streaming sessions ended
652- Fixed tab completion not updating the input field when accepting a suggestion
653- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
654- Improved terminal rendering performance with optimized screen data layout
655- Changed Bash commands to show timeout duration alongside elapsed time
656- Changed merged pull requests to show a purple status indicator in the prompt footer
657- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
658
659## 2.1.22
660
661- Fixed structured outputs for non-interactive (-p) mode
662
663## 2.1.21
664
665- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
666- Fixed shell completion cache files being truncated on exit
667- Fixed API errors when resuming sessions that were interrupted during tool execution
668- Fixed auto-compact triggering too early on models with large output token limits
669- Fixed task IDs potentially being reused after deletion
670- Fixed file search not working in VS Code extension on Windows
671- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
672- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
673- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
674- [VSCode] Fixed message action buttons having incorrect background colors
675
676## 2.1.20
677
678- Added arrow key history navigation in vim normal mode when cursor cannot move further
679- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
680- 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
681- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
682- Added ability to delete tasks via the `TaskUpdate` tool
683- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
684- Fixed agents sometimes ignoring user messages sent while actively working on a task
685- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
686- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
687- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
688- Fixed draft prompt being lost when pressing UP arrow to navigate command history
689- Fixed ghost text flickering when typing slash commands mid-input
690- Fixed marketplace source removal not properly deleting settings
691- Fixed duplicate output in some commands like `/context`
692- Fixed task list sometimes showing outside the main conversation view
693- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
694- Fixed crashes when cancelling tool use
695- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
696- Improved thinking status text with a subtle shimmer animation
697- Improved task list to dynamically adjust visible items based on terminal height
698- Improved fork conversation hint to show how to resume the original session
699- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
700- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
701- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
702- Changed the `/copy` command to be available to all users
703- Changed background agents to prompt for tool permissions before launching
704- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
705- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
706
707## 2.1.19
708
709- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
710- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
711- Fixed crashes on processors without AVX instruction support
712- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
713- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
714- Fixed resuming sessions by custom title not working when run from a different directory
715- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
716- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
717- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
718- Fixed file write preview omitting empty lines
719- Changed skills without additional permissions or hooks to be allowed without requiring approval
720- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
721- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
722- [VSCode] Enabled session forking and rewind functionality for all users
723
724## 2.1.18
725
726- 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
727
728## 2.1.17
729
730- Fixed crashes on processors without AVX instruction support
731
732## 2.1.16
733
734- Added new task management system, including new capabilities like dependency tracking
735- [VSCode] Added native plugin management support
736- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
737- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
738- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
739- Fixed session titles on the resume screen not respecting the user's language setting
740- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
741
742## 2.1.15
743
744- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
745- Improved UI rendering performance with React Compiler
746- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
747- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
748
749## 2.1.14
750
751- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
752- Added search to installed plugins list - type to filter by name or description
753- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
754- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
755- Fixed memory issues that could cause crashes when running parallel subagents
756- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
757- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
758- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
759- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
760- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
761- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
762- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
763- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
764- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
765- Improved backspace to delete pasted text as a single token instead of one character at a time
766- [VSCode] Added `/usage` command to display current plan usage
767
768## 2.1.12
769
770- Fixed message rendering bug
771
772## 2.1.11
773
774- Fixed excessive MCP connection requests for HTTP/SSE transports
775
776## 2.1.10
777
778- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
779- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
780- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
781- Improved startup to capture keystrokes typed before the REPL is fully ready
782- Improved file suggestions to show as removable attachments instead of inserting text when accepted
783- [VSCode] Added install count display to plugin listings
784- [VSCode] Added trust warning when installing plugins
785
786## 2.1.9
787
788- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
789- Added `plansDirectory` setting to customize where plan files are stored
790- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
791- Added session URL attribution to commits and PRs created from web sessions
792- Added support for `PreToolUse` hooks to return `additionalContext` to the model
793- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
794- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
795- Fixed MCP server reconnection hanging when cached connection promise never resolves
796- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
797
798## 2.1.7
799
800- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
801- Added ability to provide feedback when accepting permission prompts
802- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
803- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
804- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
805- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
806- 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)
807- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
808- Fixed terminal title animation jitter by using fixed-width braille characters
809- Fixed plugins with git submodules not being fully initialized when installed
810- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
811- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
812- 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.
813- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
814- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
815
816## 2.1.6
817
818- Added search functionality to `/config` command for quickly filtering settings
819- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
820- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
821- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
822- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
823- Added an error display when the editor fails during Ctrl+G
824- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
825- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
826- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
827- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
828- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
829- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
830- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
831- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
832- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
833- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
834- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
835- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
836- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
837- Improved @ autocomplete with icons for different suggestion types and single-line formatting
838- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
839- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
840- Changed terminal title to "Claude Code" on startup for better window identification
841- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
842- [VSCode] Fixed usage indicator not updating after manual compact
843
844## 2.1.5
845
846- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
847
848## 2.1.4
849
850- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
851- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
852
853## 2.1.3
854
855- Merged slash commands and skills, simplifying the mental model with no change in behavior
856- Added release channel (`stable` or `latest`) toggle to `/config`
857- 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
858- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
859- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
860- Fixed mismatch between background task count in status bar and items shown in tasks dialog
861- Fixed sub-agents using the wrong model during conversation compaction
862- Fixed web search in sub-agents using incorrect model
863- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
864- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
865- Improved slash command suggestion readability by truncating long descriptions to 2 lines
866- Changed tool hook execution timeout from 60 seconds to 10 minutes
867- [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)
868
869## 2.1.2
870
871- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
872- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
873- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
874- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
875- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
876- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
877- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
878- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
879- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
880- Fixed updates incorrectly claiming another installation is in progress
881- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
882- Fixed remote session URL and teleport being broken when using `/tasks` command
883- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
884- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
885- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
886- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
887- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
888- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
889- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
890- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
891- [SDK] Changed minimum zod peer dependency to ^4.0.0
892- [VSCode] Fixed usage display not updating after manual compact
893
894## 2.1.0
895
896- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
897- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
898- Added support for `agent` field in skills to specify agent type for execution
899- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
900- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
901- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
902- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
903- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
904- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
905- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
906- Fixed Esc key with queued prompts to only move them to input without canceling the running task
907- Reduced permission prompts for complex bash commands
908- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
909- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
910- Fixed plugin path resolution for file-based marketplace sources
911- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
912- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
913- Fixed Claude in Chrome support for WSL environments
914- Fixed Windows native installer silently failing when executable creation fails
915- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
916- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
917- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
918- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
919- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
920- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
921- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
922- Added hooks support for skill and slash command frontmatter
923- 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
924- Added `/plan` command shortcut to enable plan mode directly from the prompt
925- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
926- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
927- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
928- Added support for `once: true` config for hooks
929- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
930- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
931- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
932- Added left/right arrow key navigation for cycling through tabs in dialogs
933- Added real-time thinking block display in Ctrl+O transcript mode
934- Added filepath to full output in background bash task details dialog
935- Added Skills as a separate category in the context visualization
936- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
937- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
938- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
939- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
940- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
941- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
942- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
943- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
944- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
945- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
946- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
947- Fixed commands with `$()` command substitution failing with parse errors
948- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
949- 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)
950- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
951- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
952- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
953- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
954- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
955- Fixed subagents sometimes not inheriting the parent's model by default
956- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
957- Fixed duplicate Bash commands appearing in permission request option labels
958- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
959- Fixed background task completion notifications to appear proactively with bullet point
960- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
961- Fixed cursor disappearing after dismissing permission dialogs
962- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
963- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
964- Fixed images being silently dropped when queueing messages while backgrounding a task
965- Fixed large pasted images failing with "Image was too large" error
966- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
967- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
968- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
969- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
970- Fixed spinner token counter to properly accumulate tokens from subagents during execution
971- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
972- Fixed race condition where LSP tool could return "no server available" during startup
973- Fixed feedback submission hanging indefinitely when network requests timeout
974- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
975- Fixed hook success message showing trailing colon when hook has no output
976- Multiple optimizations to improve startup performance
977- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
978- Improved performance when reading Jupyter notebooks with many cells
979- Improved reliability for piped input like `cat refactor.md | claude`
980- Improved reliability for AskQuestion tool
981- Improved sed in-place edit commands to render as file edits with diff preview
982- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
983- Improved compaction reliability
984- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
985- Improved skills to show progress while executing, displaying tool uses as they happen
986- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
987- Improved skill suggestions to prioritize recently and frequently used skills
988- Improved spinner feedback when waiting for the first response token
989- Improved token count display in spinner to include tokens from background agents
990- Improved incremental output for async agents to give the main thread more control and visibility
991- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
992- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
993- Improved macOS screenshot paste reliability with TIFF format support
994- Improved `/stats` output
995- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
996- Changed "Interrupted" message color from red to grey for a less alarming appearance
997- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
998- Removed underline styling from image reference links
999- [SDK] Changed minimum zod peer dependency to ^4.0.0
1000- [VSCode] Added currently selected model name to the context menu
1001- [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")
1002- [VSCode] Fixed paragraph breaks not rendering in markdown content
1003- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
1004- [Windows] Fixed issue with improper rendering
1005
1006## 2.0.76
1007
1008- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
1009
1010## 2.0.75
1011
1012- Minor bugfixes
1013
1014## 2.0.74
1015
1016- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
1017- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
1018- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
1019- Added syntax highlighting info to theme picker
1020- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
1021- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
1022- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
1023- Fixed a potential crash when syntax highlighting isn't initialized correctly
1024- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
