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 45b5430 · on Mar 11, 2026 History
2,215 lines · 140.9 KBPreviewCodeRaw
1# Changelog
2
3## 2.1.73
4
5- Added `modelOverrides` setting to map model picker entries to custom provider model IDs (e.g. Bedrock inference profile ARNs)
6- Added actionable guidance when OAuth login or connectivity checks fail due to SSL certificate errors (corporate proxies, `NODE_EXTRA_CA_CERTS`)
7- Fixed freezes and 100% CPU loops triggered by permission prompts for complex bash commands
8- Fixed a deadlock that could freeze Claude Code when many skill files changed at once (e.g. during `git pull` in a repo with a large `.claude/skills/` directory)
9- Fixed Bash tool output being lost when running multiple Claude Code sessions in the same project directory
10- Fixed subagents with `model: opus`/`sonnet`/`haiku` being silently downgraded to older model versions on Bedrock, Vertex, and Microsoft Foundry
11- Fixed background bash processes spawned by subagents not being cleaned up when the agent exits
12- Fixed `/resume` showing the current session in the picker
13- Fixed `/ide` crashing with `onInstall is not defined` when auto-installing the extension
14- Fixed `/loop` not being available on Bedrock/Vertex/Foundry and when telemetry was disabled
15- Fixed SessionStart hooks firing twice when resuming a session via `--resume` or `--continue`
16- Fixed JSON-output hooks injecting no-op system-reminder messages into the model's context on every turn
17- Fixed voice mode session corruption when a slow connection overlaps a new recording
18- Fixed Linux sandbox failing to start with "ripgrep (rg) not found" on native builds
19- Fixed Linux native modules not loading on Amazon Linux 2 and other glibc 2.26 systems
20- Fixed "media_type: Field required" API error when receiving images via Remote Control
21- Fixed `/heapdump` failing on Windows with `EEXIST` error when the Desktop folder already exists
22- Improved Up arrow after interrupting Claude — now restores the interrupted prompt and rewinds the conversation in one step
23- Improved IDE detection speed at startup
24- Improved clipboard image pasting performance on macOS
25- Improved `/effort` to work while Claude is responding, matching `/model` behavior
26- Improved voice mode to automatically retry transient connection failures during rapid push-to-talk re-press
27- Improved the Remote Control spawn mode selection prompt with better context
28- Changed default Opus model on Bedrock, Vertex, and Microsoft Foundry to Opus 4.6 (was Opus 4.1)
29- Deprecated `/output-style` command — use `/config` instead. Output style is now fixed at session start for better prompt caching
30- VSCode: Fixed HTTP 400 errors for users behind proxies or on Bedrock/Vertex with Claude 4.5 models
31
32## 2.1.72
33
34- Fixed tool search to activate even with `ANTHROPIC_BASE_URL` as long as `ENABLE_TOOL_SEARCH` is set.
35- Added `w` key in `/copy` to write the focused selection directly to a file, bypassing the clipboard (useful over SSH)
36- Added optional description argument to `/plan` (e.g., `/plan fix the auth bug`) that enters plan mode and immediately starts
37- Added `ExitWorktree` tool to leave an `EnterWorktree` session
38- Added `CLAUDE_CODE_DISABLE_CRON` environment variable to immediately stop scheduled cron jobs mid-session
39- Added `lsof`, `pgrep`, `tput`, `ss`, `fd`, and `fdfind` to the bash auto-approval allowlist, reducing permission prompts for common read-only operations
40- Restored the `model` parameter on the Agent tool for per-invocation model overrides
41- 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
42- Improved `/config` — Escape now cancels changes, Enter saves and closes, Space toggles settings
43- Improved up-arrow history to show current session's messages first when running multiple concurrent sessions
44- Improved voice input transcription accuracy for repo names and common dev terms (regex, OAuth, JSON)
45- Improved bash command parsing by switching to a native module — faster initialization and no memory leak
46- Reduced bundle size by ~510 KB
47- Changed CLAUDE.md HTML comments (`<!-- ... -->`) to be hidden from Claude when auto-injected. Comments remain visible when read with the Read tool
48- Fixed slow exits when background tasks or hooks were slow to respond
49- Fixed agent task progress stuck on "Initializing…"
50- Fixed skill hooks firing twice per event when a hooks-enabled skill is invoked by the model
51- 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
52- Fixed `--continue` not resuming from the most recent point after `--compact`
53- Fixed bash security parsing edge cases
54- Added support for marketplace git URLs without `.git` suffix (Azure DevOps, AWS CodeCommit)
55- Improved marketplace clone failure messages to show diagnostic info even when git produces no stderr
56- 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
57- Fixed feedback survey appearing too frequently in long sessions
58- Fixed `--effort` CLI flag being reset by unrelated settings writes on startup
59- Fixed backgrounded Ctrl+B queries losing their transcript or corrupting the new conversation after `/clear`
60- Fixed `/clear` killing background agent/bash tasks — only foreground tasks are now cleared
61- Fixed worktree isolation issues: Task tool resume not restoring cwd, and background task notifications missing `worktreePath` and `worktreeBranch`
62- Fixed `/model` not displaying results when run while Claude is working
63- Fixed digit keys selecting menu options instead of typing in plan mode permission prompt's text input
64- Fixed sandbox permission issues: certain file write operations incorrectly allowed without prompting, and output redirections to allowlisted directories (like `/tmp/claude/`) prompting unnecessarily
65- Improved CPU utilization in long sessions
66- Fixed prompt cache invalidation in SDK `query()` calls, reducing input token costs up to 12x
67- Fixed Escape key becoming unresponsive after cancelling a query
68- Fixed double Ctrl+C not exiting when background agents or tasks are running
69- Fixed team agents to inherit the leader's model
70- Fixed "Always Allow" saving permission rules that never match again
71- 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
72- Fixed session crashes in Desktop/SDK when Read returned files containing U+2028/U+2029 characters
73- Fixed terminal title being cleared on exit even when `CLAUDE_CODE_DISABLE_TERMINAL_TITLE` was set
74- 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
75- Fixed oversized and truncated images from Bash data-URL output
76- Fixed a crash when resuming sessions that contained Bedrock API errors