1025- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
1026- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
1027- [Windows] Fixed issue with improper rendering
1028- [VSCode] Added gift tag pictogram for year-end promotion message
1029
1030## 2.0.73
1031
1032- Added clickable `[Image #N]` links that open attached images in the default viewer
1033- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
1034- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
1035- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
1036- Fixed slow input history cycling and race condition that could overwrite text after message submission
1037- Improved `/theme` command to open theme picker directly
1038- Improved theme picker UI
1039- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
1040- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
1041
1042## 2.0.72
1043
1044- 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
1045- Reduced terminal flickering
1046- Added scannable QR code to mobile app tip for quick app downloads
1047- Added loading indicator when resuming conversations for better feedback
1048- Fixed `/context` command not respecting custom system prompts in non-interactive mode
1049- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
1050- Improved @ mention file suggestion speed (~3× faster in git repositories)
1051- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
1052- Improved settings validation errors to be more prominent
1053- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
1054
1055## 2.0.71
1056
1057- Added /config toggle to enable/disable prompt suggestions
1058- Added `/settings` as an alias for the `/config` command
1059- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
1060- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
1061- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
1062- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
1063- New syntax highlighting engine for native build
1064
1065## 2.0.70
1066
1067- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
1068- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
1069- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
1070- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
1071- Fixed input being cleared when processing queued commands while the user was typing
1072- Fixed prompt suggestions replacing typed input when pressing Tab
1073- Fixed diff view not updating when terminal is resized
1074- Improved memory usage by 3x for large conversations
1075- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
1076- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
1077- Fix thinking mode toggle in /config not persisting correctly
1078- Improve UI for file creation permission dialog
1079
1080## 2.0.69
1081
1082- Minor bugfixes
1083
1084## 2.0.68
1085
1086- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
1087- Fixed a bug where disallowed MCP tools were visible to the model
1088- Fixed an issue where steering messages could be lost while a subagent is working
1089- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
1090- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
1091- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
1092
1093## 2.0.67
1094
1095- Thinking mode is now enabled by default for Opus 4.5
1096- Thinking mode configuration has moved to /config
1097- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
1098- Show reason why autoupdater is disabled in `/doctor`
1099- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
1100- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
1101- Fixed scroll position resetting after deleting a permission rule in `/permissions`
1102- 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
1103- Fixed `claude install --force` not bypassing stale lock files
1104- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
1105- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
1106
1107## 2.0.65
1108
1109- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
1110- Added context window information to status line input
1111- Added `fileSuggestion` setting for custom `@` file search commands
1112- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
1113- Fixed prompt not being saved to history when aborting a query with Escape
1114- Fixed Read tool image handling to identify format from bytes instead of file extension
1115
1116## 2.0.64
1117
1118- Made auto-compacting instant
1119- Agents and bash commands can run asynchronously and send messages to wake up the main agent
1120- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
1121- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
1122- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
1123- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
1124- Fixed auto-loading .env when using native installer
1125- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
1126- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
1127- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
1128- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
1129- Bedrock: Improve efficiency of token counting
1130- Bedrock: Add support for `aws login` AWS Management Console credentials
1131- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
1132
1133## 2.0.62
1134
1135- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
1136- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
1137- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
1138- Fixed slash command selection not working when multiple commands share the same name
1139- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
1140- Fixed running versions getting removed because lock file incorrectly going stale
1141- Fixed IDE diff tab not closing when rejecting file changes
1142
1143## 2.0.61
1144
1145- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
1146
1147## 2.0.60
1148
1149- Added background agent support. Agents run in the background while you work
1150- Added --disable-slash-commands CLI flag to disable all slash commands
1151- Added model name to "Co-Authored-By" commit messages
1152- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
1153- Updated Fetch to skip summarization for pre-approved websites
1154- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
1155
1156## 2.0.59
1157
1158- Added --agent CLI flag to override the agent setting for the current session
1159- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
1160- VS Code: Fixed .claude.json config file being read from incorrect location
1161
1162## 2.0.58
1163
1164- Pro users now have access to Opus 4.5 as part of their subscription!
1165- Fixed timer duration showing "11m 60s" instead of "12m 0s"
1166- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
1167
1168## 2.0.57
1169
1170- Added feedback input when rejecting plans, allowing users to tell Claude what to change
1171- VSCode: Added streaming message support for real-time response display
1172
1173## 2.0.56
1174
1175- Added setting to enable/disable terminal progress bar (OSC 9;4)
1176- 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.
1177
1178## 2.0.55
1179
1180- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
1181- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
1182- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
1183- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
1184
1185## 2.0.54
1186
1187- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
1188- Fix issue with excessive iTerm notifications
1189
1190## 2.0.52
1191
1192- Fixed duplicate message display when starting Claude with a command line argument
1193- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
1194- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
1195- Permit some uses of `$!` in bash commands
1196
1197## 2.0.51
1198
1199- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
1200- Introducing Claude Code for Desktop: https://claude.com/download
1201- 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
1202- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
1203- Plan Mode now builds more precise plans and executes more thoroughly
1204- Usage limit notifications now easier to understand
1205- Switched `/usage` back to "% used"
1206- Fixed handling of thinking errors
1207- Fixed performance regression
1208
1209## 2.0.50
1210
1211- Fixed bug preventing calling MCP tools that have nested references in their input schemas
1212- Silenced a noisy but harmless error during upgrades
1213- Improved ultrathink text display
1214- Improved clarity of 5-hour session limit warning message
1215
1216## 2.0.49
1217
1218- Added readline-style ctrl-y for pasting deleted text
1219- Improved clarity of usage limit warning message
1220- Fixed handling of subagent permissions
1221
1222## 2.0.47
1223
1224- Improved error messages and validation for `claude --teleport`
1225- Improved error handling in `/usage`
1226- Fixed race condition with history entry not getting logged at exit
1227- Fixed Vertex AI configuration not being applied from `settings.json`
1228
1229## 2.0.46
1230
1231- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
1232
1233## 2.0.45
1234
1235- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
1236- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
1237- Send background tasks to Claude Code on the web by starting a message with `&`
1238
1239## 2.0.43
1240
1241- Added `permissionMode` field for custom agents
1242- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
1243- Added skills frontmatter field to declare skills to auto-load for subagents
1244- Added the `SubagentStart` hook event
1245- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
1246- Fixed duplicate rendering of some messages in the UI
1247- Fixed some visual flickers
1248- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
1249
1250## 2.0.42
1251
1252- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
1253
1254## 2.0.41
1255
1256- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
1257- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
1258- Fixed incorrect labeling of user settings vs project settings in command descriptions
1259- Fixed crash when plugin command hooks timeout during execution
1260- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
1261- Fixed broken security documentation links in trust dialogs and onboarding
1262- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
1263- ctrl-r history search landing on a slash command no longer cancels the search
1264- SDK: Support custom timeouts for hooks
1265- Allow more safe git commands to run without approval
1266- Plugins: Added support for sharing and installing output styles
1267- Teleporting a session from web will automatically set the upstream branch
1268
1269## 2.0.37
1270
1271- Fixed how idleness is computed for notifications
1272- Hooks: Added matcher values for Notification hook events
1273- Output Styles: Added `keep-coding-instructions` option to frontmatter
1274
1275## 2.0.36
1276
1277- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
1278- Fixed queued messages being incorrectly executed as bash commands
1279- Fixed input being lost when typing while a queued message is processed
1280
1281## 2.0.35
1282
1283- Improve fuzzy search results when searching commands
1284- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
1285- 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
1286- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
1287- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
1288
1289## 2.0.34
1290
1291- VSCode Extension: Added setting to configure the initial permission mode for new conversations
1292- Improved file path suggestion performance with native Rust-based fuzzy finder
1293- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
1294- Fixed memory crash when reading or writing large files (especially base64-encoded images)
1295
1296## 2.0.33
1297
1298- Native binary installs now launch quicker.