77- Fixed intermittent "expected boolean, received string" validation errors on Edit, Bash, and Grep tool inputs
78- Fixed multi-line session titles when forking from a conversation whose first message contained newlines
79- Fixed queued messages not showing attached images, and images being lost when pressing ↑ to edit a queued message
80- Fixed parallel tool calls where a failed Read/WebFetch/Glob would cancel its siblings — only Bash errors now cascade
81- VSCode: Fixed scroll speed in integrated terminals not matching native terminals
82- VSCode: Fixed Shift+Enter submitting input instead of inserting a newline for users with older keybindings
83- VSCode: Added effort level indicator on the input border
84- VSCode: Added `vscode://anthropic.claude-code/open` URI handler to open a new Claude Code tab programmatically, with optional `prompt` and `session` query parameters
85
86## 2.1.71
87
88- Added `/loop` command to run a prompt or slash command on a recurring interval (e.g. `/loop 5m check the deploy`)
89- Added cron scheduling tools for recurring prompts within a session
90- 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
91- Added `fmt`, `comm`, `cmp`, `numfmt`, `expr`, `test`, `printf`, `getconf`, `seq`, `tsort`, and `pr` to the bash auto-approval allowlist
92- Fixed stdin freeze in long-running sessions where keystrokes stop being processed but the process stays alive
93- Fixed a 5–8 second startup freeze for users with voice mode enabled, caused by CoreAudio initialization blocking the main thread after system wake
94- Fixed startup UI freeze when many claude.ai proxy connectors refresh an expired OAuth token simultaneously
95- Fixed forked conversations (`/fork`) sharing the same plan file, which caused plan edits in one fork to overwrite the other
96- Fixed the Read tool putting oversized images into context when image processing failed, breaking subsequent turns in long image-heavy sessions
97- Fixed false-positive permission prompts for compound bash commands containing heredoc commit messages
98- Fixed plugin installations being lost when running multiple Claude Code instances
99- Fixed claude.ai connectors failing to reconnect after OAuth token refresh
100- Fixed claude.ai MCP connector startup notifications appearing for every org-configured connector instead of only previously connected ones
101- Fixed background agent completion notifications missing the output file path, which made it difficult for parent agents to recover agent results after context compaction
102- Fixed duplicate output in Bash tool error messages when commands exit with non-zero status
103- Fixed Chrome extension auto-detection getting permanently stuck on "not installed" after running on a machine without local Chrome
104- Fixed `/plugin marketplace update` failing with merge conflicts when the marketplace is pinned to a branch/tag ref
105- Fixed `/plugin marketplace add owner/repo@ref` incorrectly parsing `@` — previously only `#` worked as a ref separator, causing undiagnosable errors with `strictKnownMarketplaces`
106- Fixed duplicate entries in `/permissions` Workspace tab when the same directory is added with and without a trailing slash
107- Fixed `--print` hanging forever when team agents are configured — the exit loop no longer waits on long-lived `in_process_teammate` tasks
108- Fixed "❯ Tool loaded." appearing in the REPL after every `ToolSearch` call
109- Fixed prompting for `cd <cwd> && git ...` on Windows when the model uses a mingw-style path
110- Improved startup time by deferring native image processor loading to first use
111- Improved bridge session reconnection to complete within seconds after laptop wake from sleep, instead of waiting up to 10 minutes
112- 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
113- 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.
114- Updated `/debug` to toggle debug logging on mid-session, since debug logs are no longer written by default
115- Removed startup notification noise for unauthenticated org-registered claude.ai connectors
116
117## 2.1.70
118
119- 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
120- 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
121- 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
122- Fixed prompt-cache bust when an MCP server with `instructions` connects after the first turn
123- Fixed Enter inserting a newline instead of submitting when typing over a slow SSH connection
124- Fixed clipboard corrupting non-ASCII text (CJK, emoji) on Windows/WSL by using PowerShell `Set-Clipboard`
125- Fixed extra VS Code windows opening at startup on Windows when running from the VS Code integrated terminal
126- Fixed voice mode failing on Windows native binary with "native audio module could not be loaded"
127- Fixed push-to-talk not activating on session start when `voiceEnabled: true` was set in settings
128- Fixed markdown links containing `#NNN` references incorrectly pointing to the current repository instead of the linked URL
129- Fixed repeated "Model updated to Opus 4.6" notification when a project's `.claude/settings.json` has a legacy Opus model string pinned
130- Fixed plugins showing as inaccurately installed in `/plugin`
131- Fixed plugins showing "not found in marketplace" errors on fresh startup by auto-refreshing after marketplace installation
132- Fixed `/security-review` command failing with `unknown option merge-base` on older git versions
133- 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
134- Fixed a performance regression in the `AskUserQuestion` preview dialog that re-ran markdown rendering on every keystroke in the notes input
135- Fixed feature flags read during early startup never refreshing their disk cache, causing stale values to persist across sessions
136- Fixed `permissions.defaultMode` settings values other than `acceptEdits` or `plan` being applied in Claude Code Remote environments — they are now ignored
137- Fixed skill listing being re-injected on every `--resume` (~600 tokens saved per resume)
138- Fixed teleport marker not rendering in VS Code teleported sessions
139- Improved error message when microphone captures silence to distinguish from "no speech detected"
140- Improved compaction to preserve images in the summarizer request, allowing prompt cache reuse for faster and cheaper compaction
141- Improved `/rename` to work while Claude is processing, instead of being silently queued
142- Reduced prompt input re-renders during turns by ~74%
143- Reduced startup memory by ~426KB for users without custom CA certificates
144- 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.
145- [VSCode] Added spark icon in VS Code activity bar that lists all Claude Code sessions, with sessions opening as full editors
146- [VSCode] Added full markdown document view for plans in VS Code, with support for adding comments to provide feedback
147- [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
148
149## 2.1.69
150
151- Added the `/claude-api` skill for building applications with the Claude API and Anthropic SDK
152- Added Ctrl+U on an empty bash prompt (`!`) to exit bash mode, matching `escape` and `backspace`
153- Added numeric keypad support for selecting options in Claude's interview questions (previously only the number row above QWERTY worked)
154- 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
155- Added Voice STT support for 10 new languages (20 total) — Russian, Polish, Turkish, Dutch, Ukrainian, Greek, Czech, Danish, Swedish, Norwegian
156- Added effort level display (e.g., "with low effort") to the logo and spinner, making it easier to see which effort setting is active
157- Added agent name display in terminal title when using `claude --agent`
158- 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`
159- 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
160- Added `/reload-plugins` command to activate pending plugin changes without restarting
161- Added a one-time startup prompt suggesting Claude Code Desktop on macOS and Windows (max 3 showings, dismissible)
162- Added `${CLAUDE_SKILL_DIR}` variable for skills to reference their own directory in SKILL.md content
163- Added `InstructionsLoaded` hook event that fires when CLAUDE.md or `.claude/rules/*.md` files are loaded into context
164- Added `agent_id` (for subagents) and `agent_type` (for subagents and `--agent`) to hook events
165- Added `worktree` field to status line hook commands with name, path, branch, and original repo directory when running in a `--worktree` session
166- Added `pluginTrustMessage` in managed settings to append organization-specific context to the plugin trust warning shown before installation
167- Added policy limit fetching (e.g., remote control restrictions) for Team plan OAuth users, not just Enterprise
168- Added `pathPattern` to `strictKnownMarketplaces` for regex-matching file/directory marketplace sources alongside `hostPattern` restrictions
169- Added plugin source type `git-subdir` to point to a subdirectory within a git repo
170- Added `oauth.authServerMetadataUrl` config option for MCP servers to specify a custom OAuth metadata discovery URL when standard discovery fails
171- Fixed a security issue where nested skill discovery could load skills from gitignored directories like `node_modules`
172- Fixed trust dialog silently enabling all `.mcp.json` servers on first run. You'll now see the per-server approval dialog as expected
173- Fixed `claude remote-control` crashing immediately on npm installs with "bad option: --sdk-url" (anthropics/claude-code#28334)
174- Fixed `--model claude-opus-4-0` and `--model claude-opus-4-1` resolving to deprecated Opus versions instead of current
175- 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.
176- 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)
177- Fixed ghost dotfiles (`.bashrc`, `HEAD`, etc.) appearing as untracked files in the working directory after sandboxed Bash commands on Linux
178- Fixed Shift+Enter printing `[27;2;13~` instead of inserting a newline in Ghostty over SSH
179- Fixed stash (Ctrl+S) being cleared when submitting a message while Claude is working
180- Fixed ctrl+o (transcript toggle) freezing for many seconds in long sessions with lots of file edits
181- Fixed plan mode feedback input not supporting multi-line text entry (backslash+Enter and Shift+Enter now insert newlines)
182- Fixed cursor not moving down into blank lines at the top of the input box
183- Fixed `/stats` crash when transcript files contain entries with missing or malformed timestamps
184- 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)
185- Fixed `--setting-sources user` not blocking dynamically discovered project skills
186- Fixed duplicate CLAUDE.md, slash commands, agents, and rules when running from a worktree nested inside its main repo (e.g. `claude -w`)
187- Fixed plugin Stop/SessionEnd/etc hooks not firing after any `/plugin` operation
188- Fixed plugin hooks being silently dropped when two plugins use the same `${CLAUDE_PLUGIN_ROOT}/...` command template
189- Fixed memory leak in long-running SDK/CCR sessions where conversation messages were retained unnecessarily
190- Fixed API 400 errors in forked agents (autocompact, summarization) when resuming sessions that were interrupted mid-tool-batch
191- Fixed "unexpected tool_use_id found in tool_result blocks" error when resuming conversations that start with an orphaned tool result
192- Fixed teammates accidentally spawning nested teammates via the Agent tool's `name` parameter
193- Fixed `CLAUDE_CODE_MAX_OUTPUT_TOKENS` being ignored during conversation compaction
194- Fixed `/compact` summary rendering as a user bubble in SDK consumers (Claude Code Remote web UI, VSCode extension)
195- Fixed voice space bar getting stuck after a failed voice activation (module loading race, cold GrowthBook)
196- Fixed worktree file copy on Windows
197- Fixed global `.claude` folder detection on Windows
198- Fixed symlink bypass where writing new files through a symlinked parent directory could escape the working directory in `acceptEdits` mode
199- 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
200- 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
201- Fixed multi-GB memory spike when committing with large untracked binary files in the working tree
202- 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.
203- Fixed Android app crash when running local slash commands (`/voice`, `/cost`) in Remote Control sessions
204- Fixed a memory leak where old message array versions accumulated in React Compiler `memoCache` over long sessions
205- Fixed a memory leak where REPL render scopes accumulated over long sessions (~35MB over 1000 turns)
206- 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
207- Fixed a memory leak in interactive mode where hook events could accumulate unboundedly during long sessions
208- Fixed hang when `--mcp-config` points to a corrupted file
209- Fixed slow startup when many skills/plugins are installed
210- Fixed `cd <outside-dir> && <cmd>` permission prompt to surface the chained command instead of only showing "Yes, allow reading from <dir>/"
211- Fixed conditional `.claude/rules/*.md` files (with `paths:` frontmatter) and nested CLAUDE.md files not loading in print mode (`claude -p`)
212- Fixed `/clear` not fully clearing all session caches, reducing memory retention in long sessions
213- Fixed terminal flicker caused by animated elements at the scrollback boundary
214- Fixed UI frame drops on macOS when using MCP servers with OAuth (regression from 2.1.x)
215- Fixed occasional frame stalls during typing caused by synchronous debug log flushes
216- Fixed `TeammateIdle` and `TaskCompleted` hooks to support `{"continue": false, "stopReason": "..."}` to stop the teammate, matching `Stop` hook behavior
217- Fixed `WorktreeCreate` and `WorktreeRemove` plugin hooks being silently ignored
218- Fixed skill descriptions with colons (e.g., "Triggers include: X, Y, Z") failing to load from SKILL.md frontmatter
219- Fixed project skills without a `description:` frontmatter field not appearing in Claude's available skills list
220- Fixed `/context` showing identical token counts for all MCP tools from a server
221- Fixed literal `nul` file creation on Windows when the model uses CMD-style `2>nul` redirection in Git Bash
222- Fixed extra blank lines appearing below each tool call in the expanded subagent transcript view (Ctrl+O)
223- Fixed Tab/arrow keys not cycling Settings tabs when `/config` search box is focused but empty
224- Fixed service key OAuth sessions (CCR containers) spamming `[ERROR]` logs with 403s from profile-scoped endpoints
225- Fixed inconsistent color for "Remote Control active" status indicator
226- Fixed Voice waveform cursor covering the first suffix letter when dictating mid-input
227- Fixed Voice input showing all 5 spaces during warmup instead of capping at ~2 (aligning with the "keep holding…" hint)
228- Improved spinner performance by isolating the 50ms animation loop from the surrounding shell, reducing render and CPU overhead during turns
229- Improved UI rendering performance in native binaries with React Compiler
230- Improved `--worktree` startup by eliminating a git subprocess on the startup path
231- Improved macOS startup by eliminating redundant settings-file reloads when managed settings resolve
232- Improved macOS startup for Claude.ai enterprise/team users by skipping an unnecessary keychain lookup
233- Improved MCP `-p` startup by pipelining claude.ai config fetch with local connections and using a concurrency pool instead of sequential batching
234- Improved voice startup by removing imperceptible warmup pulse animations that were causing re-render stutter
235- 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.