1299- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
1300- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
1301
1302## 2.0.32
1303
1304- Un-deprecate output styles based on community feedback
1305- Added `companyAnnouncements` setting for displaying announcements on startup
1306- Fixed hook progress messages not updating correctly during PostToolUse hook execution
1307
1308## 2.0.31
1309
1310- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
1311- Vertex: add support for Web Search on supported models
1312- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1313- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
1314- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
1315- Fixed plugin uninstall not removing plugins
1316
1317## 2.0.30
1318
1319- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
1320- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
1321- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
1322- Added prompt-based stop hooks
1323- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1324- Enabled SSE MCP servers on native build
1325- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
1326- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
1327- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
1328- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
1329- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
1330- Fixed bug that saved session permissions to local settings
1331- Fixed MCP tools not being available to sub-agents
1332- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
1333- Fixed delay when navigating through typeahead suggestions with arrow keys
1334- VSCode: Restored selection indicator in input footer showing current file or code selection status
1335
1336## 2.0.28
1337
1338- Plan mode: introduced new Plan subagent
1339- Subagents: claude can now choose to resume subagents
1340- Subagents: claude can dynamically choose the model used by its subagents
1341- SDK: added --max-budget-usd flag
1342- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
1343- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
1344- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
1345- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
1346- Various other bug fixes
1347
1348## 2.0.27
1349
1350- New UI for permission prompts
1351- Added current branch filtering and search to session resume screen for easier navigation
1352- Fixed directory @-mention causing "No assistant message found" error
1353- VSCode Extension: Add config setting to include .gitignored files in file searches
1354- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
1355
1356## 2.0.25
1357
1358- 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
1359
1360## 2.0.24
1361
1362- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
1363- Claude Code Web: Support for Web -> CLI teleport
1364- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
1365- Bedrock: Display awsAuthRefresh output when auth is required
1366
1367## 2.0.22
1368
1369- Fixed content layout shift when scrolling through slash commands
1370- IDE: Add toggle to enable/disable thinking.
1371- Fix bug causing duplicate permission prompts with parallel tool calls
1372- Add support for enterprise managed MCP allowlist and denylist
1373
1374## 2.0.21
1375
1376- Support MCP `structuredContent` field in tool responses
1377- Added an interactive question tool
1378- Claude will now ask you questions more often in plan mode
1379- Added Haiku 4.5 as a model option for Pro users
1380- Fixed an issue where queued commands don't have access to previous messages' output
1381
1382## 2.0.20
1383
1384- Added support for Claude Skills
1385
1386## 2.0.19
1387
1388- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
1389- Fixed a bug where Haiku was unnecessarily called in print mode
1390
1391## 2.0.17
1392
1393- Added Haiku 4.5 to model selector!
1394- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
1395- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
1396- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
1397- OTEL: support HTTP_PROXY and HTTPS_PROXY
1398- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
1399
1400## 2.0.15
1401
1402- Fixed bug with resuming where previously created files needed to be read again before writing
1403- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
1404
1405## 2.0.14
1406
1407- Fix @-mentioning MCP servers to toggle them on/off
1408- Improve permission checks for bash with inline env vars
1409- Fix ultrathink + thinking toggle
1410- Reduce unnecessary logins
1411- Document --system-prompt
1412- Several improvements to rendering
1413- Plugins UI polish
1414
1415## 2.0.13
1416
1417- Fixed `/plugin` not working on native build
1418
1419## 2.0.12
1420
1421- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
1422- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
1423- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
1424- `/plugin validate` command for validating plugin structure and configuration
1425- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
1426- Plugin documentation available at https://code.claude.com/docs/en/plugins
1427- Comprehensive error messages and diagnostics via `/doctor` command
1428- Avoid flickering in `/model` selector
1429- Improvements to `/help`
1430- Avoid mentioning hooks in `/resume` summaries
1431- Changes to the "verbose" setting in `/config` now persist across sessions
1432
1433## 2.0.11
1434
1435- Reduced system prompt size by 1.4k tokens
1436- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
1437- Fixed Opus fallback rate limit errors appearing incorrectly
1438- Fixed /add-dir command selecting wrong default tab
1439
1440## 2.0.10
1441
1442- Rewrote terminal renderer for buttery smooth UI
1443- Enable/disable MCP servers by @mentioning, or in /mcp
1444- Added tab completion for shell commands in bash mode
1445- PreToolUse hooks can now modify tool inputs
1446- Press Ctrl-G to edit your prompt in your system's configured text editor
1447- Fixes for bash permission checks with environment variables in the command
1448
1449## 2.0.9
1450
1451- Fix regression where bash backgrounding stopped working
1452
1453## 2.0.8
1454
1455- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
1456- IDE: Add drag-and-drop support for files and folders in chat
1457- /context: Fix counting for thinking blocks
1458- Improve message rendering for users with light themes on dark terminals
1459- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
1460
1461## 2.0.5
1462
1463- IDE: Fix IME unintended message submission with Enter and Tab
1464- IDE: Add "Open in Terminal" link in login screen
1465- Fix unhandled OAuth expiration 401 API errors
1466- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
1467
1468## 2.0.1
1469
1470- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
1471- Various bug fixes and presentation improvements
1472
1473## 2.0.0
1474
1475- New native VS Code extension
1476- Fresh coat of paint throughout the whole app
1477- /rewind a conversation to undo code changes
1478- /usage command to see plan limits
1479- Tab to toggle thinking (sticky across sessions)
1480- Ctrl-R to search history
1481- Unshipped claude config command
1482- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
1483- SDK: The Claude Code SDK is now the Claude Agent SDK
1484- Add subagents dynamically with `--agents` flag
1485
1486## 1.0.126
1487
1488- Enable /context command for Bedrock and Vertex
1489- Add mTLS support for HTTP-based OpenTelemetry exporters
1490
1491## 1.0.124
1492
1493- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1494- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1495- No longer inform Claude of the list of allowed tools when permission is denied
1496- Fixed security vulnerability in Bash tool permission checks
1497- Improved VSCode extension performance for large files
1498
1499## 1.0.123
1500
1501- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1502- Fixed thinking mode triggering on negation phrases like "don't think"
1503- Fixed rendering performance degradation during token streaming
1504- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1505- Enhanced BashTool environment snapshot logging
1506- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1507- Migrated --debug logging to a file, to enable easy tailing & filtering
1508
1509## 1.0.120
1510
1511- Fix input lag during typing, especially noticeable with large prompts
1512- Improved VSCode extension command registry and sessions dialog user experience
1513- Enhanced sessions dialog responsiveness and visual feedback
1514- Fixed IDE compatibility issue by removing worktree support check
1515- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1516
1517## 1.0.119
1518
1519- Fix Windows issue where process visually freezes on entering interactive mode
1520- Support dynamic headers for MCP servers via headersHelper configuration
1521- Fix thinking mode not working in headless sessions
1522- Fix slash commands now properly update allowed tools instead of replacing them
1523
1524## 1.0.117
1525
1526- Add Ctrl-R history search to recall previous commands like bash/zsh
1527- Fix input lag while typing, especially on Windows
1528- Add sed command to auto-allowed commands in acceptEdits mode
1529- Fix Windows PATH comparison to be case-insensitive for drive letters
1530- Add permissions management hint to /add-dir output
1531
1532## 1.0.115
1533
1534- Improve thinking mode display with enhanced visual effects
1535- Type /t to temporarily disable thinking mode in your prompt
1536- Improve path validation for glob and grep tools
1537- Show condensed output for post-tool hooks to reduce visual clutter
1538- Fix visual feedback when loading state completes
1539- Improve UI consistency for permission request dialogs
1540
1541## 1.0.113
1542
1543- Deprecated piped input in interactive mode
1544- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1545
1546## 1.0.112
1547
1548- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1549- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1550- Hooks: Added systemMessage support for SessionEnd hooks
1551- Added `spinnerTipsEnabled` setting to disable spinner tips
1552- IDE: Various improvements and bug fixes
1553
1554## 1.0.111
1555
1556- /model now validates provided model names
1557- Fixed Bash tool crashes caused by malformed shell syntax parsing
1558
1559## 1.0.110
1560
1561- /terminal-setup command now supports WezTerm
1562- MCP: OAuth tokens now proactively refresh before expiration
1563- Fixed reliability issues with background Bash processes
1564
1565## 1.0.109
1566
1567- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1568
1569## 1.0.106
1570
1571- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1572
1573## 1.0.97
1574
1575- Settings: /doctor now validates permission rule syntax and suggests corrections
1576
1577## 1.0.94
1578
1579- Vertex: add support for global endpoints for supported models
1580- /memory command now allows direct editing of all imported memory files
1581- SDK: Add custom tools as callbacks
1582- Added /todos command to list current todo items
1583
1584## 1.0.93
1585
1586- Windows: Add alt + v shortcut for pasting images from clipboard
1587- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1588
1589## 1.0.90
1590
1591- Settings file changes take effect immediately - no restart required
1592
1593## 1.0.88
1594
1595- Fixed issue causing "OAuth authentication is currently not supported"
1596- Status line input now includes `exceeds_200k_tokens`
1597- Fixed incorrect usage tracking in /cost.