236- Improved memory usage in long sessions by stabilizing `onSubmit` across message updates
237- Improved LSP tool rendering and memory context building to no longer read entire files
238- Improved session upload and memory sync to avoid reading large files into memory before size/binary checks
239- Improved file operation performance by avoiding reading file contents for existence checks (6 sites)
240- Improved documentation to clarify that `--append-system-prompt-file` and `--system-prompt-file` work in interactive mode (the docs previously said print mode only)
241- Reduced baseline memory by ~16MB by deferring Yoga WASM preloading
242- Reduced memory footprint for SDK and CCR sessions using stream-json output
243- Reduced memory usage when resuming large sessions (including compacted history)
244- Reduced token usage on multi-agent tasks with more concise subagent final reports
245- Changed Sonnet 4.5 users on Pro/Max/Team Premium to be automatically migrated to Sonnet 4.6
246- Changed the `/resume` picker to show your most recent prompt instead of the first one. This also resolves some titles appearing as `(session)`.
247- Changed claude.ai MCP connector failures to show a notification instead of silently disappearing from the tool list
248- Changed example command suggestions to be generated deterministically instead of calling Haiku
249- Changed resuming after compaction to no longer produce a preamble recap before continuing
250- [SDK] Changed task creation to no longer require the `activeForm` field — the spinner falls back to the task subject
251- [VSCode] Added compaction display as a collapsible "Compacted chat" card with the summary inside
252- [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
253- [VSCode] Fixed RTL text (Arabic, Hebrew, Persian) rendering reversed in the chat panel (regression in v2.1.63)
254
255## 2.1.68
256
257- 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`
258- Re-introduced the "ultrathink" keyword to enable high effort for the next turn
259- 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
260
261## 2.1.66
262
263- Reduced spurious error logging
264
265## 2.1.63
266
267- Added `/simplify` and `/batch` bundled slash commands
268- Fixed local slash command output like /cost appearing as user-sent messages instead of system messages in the UI
269- Project configs & auto memory now shared across git worktrees of the same repository
270- Added `ENABLE_CLAUDEAI_MCP_SERVERS=false` env var to opt out from making claude.ai MCP servers available
271- Improved `/model` command to show the currently active model in the slash command menu
272- Added HTTP hooks, which can POST JSON to a URL and receive JSON instead of running a shell command
273- Fixed listener leak in bridge polling loop
274- Fixed listener leak in MCP OAuth flow cleanup
275- 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.
276- Fixed memory leak when navigating hooks configuration menu
277- Fixed listener leak in interactive permission handler during auto-approvals
278- Fixed file count cache ignoring glob ignore patterns
279- Fixed memory leak in bash command prefix cache
280- Fixed MCP tool/resource cache leak on server reconnect
281- Fixed IDE host IP detection cache incorrectly sharing results across ports
282- Fixed WebSocket listener leak on transport reconnect
283- Fixed memory leak in git root detection cache that could cause unbounded growth in long-running sessions
284- Fixed memory leak in JSON parsing cache that grew unbounded over long sessions
285- VSCode: Fixed remote sessions not appearing in conversation history
286- 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.
287- Fixed memory leak where long-running teammates retained all messages in AppState even after conversation compaction
288- Fixed a memory leak where MCP server fetch caches were not cleared on disconnect, causing growing memory usage with servers that reconnect frequently
289- Improved memory usage in long sessions with subagents by stripping heavy progress message payloads during context compaction
290- 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.
291- VSCode: Added session rename and remove actions to the sessions list
292- Fixed `/clear` not resetting cached skills, which could cause stale skill content to persist in the new conversation
293
294## 2.1.62
295
296- Fixed prompt suggestion cache regression that reduced cache hit rates
297
298## 2.1.61
299
300- Fixed concurrent writes corrupting config file on Windows
301
302## 2.1.59
303
304- Claude automatically saves useful context to auto-memory. Manage with /memory
305- Added `/copy` command to show an interactive picker when code blocks are present, allowing selection of individual code blocks or the full response.
306- 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
307- Improved ordering of short task lists
308- Improved memory usage in multi-agent sessions by releasing completed subagent task state
309- Fixed MCP OAuth token refresh race condition when running multiple Claude Code instances simultaneously
310- Fixed shell commands not showing a clear error message when the working directory has been deleted
311- Fixed config file corruption that could wipe authentication when multiple Claude Code instances ran simultaneously
312
313## 2.1.58
314
315- Expand Remote Control to more users
316
317## 2.1.56
318
319- VS Code: Fixed another cause of "command 'claude-vscode.editor.openLast' not found" crashes
320
321## 2.1.55
322
323- Fixed BashTool failing on Windows with EINVAL error
324
325## 2.1.53
326
327- Fixed a UI flicker where user input would briefly disappear after submission before the message rendered
328- Fixed bulk agent kill (ctrl+f) to send a single aggregate notification instead of one per agent, and to properly clear the command queue
329- Fixed graceful shutdown sometimes leaving stale sessions when using Remote Control by parallelizing teardown network calls
330- Fixed `--worktree` sometimes being ignored on first launch
331- Fixed a panic ("switch on corrupted value") on Windows
332- Fixed a crash that could occur when spawning many processes on Windows
333- Fixed a crash in the WebAssembly interpreter on Linux x64 & Windows x64
334- Fixed a crash that sometimes occurred after 2 minutes on Windows ARM64
335
336## 2.1.52
337
338- VS Code: Fixed extension crash on Windows ("command 'claude-vscode.editor.openLast' not found")
339
340## 2.1.51
341
342- Added `claude remote-control` subcommand for external builds, enabling local environment serving for all users.