1598- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1599- Bedrock: Updated default Sonnet model to Sonnet 4
1600
1601## 1.0.86
1602
1603- Added /context to help users self-serve debug context issues
1604- SDK: Added UUID support for all SDK messages
1605- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1606
1607## 1.0.85
1608
1609- Status line input now includes session cost info
1610- Hooks: Introduced SessionEnd hook
1611
1612## 1.0.84
1613
1614- Fix tool_use/tool_result id mismatch error when network is unstable
1615- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1616- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1617- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1618
1619## 1.0.83
1620
1621- @-mention: Support files with spaces in path
1622- New shimmering spinner
1623
1624## 1.0.82
1625
1626- SDK: Add request cancellation support
1627- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1628- Settings: Validation prevents invalid fields in .claude/settings.json files
1629- MCP: Improve tool name consistency
1630- Bash: Fix crash when Claude tries to automatically read large files
1631
1632## 1.0.81
1633
1634- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1635- Agents: Fix custom agent loading when agent files are unparsable
1636
1637## 1.0.80
1638
1639- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1640
1641## 1.0.77
1642
1643- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1644- SDK: Add session support and permission denial tracking
1645- Fix token limit errors in conversation summarization
1646- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1647
1648## 1.0.73
1649
1650- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1651- MCP: Press Esc to cancel OAuth authentication flows
1652- Bash: Improved command validation and reduced false security warnings
1653- UI: Enhanced spinner animations and status line visual hierarchy
1654- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1655
1656## 1.0.72
1657
1658- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1659
1660## 1.0.71
1661
1662- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1663- Customizable status line: add your terminal prompt to Claude Code with /statusline
1664
1665## 1.0.70
1666
1667- Performance: Optimized message rendering for better performance with large contexts
1668- Windows: Fixed native file search, ripgrep, and subagent functionality
1669- Added support for @-mentions in slash command arguments
1670
1671## 1.0.69
1672
1673- Upgraded Opus to version 4.1
1674
1675## 1.0.68
1676
1677- Fix incorrect model names being used for certain commands like `/pr-comments`
1678- 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.
1679- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1680- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1681- SDK: Added canUseTool callback support for tool confirmation
1682- Added `disableAllHooks` setting
1683- Improved file suggestions performance in large repos
1684
1685## 1.0.65
1686
1687- IDE: Fixed connection stability issues and error handling for diagnostics
1688- Windows: Fixed shell environment setup for users without .bashrc files
1689
1690## 1.0.64
1691
1692- Agents: Added model customization support - you can now specify which model an agent should use
1693- Agents: Fixed unintended access to the recursive agent tool
1694- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1695- SDK: Fixed user input tracking across multi-turn conversations
1696- Added hidden files to file search and @-mention suggestions
1697
1698## 1.0.63
1699
1700- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1701
1702## 1.0.62
1703
1704- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1705- Hooks: Added SessionStart hook for new session initialization
1706- /add-dir command now supports typeahead for directory paths
1707- Improved network connectivity check reliability
1708
1709## 1.0.61
1710
1711- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1712- Settings: Added `--settings` flag to load settings from a JSON file
1713- Settings: Fixed resolution of settings files paths that are symlinks
1714- OTEL: Fixed reporting of wrong organization after authentication changes
1715- Slash commands: Fixed permissions checking for allowed-tools with Bash
1716- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1717- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1718- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1719
1720## 1.0.60
1721
1722- You can now create custom subagents for specialized tasks! Run /agents to get started
1723
1724## 1.0.59
1725
1726- SDK: Added tool confirmation support with canUseTool callback
1727- SDK: Allow specifying env for spawned process
1728- Hooks: Exposed PermissionDecision to hooks (including "ask")
1729- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1730- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1731
1732## 1.0.58
1733
1734- Added support for reading PDFs
1735- MCP: Improved server health status display in 'claude mcp list'
1736- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1737
1738## 1.0.57
1739
1740- Added support for specifying a model in slash commands
1741- Improved permission messages to help Claude understand allowed tools
1742- Fix: Remove trailing newlines from bash output in terminal wrapping
1743
1744## 1.0.56
1745
1746- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1747- Fixes for WSL IDE detection
1748- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1749
1750## 1.0.55
1751
1752- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1753- Windows: fixed Ctrl+Z crash
1754- SDK: Added ability to capture error logging
1755- Add --system-prompt-file option to override system prompt in print mode
1756
1757## 1.0.54
1758
1759- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1760- Custom slash commands: Added argument-hint to frontmatter
1761- Windows: OAuth uses port 45454 and properly constructs browser URL
1762- Windows: mode switching now uses alt + m, and plan mode renders properly
1763- Shell: Switch to in-memory shell snapshot to fix file-related errors
1764
1765## 1.0.53
1766
1767- Updated @-mention file truncation from 100 lines to 2000 lines
1768- 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).