343- Updated plugin marketplace default git timeout from 30s to 120s and added `CLAUDE_CODE_PLUGIN_GIT_TIMEOUT_MS` to configure.
344- Added support for custom npm registries and specific version pinning when installing plugins from npm sources
345- 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`.
346- Fixed a security issue where `statusLine` and `fileSuggestion` hook commands could execute without workspace trust acceptance in interactive mode.
347- Tool results larger than 50K characters are now persisted to disk (previously 100K). This reduces context window usage and improves conversation longevity.
348- 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.
349- 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.
350- Fixed slash command autocomplete crashing when a plugin's SKILL.md description is a YAML array or other non-string type
351- 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.
352- Managed settings can now be set via macOS plist or Windows Registry. Learn more at https://code.claude.com/docs/en/settings#settings-files
353
354## 2.1.50
355
356- Added support for `startupTimeout` configuration for LSP servers
357- Added `WorktreeCreate` and `WorktreeRemove` hook events, enabling custom VCS setup and teardown when agent worktree isolation creates or removes worktrees.
358- 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.
359- Linux: Fixed native modules not loading on systems with glibc older than 2.30 (e.g., RHEL 8)
360- Fixed memory leak in agent teams where completed teammate tasks were never garbage collected from session state
361- Fixed `CLAUDE_CODE_SIMPLE` to fully strip down skills, session memory, custom agents, and CLAUDE.md token counting
362- Fixed `/mcp reconnect` freezing the CLI when given a server name that doesn't exist
363- Fixed memory leak where completed task state objects were never removed from AppState
364- Added support for `isolation: worktree` in agent definitions, allowing agents to declaratively run in isolated git worktrees.
365- `CLAUDE_CODE_SIMPLE` mode now also disables MCP tools, attachments, hooks, and CLAUDE.md file loading for a fully minimal experience.
366- Fixed bug where MCP tools were not discovered when tool search is enabled and a prompt is passed in as a launch argument
367- Improved memory usage during long sessions by clearing internal caches after compaction
368- Added `claude agents` CLI command to list all configured agents
369- Improved memory usage during long sessions by clearing large tool results after they have been processed
370- Fixed a memory leak where LSP diagnostic data was never cleaned up after delivery, causing unbounded memory growth in long sessions
371- Fixed a memory leak where completed task output was not freed from memory, reducing memory usage in long sessions with many tasks
372- Improved startup performance for headless mode (`-p` flag) by deferring Yoga WASM and UI component imports
373- Fixed prompt suggestion cache regression that reduced cache hit rates
374- Fixed unbounded memory growth in long sessions by capping file history snapshots
375- Added `CLAUDE_CODE_DISABLE_1M_CONTEXT` environment variable to disable 1M context window support
376- Opus 4.6 (fast mode) now includes the full 1M context window
377- VSCode: Added `/extra-usage` command support in VS Code sessions
378- Fixed memory leak where TaskOutput retained recent lines after cleanup
379- Fixed memory leak in CircularBuffer where cleared items were retained in the backing array
380- Fixed memory leak in shell command execution where ChildProcess and AbortController references were retained after cleanup
381
382## 2.1.49
383
384- Improved MCP OAuth authentication with step-up auth support and discovery caching, reducing redundant network requests during server connections
385- Added `--worktree` (`-w`) flag to start Claude in an isolated git worktree
386- Subagents support `isolation: "worktree"` for working in a temporary git worktree
387- Added Ctrl+F keybinding to kill background agents (two-press confirmation)
388- Agent definitions support `background: true` to always run as a background task
389- Plugins can ship `settings.json` for default configuration
390- Fixed file-not-found errors to suggest corrected paths when the model drops the repo folder
391- 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.
392- Fixed prompt suggestion cache regression that reduced cache hit rates.
393- Fixed `plugin enable` and `plugin disable` to auto-detect the correct scope when `--scope` is not specified, instead of always defaulting to user scope
394- Simple mode (`CLAUDE_CODE_SIMPLE`) now includes the file edit tool in addition to the Bash tool, allowing direct file editing in simple mode.
395- Permission suggestions are now populated when safety checks trigger an ask response, enabling SDK consumers to display permission options
396- 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.
397- Fixed verbose mode not updating thinking block display when toggled via `/config` — memo comparators now correctly detect verbose changes
398- Fixed unbounded WASM memory growth during long sessions by periodically resetting the tree-sitter parser
399- Fixed potential rendering issues caused by stale yoga layout references
400- Improved performance in non-interactive mode (`-p`) by skipping unnecessary API calls during startup
401- Improved performance by caching authentication failures for HTTP and SSE MCP servers, avoiding repeated connection attempts to servers requiring auth
402- Fixed unbounded memory growth during long-running sessions caused by Yoga WASM linear memory never shrinking
403- SDK model info now includes `supportsEffort`, `supportedEffortLevels`, and `supportsAdaptiveThinking` fields so consumers can discover model capabilities.
404- Added `ConfigChange` hook event that fires when configuration files change during a session, enabling enterprise security auditing and optional blocking of settings changes.
405- Improved startup performance by caching MCP auth failures to avoid redundant connection attempts
406- Improved startup performance by reducing HTTP calls for analytics token counting
407- Improved startup performance by batching MCP tool token counting into a single API call
408- Fixed `disableAllHooks` setting to respect managed settings hierarchy — non-managed settings can no longer disable managed hooks set by policy (#26637)
409- 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.
410- 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
411
412## 2.1.47
413
414- Fixed FileWriteTool line counting to preserve intentional trailing blank lines instead of stripping them with `trimEnd()`.
415- 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.
416- 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.
417- 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.
418- 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)
419- Improved memory usage in long-running sessions by releasing API stream buffers, agent context, and skill state after use
420- Improved startup performance by deferring SessionStart hook execution, reducing time-to-interactive by ~500ms.
421- Fixed an issue where bash tool output was silently discarded on Windows when using MSYS2 or Cygwin shells.
422- 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.
423- Improved memory usage by trimming agent task message history after tasks complete
424- Improved memory usage during long agent sessions by eliminating O(n²) message accumulation in progress updates
425- Fixed the bash permission classifier to validate that returned match descriptions correspond to actual input rules, preventing hallucinated descriptions from incorrectly granting permissions
426- Fixed user-defined agents only loading one file on NFS/FUSE filesystems that report zero inodes (anthropics/claude-code#26044)
427- Fixed plugin agent skills silently failing to load when referenced by bare name instead of fully-qualified plugin name (anthropics/claude-code#25834)
428- Search patterns in collapsed tool results are now displayed in quotes for clarity
429- Windows: Fixed CWD tracking temp files never being cleaned up, causing them to accumulate indefinitely (anthropics/claude-code#17600)
430- 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.
431- 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.
432- Simplified teammate navigation to use only Shift+Down (with wrapping) instead of both Shift+Up and Shift+Down.
433- 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.
434- 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.
435- Fixed custom session titles set via `/rename` being lost after resuming a conversation (anthropics/claude-code#23610)
436- Fixed collapsed read/search hint text overflowing on narrow terminals by truncating from the start.
437- 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.
438- 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)
439- Fixed MCP servers not appearing in the MCP Management Dialog after deferred loading
440- Fixed session name persisting in status bar after `/clear` command (anthropics/claude-code#26082)
441- 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)
442- Fixed /resume silently dropping sessions when the first message exceeds 16KB or uses array-format content (anthropics/claude-code#25721)
443- Added `chat:newline` keybinding action for configurable multi-line input (anthropics/claude-code#26075)
444- Added `added_dirs` to the statusline JSON `workspace` section, exposing directories added via `/add-dir` to external scripts (anthropics/claude-code#26096)
445- Fixed `claude doctor` misclassifying mise and asdf-managed installations as native installs (anthropics/claude-code#26033)
446- Fixed zsh heredoc failing with "read-only file system" error in sandboxed commands (anthropics/claude-code#25990)
447- Fixed agent progress indicator showing inflated tool use count (anthropics/claude-code#26023)
448- Fixed image pasting not working on WSL2 systems where Windows copies images as BMP format (anthropics/claude-code#25935)
449- Fixed background agent results returning raw transcript data instead of the agent's final answer (anthropics/claude-code#26012)
450- Fixed Warp terminal incorrectly prompting for Shift+Enter setup when it supports it natively (anthropics/claude-code#25957)
451- Fixed CJK wide characters causing misaligned timestamps and layout elements in the TUI (anthropics/claude-code#26084)
452- Fixed custom agent `model` field in `.claude/agents/*.md` being ignored when spawning team teammates (anthropics/claude-code#26064)
453- Fixed plan mode being lost after context compaction, causing the model to switch from planning to implementation mode (anthropics/claude-code#26061)
454- Fixed `alwaysThinkingEnabled: true` in settings.json not enabling thinking mode on Bedrock and Vertex providers (anthropics/claude-code#26074)
455- Fixed `tool_decision` OTel telemetry event not being emitted in headless/SDK mode (anthropics/claude-code#26059)
456- Fixed session name being lost after context compaction — renamed sessions now preserve their custom title through compaction (anthropics/claude-code#26121)
457- Increased initial session count in resume picker from 10 to 50 for faster session discovery (anthropics/claude-code#26123)
458- Windows: fixed worktree session matching when drive letter casing differs (anthropics/claude-code#26123)
459- Fixed `/resume <session-id>` failing to find sessions whose first message exceeds 16KB (anthropics/claude-code#25920)
460- Fixed "Always allow" on multiline bash commands creating invalid permission patterns that corrupt settings (anthropics/claude-code#25909)
461- 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)
462- 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)
463- Fixed read-only git commands triggering FSEvents file watcher loops on macOS by adding --no-optional-locks flag (anthropics/claude-code#25750)
464- 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)
465- Fixed non-interactive subcommands like `claude doctor` and `claude plugin validate` being blocked inside nested Claude sessions (anthropics/claude-code#25803)
466- Windows: Fixed the same CLAUDE.md file being loaded twice when drive letter casing differs between paths (anthropics/claude-code#25756)
467- Fixed inline code spans in markdown being incorrectly parsed as bash commands (anthropics/claude-code#25792)
468- Fixed teammate spinners not respecting custom spinnerVerbs from settings (anthropics/claude-code#25748)
469- Fixed shell commands permanently failing after a command deletes its own working directory (anthropics/claude-code#26136)
470- Fixed hooks (PreToolUse, PostToolUse) silently failing to execute on Windows by using Git Bash instead of cmd.exe (anthropics/claude-code#25981)
471- Fixed LSP `findReferences` and other location-based operations returning results from gitignored files (e.g., `node_modules/`, `venv/`) (anthropics/claude-code#26051)
472- Moved config backup files from home directory root to `~/.claude/backups/` to reduce home directory clutter (anthropics/claude-code#26130)
473- Fixed sessions with large first prompts (>16KB) disappearing from the /resume list (anthropics/claude-code#26140)
474- Fixed shell functions with double-underscore prefixes (e.g., `__git_ps1`) not being preserved across shell sessions (anthropics/claude-code#25824)
475- Fixed spinner showing "0 tokens" counter before any tokens have been received (anthropics/claude-code#26105)
476- VSCode: Fixed conversation messages appearing dimmed while the AskUserQuestion dialog is open (anthropics/claude-code#26078)
477- 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)
478- Fixed Right Alt key leaving visible `[25~` escape sequence residue in the input field on Windows/Git Bash terminals (anthropics/claude-code#25943)
479- The `/rename` command now updates the terminal tab title by default (anthropics/claude-code#25789)
480- 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)
481- Fixed OSC 8 hyperlinks only being clickable on the first line when link text wraps across multiple terminal lines.