1769
1770## 1.0.52
1771
1772- Added support for MCP server instructions
1773
1774## 1.0.51
1775
1776- Added support for native Windows (requires Git for Windows)
1777- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1778- Settings: /doctor can now help you identify and fix invalid setting files
1779- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1780- Increased auto-compact warning threshold from 60% to 80%
1781- Fixed an issue with handling user directories with spaces for shell snapshots
1782- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1783- Custom slash commands: Fixed user-level commands in subdirectories
1784- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1785
1786## 1.0.48
1787
1788- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1789- Added progress messages to Bash tool based on the last 5 lines of command output
1790- Added expanding variables support for MCP server configuration
1791- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1792- Improved IDE extension path handling when Claude Code runs in WSL
1793- Hooks: Added a PreCompact hook
1794- Vim mode: Added c, f/F, t/T
1795
1796## 1.0.45
1797
1798- Redesigned Search (Grep) tool with new tool input parameters and features
1799- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1800- Fixed config file corruption issue by enforcing atomic writes
1801- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1802- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1803- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1804
1805## 1.0.44
1806
1807- New /export command lets you quickly export a conversation for sharing
1808- MCP: resource_link tool results are now supported
1809- MCP: tool annotations and tool titles now display in /mcp view
1810- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1811
1812## 1.0.43
1813
1814- Fixed a bug where the theme selector was saving excessively
1815- Hooks: Added EPIPE system error handling
1816
1817## 1.0.42
1818
1819- Added tilde (`~`) expansion support to `/add-dir` command
1820
1821## 1.0.41
1822
1823- Hooks: Split Stop hook triggering into Stop and SubagentStop
1824- Hooks: Enabled optional timeout configuration for each command
1825- Hooks: Added "hook_event_name" to hook input
1826- Fixed a bug where MCP tools would display twice in tool list
1827- New tool parameters JSON for Bash tool in `tool_decision` event
1828
1829## 1.0.40
1830
1831- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1832
1833## 1.0.39
1834
1835- New Active Time metric in OpenTelemetry logging
1836
1837## 1.0.38
1838
1839- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1840
1841## 1.0.37
1842
1843- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1844
1845## 1.0.36
1846
1847- Web search now takes today's date into context
1848- Fixed a bug where stdio MCP servers were not terminating properly on exit
1849
1850## 1.0.35
1851
1852- Added support for MCP OAuth Authorization Server discovery
1853
1854## 1.0.34
1855
1856- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1857
1858## 1.0.33
1859
1860- Improved logging functionality with session ID support
1861- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1862- Improvements to plan mode
1863
1864## 1.0.32
1865
1866- Updated loopback config for litellm
1867- Added forceLoginMethod setting to bypass login selection screen
1868
1869## 1.0.31
1870
1871- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1872
1873## 1.0.30
1874
1875- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1876- Improved file path autocomplete with filename matching
1877- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1878- Enhanced jq regex support for complex filters with pipes and select
1879
1880## 1.0.29
1881
1882- Improved CJK character support in cursor navigation and rendering
1883
1884## 1.0.28
1885
1886- Slash commands: Fix selector display during history navigation
1887- Resizes images before upload to prevent API size limit errors
1888- Added XDG_CONFIG_HOME support to configuration directory
1889- Performance optimizations for memory usage
1890- New attributes (terminal.type, language) in OpenTelemetry logging
1891
1892## 1.0.27
1893
1894- Streamable HTTP MCP servers are now supported
1895- Remote MCP servers (SSE and HTTP) now support OAuth
1896- MCP resources can now be @-mentioned
1897- /resume slash command to switch conversations within Claude Code
1898
1899## 1.0.25
1900
1901- Slash commands: moved "project" and "user" prefixes to descriptions
1902- Slash commands: improved reliability for command discovery
1903- Improved support for Ghostty
1904- Improved web search reliability
1905
1906## 1.0.24
1907
1908- Improved /mcp output
1909- Fixed a bug where settings arrays got overwritten instead of merged
1910
1911## 1.0.23
1912
1913- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1914- Released Python SDK: pip install claude-code-sdk to get started
1915
1916## 1.0.22
1917
1918- SDK: Renamed `total_cost` to `total_cost_usd`
1919
1920## 1.0.21
1921
1922- Improved editing of files with tab-based indentation
1923- Fix for tool_use without matching tool_result errors
1924- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1925
1926## 1.0.18
1927
1928- Added --add-dir CLI argument for specifying additional working directories
1929- Added streaming input support without require -p flag
1930- Improved startup performance and session storage performance
1931- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1932- Added detailed MCP server tools display (/mcp)
1933- MCP authentication and permission improvements
1934- Added auto-reconnection for MCP SSE connections on disconnect
1935- Fixed issue where pasted content was lost when dialogs appeared
1936
1937## 1.0.17
1938
1939- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1940- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1941- MCP server list UI improvements
1942- Update Claude Code process title to display "claude" instead of "node"
1943
1944## 1.0.11
1945
1946- Claude Code can now also be used with a Claude Pro subscription
1947- Added /upgrade for smoother switching to Claude Max plans
1948- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1949- Improved shell configuration error handling
1950- Improved todo list handling during compaction
1951
1952## 1.0.10
1953
1954- Added markdown table support
1955- Improved streaming performance
1956
1957## 1.0.8
1958
1959- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1960- Increased default otel interval from 1s -> 5s
1961- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1962- Fixed a regression where search tools unnecessarily asked for permissions
1963- Added support for triggering thinking non-English languages
1964- Improved compacting UI
1965
1966## 1.0.7
1967
1968- Renamed /allowed-tools -> /permissions