482
483## 2.1.46
484
485- Fixed orphaned CC processes after terminal disconnect on macOS
486- Added support for using claude.ai MCP connectors in Claude Code
487
488## 2.1.45
489
490- Added support for Claude Sonnet 4.6
491- Added support for reading `enabledPlugins` and `extraKnownMarketplaces` from `--add-dir` directories
492- 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
493- Added `SDKRateLimitInfo` and `SDKRateLimitEvent` types to the SDK, enabling consumers to receive rate limit status updates including utilization, reset times, and overage information
494- Fixed Agent Teams teammates failing on Bedrock, Vertex, and Foundry by propagating API provider environment variables to tmux-spawned processes (anthropics/claude-code#23561)
495- Fixed sandbox "operation not permitted" errors when writing temporary files on macOS by using the correct per-user temp directory (anthropics/claude-code#21654)
496- Fixed Task tool (backgrounded agents) crashing with a `ReferenceError` on completion (anthropics/claude-code#22087)
497- Fixed autocomplete suggestions not being accepted on Enter when images are pasted in the input
498- Fixed skills invoked by subagents incorrectly appearing in main session context after compaction
499- Fixed excessive `.claude.json.backup` files accumulating on every startup
500- Fixed plugin-provided commands, agents, and hooks not being available immediately after installation without requiring a restart
501- Improved startup performance by removing eager loading of session history for stats caching
502- Improved memory usage for shell commands that produce large output — RSS no longer grows unboundedly with command output size
503- Improved collapsed read/search groups to show the current file or search pattern being processed beneath the summary line while active
504- [VSCode] Improved permission destination choice (project/user/session) to persist across sessions
505
506## 2.1.44
507
508- Fixed ENAMETOOLONG errors for deeply-nested directory paths
509- Fixed auth refresh errors
510
511## 2.1.43
512
513- Fixed AWS auth refresh hanging indefinitely by adding a 3-minute timeout
514- Fixed spurious warnings for non-agent markdown files in `.claude/agents/` directory
515- Fixed structured-outputs beta header being sent unconditionally on Vertex/Bedrock
516
517## 2.1.42
518
519- Improved startup performance by deferring Zod schema construction
520- Improved prompt cache hit rates by moving date out of system prompt
521- Added one-time Opus 4.6 effort callout for eligible users
522- Fixed /resume showing interrupt messages as session titles
523- Fixed image dimension limit errors to suggest /compact
524
525## 2.1.41
526
527- Added guard against launching Claude Code inside another Claude Code session
528- Fixed Agent Teams using wrong model identifier for Bedrock, Vertex, and Foundry customers
529- Fixed a crash when MCP tools return image content during streaming
530- Fixed /resume session previews showing raw XML tags instead of readable command names
531- Improved model error messages for Bedrock/Vertex/Foundry users with fallback suggestions
532- Fixed plugin browse showing misleading "Space to Toggle" hint for already-installed plugins
533- Fixed hook blocking errors (exit code 2) not showing stderr to the user
534- Added `speed` attribute to OTel events and trace spans for fast mode visibility
535- Added `claude auth login`, `claude auth status`, and `claude auth logout` CLI subcommands
536- Added Windows ARM64 (win32-arm64) native binary support
537- Improved `/rename` to auto-generate session name from conversation context when called without arguments
538- Improved narrow terminal layout for prompt footer
539- Fixed file resolution failing for @-mentions with anchor fragments (e.g., `@README.md#installation`)
540- Fixed FileReadTool blocking the process on FIFOs, `/dev/stdin`, and large files
541- Fixed background task notifications not being delivered in streaming Agent SDK mode
542- Fixed cursor jumping to end on each keystroke in classifier rule input
543- Fixed markdown link display text being dropped for raw URL
544- Fixed auto-compact failure error notifications being shown to users
545- Fixed permission wait time being included in subagent elapsed time display
546- Fixed proactive ticks firing while in plan mode
547- Fixed clear stale permission rules when settings change on disk
548- Fixed hook blocking errors showing stderr content in UI
549
550## 2.1.39
551
552- Improved terminal rendering performance
553- Fixed fatal errors being swallowed instead of displayed
554- Fixed process hanging after session close
555- Fixed character loss at terminal screen boundary
556- Fixed blank lines in verbose transcript view
557
558## 2.1.38
559
560- Fixed VS Code terminal scroll-to-top regression introduced in 2.1.37
561- Fixed Tab key queueing slash commands instead of autocompleting
562- Fixed bash permission matching for commands using environment variable wrappers
563- Fixed text between tool uses disappearing when not using streaming
564- Fixed duplicate sessions when resuming in VS Code extension
565- Improved heredoc delimiter parsing to prevent command smuggling
566- Blocked writes to `.claude/skills` directory in sandbox mode
567
568## 2.1.37
569
570- Fixed an issue where /fast was not immediately available after enabling /extra-usage
571
572## 2.1.36
573
574- Fast mode is now available for Opus 4.6. Learn more at https://code.claude.com/docs/en/fast-mode
575
576## 2.1.34
577
578- Fixed a crash when agent teams setting changed between renders
579- Fixed a bug where commands excluded from sandboxing (via `sandbox.excludedCommands` or `dangerouslyDisableSandbox`) could bypass the Bash ask permission rule when `autoAllowBashIfSandboxed` was enabled
580
581## 2.1.33
582
583- Fixed agent teammate sessions in tmux to send and receive messages
584- Fixed warnings about agent teams not being available on your current plan
585- Added `TeammateIdle` and `TaskCompleted` hook events for multi-agent workflows
586- Added support for restricting which sub-agents can be spawned via `Task(agent_type)` syntax in agent "tools" frontmatter
587- Added `memory` frontmatter field support for agents, enabling persistent memory with `user`, `project`, or `local` scope
588- Added plugin name to skill descriptions and `/skills` menu for better discoverability
589- Fixed an issue where submitting a new message while the model was in extended thinking would interrupt the thinking phase
590- 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
591- Fixed API proxy compatibility issue where 404 errors on streaming endpoints no longer triggered non-streaming fallback
592- 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
593- Fixed `/resume` session picker showing raw XML markup instead of clean titles for sessions started with slash commands
594- Improved error messages for API connection failures — now shows specific cause (e.g., ECONNREFUSED, SSL errors) instead of generic "Connection error"
595- Errors from invalid managed settings are now surfaced
596- VSCode: Added support for remote sessions, allowing OAuth users to browse and resume sessions from claude.ai
597- VSCode: Added git branch and message count to the session picker, with support for searching by branch name
598- VSCode: Fixed scroll-to-bottom under-scrolling on initial session load and session switch
599
600## 2.1.32
601
602- Claude Opus 4.6 is now available!