1969- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1970- Deprecated claude config commands in favor of editing settings.json
1971- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1972- Improved error handling for /install-github-app
1973- Bugfixes, UI polish, and tool reliability improvements
1974
1975## 1.0.6
1976
1977- Improved edit reliability for tab-indented files
1978- Respect CLAUDE_CONFIG_DIR everywhere
1979- Reduced unnecessary tool permission prompts
1980- Added support for symlinks in @file typeahead
1981- Bugfixes, UI polish, and tool reliability improvements
1982
1983## 1.0.4
1984
1985- Fixed a bug where MCP tool errors weren't being parsed correctly
1986
1987## 1.0.1
1988
1989- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
1990- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
1991- Updated documentation links and OAuth process descriptions
1992
1993## 1.0.0
1994
1995- Claude Code is now generally available
1996- Introducing Sonnet 4 and Opus 4 models
1997
1998## 0.2.125
1999
2000- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
2001- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
2002
2003## 0.2.117
2004
2005- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
2006- Introduced settings.cleanupPeriodDays
2007- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
2008- Introduced --debug mode
2009
2010## 0.2.108
2011
2012- You can now send messages to Claude while it works to steer Claude in real-time
2013- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
2014- Fixed a bug where thinking was not working in -p mode
2015- Fixed a regression in /cost reporting
2016- Deprecated MCP wizard interface in favor of other MCP commands
2017- Lots of other bugfixes and improvements
2018
2019## 0.2.107
2020
2021- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
2022
2023## 0.2.106
2024
2025- MCP SSE server configs can now specify custom headers
2026- Fixed a bug where MCP permission prompt didn't always show correctly
2027
2028## 0.2.105
2029
2030- Claude can now search the web
2031- Moved system & account status to /status
2032- Added word movement keybindings for Vim
2033- Improved latency for startup, todo tool, and file edits
2034
2035## 0.2.102
2036
2037- Improved thinking triggering reliability
2038- Improved @mention reliability for images and folders
2039- You can now paste multiple large chunks into one prompt
2040
2041## 0.2.100
2042
2043- Fixed a crash caused by a stack overflow error
2044- Made db storage optional; missing db support disables --continue and --resume
2045
2046## 0.2.98
2047
2048- Fixed an issue where auto-compact was running twice
2049
2050## 0.2.96
2051
2052- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
2053
2054## 0.2.93
2055
2056- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
2057- Claude now has access to a Todo list that helps it stay on track and be more organized
2058
2059## 0.2.82
2060
2061- Added support for --disallowedTools
2062- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
2063
2064## 0.2.75
2065
2066- Hit Enter to queue up additional messages while Claude is working
2067- Drag in or copy/paste image files directly into the prompt
2068- @-mention files to directly add them to context
2069- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
2070- Improved performance for filename auto-complete
2071
2072## 0.2.74
2073
2074- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
2075- Task tool can now perform writes and run bash commands
2076
2077## 0.2.72
2078
2079- Updated spinner to indicate tokens loaded and tool usage
2080
2081## 0.2.70
2082
2083- Network commands like curl are now available for Claude to use
2084- Claude can now run multiple web queries in parallel
2085- Pressing ESC once immediately interrupts Claude in Auto-accept mode
2086
2087## 0.2.69
2088
2089- Fixed UI glitches with improved Select component behavior
2090- Enhanced terminal output display with better text truncation logic
2091
2092## 0.2.67
2093
2094- Shared project permission rules can be saved in .claude/settings.json
2095
2096## 0.2.66
2097
2098- Print mode (-p) now supports streaming output via --output-format=stream-json
2099- Fixed issue where pasting could trigger memory or bash mode unexpectedly
2100
2101## 0.2.63
2102
2103- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
2104
2105## 0.2.61
2106
2107- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
2108- Enhanced image detection for more reliable clipboard paste functionality
2109- Fixed an issue where ESC key could crash the conversation history selector
2110
2111## 0.2.59
2112
2113- Copy+paste images directly into your prompt
2114- Improved progress indicators for bash and fetch tools
2115- Bugfixes for non-interactive mode (-p)
2116
2117## 0.2.54
2118
2119- Quickly add to Memory by starting your message with '#'
2120- Press ctrl+r to see full output for long tool results
2121- Added support for MCP SSE transport
2122
2123## 0.2.53
2124
2125- New web fetch tool lets Claude view URLs that you paste in
2126- Fixed a bug with JPEG detection
2127
2128## 0.2.50
2129
2130- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
2131
2132## 0.2.49
2133
2134- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
2135
2136## 0.2.47
2137
2138- Press Tab to auto-complete file and folder names
2139- Press Shift + Tab to toggle auto-accept for file edits
2140- Automatic conversation compaction for infinite conversation length (toggle with /config)
2141
2142## 0.2.44
2143
2144- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
2145
2146## 0.2.41
2147
2148- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
2149- MCP server startup no longer blocks the app from starting up
2150
2151## 0.2.37
2152
2153- New /release-notes command lets you view release notes at any time
2154- `claude config add/remove` commands now accept multiple values separated by commas or spaces
2155
2156## 0.2.36
2157
2158- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
2159- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
2160
2161## 0.2.34
2162
2163- Vim bindings for text input - enable with /vim or /config
2164
2165## 0.2.32
2166
2167- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
2168- Fix for some PersistentShell issues
2169
2170## 0.2.31
2171
2172- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
2173- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
2174
2175## 0.2.30
2176
2177- Added ANSI color theme for better terminal compatibility
2178- Fixed issue where slash command arguments weren't being sent properly
2179- (Mac-only) API keys are now stored in macOS Keychain
2180
2181## 0.2.26
2182
2183- New /approved-tools command for managing tool permissions
2184- Word-level diff display for improved code readability
2185- Fuzzy matching for slash commands
2186
2187## 0.2.21
2188
2189- Fuzzy matching for /commands