603- Added research preview agent teams feature for multi-agent collaboration (token-intensive feature, requires setting CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1)
604- Claude now automatically records and recalls memories as it works
605- Added "Summarize from here" to the message selector, allowing partial conversation summarization.
606- Skills defined in `.claude/skills/` within additional directories (`--add-dir`) are now loaded automatically.
607- Fixed `@` file completion showing incorrect relative paths when running from a subdirectory
608- Updated --resume to re-use --agent value specified in previous conversation by default.
609- Fixed: Bash tool no longer throws "Bad substitution" errors when heredocs contain JavaScript template literals like `${index + 1}`, which previously interrupted tool execution
610- Skill character budget now scales with context window (2% of context), so users with larger context windows can see more skill descriptions without truncation
611- Fixed Thai/Lao spacing vowels (สระ า, ำ) not rendering correctly in the input field
612- VSCode: Fixed slash commands incorrectly being executed when pressing Enter with preceding text in the input field
613- VSCode: Added spinner when loading past conversations list
614
615## 2.1.31
616
617- Added session resume hint on exit, showing how to continue your conversation later
618- Added support for full-width (zenkaku) space input from Japanese IME in checkbox selection
619- Fixed PDF too large errors permanently locking up sessions, requiring users to start a new conversation
620- Fixed bash commands incorrectly reporting failure with "Read-only file system" errors when sandbox mode was enabled
621- Fixed a crash that made sessions unusable after entering plan mode when project config in `~/.claude.json` was missing default fields
622- 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
623- Fixed LSP shutdown/exit compatibility with strict language servers that reject null params
624- 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
625- Improved PDF and request size error messages to show actual limits (100 pages, 20MB)
626- Reduced layout jitter in the terminal when the spinner appears and disappears during streaming
627- Removed misleading Anthropic API pricing from model selector for third-party provider (Bedrock, Vertex, Foundry) users
628
629## 2.1.30
630
631- 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.
632- 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`.
633- Added `/debug` for Claude to help troubleshoot the current session
634- Added support for additional `git log` and `git show` flags in read-only mode (e.g., `--topo-order`, `--cherry-pick`, `--format`, `--raw`)
635- Added token count, tool uses, and duration metrics to Task tool results
636- Added reduced motion mode to the config
637- Fixed phantom "(no content)" text blocks appearing in API conversation history, reducing token waste and potential model confusion
638- Fixed prompt cache not correctly invalidating when tool descriptions or input schemas changed, only when tool names changed
639- Fixed 400 errors that could occur after running `/login` when the conversation contained thinking blocks
640- Fixed a hang when resuming sessions with corrupted transcript files containing `parentUuid` cycles
641- Fixed rate limit message showing incorrect "/upgrade" suggestion for Max 20x users when extra-usage is unavailable
642- Fixed permission dialogs stealing focus while actively typing
643- Fixed subagents not being able to access SDK-provided MCP tools because they were not synced to the shared application state
644- Fixed a regression where Windows users with a `.bashrc` file could not run bash commands
645- 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
646- Improved `TaskStop` tool to display the stopped command/task description in the result line instead of a generic "Task stopped" message
647- Changed `/model` to execute immediately instead of being queued
648- [VSCode] Added multiline input support to the "Other" text input in question dialogs (use Shift+Enter for new lines)
649- [VSCode] Fixed duplicate sessions appearing in the session list when starting a new conversation
650
651## 2.1.29
652
653- Fixed startup performance issues when resuming sessions that have `saved_hook_context`
654
655## 2.1.27
656
657- Added tool call failures and denials to debug logs
658- Fixed context management validation error for gateway users, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
659- Added `--from-pr` flag to resume sessions linked to a specific GitHub PR number or URL
660- Sessions are now automatically linked to PRs when created via `gh pr create`
661- Fixed /context command not displaying colored output
662- Fixed status bar duplicating background task indicator when PR status was shown
663- Windows: Fixed bash command execution failing for users with `.bashrc` files
664- Windows: Fixed console windows flashing when spawning child processes
665- VSCode: Fixed OAuth token expiration causing 401 errors after extended sessions
666
667## 2.1.25
668
669- Fixed beta header validation error for gateway users on Bedrock and Vertex, ensuring `CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1` avoids the error
670
671## 2.1.23
672
673- Added customizable spinner verbs setting (`spinnerVerbs`)
674- Fixed mTLS and proxy connectivity for users behind corporate proxies or using client certificates
675- Fixed per-user temp directory isolation to prevent permission conflicts on shared systems
676- Fixed a race condition that could cause 400 errors when prompt caching scope was enabled
677- Fixed pending async hooks not being cancelled when headless streaming sessions ended
678- Fixed tab completion not updating the input field when accepting a suggestion
679- Fixed ripgrep search timeouts silently returning empty results instead of reporting errors
680- Improved terminal rendering performance with optimized screen data layout
681- Changed Bash commands to show timeout duration alongside elapsed time
682- Changed merged pull requests to show a purple status indicator in the prompt footer
683- [IDE] Fixed model options displaying incorrect region strings for Bedrock users in headless mode
684
685## 2.1.22
686
687- Fixed structured outputs for non-interactive (-p) mode
688
689## 2.1.21
690
691- Added support for full-width (zenkaku) number input from Japanese IME in option selection prompts
692- Fixed shell completion cache files being truncated on exit
693- Fixed API errors when resuming sessions that were interrupted during tool execution
694- Fixed auto-compact triggering too early on models with large output token limits
695- Fixed task IDs potentially being reused after deletion
696- Fixed file search not working in VS Code extension on Windows
697- Improved read/search progress indicators to show "Reading…" while in progress and "Read" when complete
698- Improved Claude to prefer file operation tools (Read, Edit, Write) over bash equivalents (cat, sed, awk)
699- [VSCode] Added automatic Python virtual environment activation, ensuring `python` and `pip` commands use the correct interpreter (configurable via `claudeCode.usePythonEnvironment` setting)
700- [VSCode] Fixed message action buttons having incorrect background colors
701
702## 2.1.20
703
704- Added arrow key history navigation in vim normal mode when cursor cannot move further
705- Added external editor shortcut (Ctrl+G) to the help menu for better discoverability
706- 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
707- Added support for loading `CLAUDE.md` files from additional directories specified via `--add-dir` flag (requires setting `CLAUDE_CODE_ADDITIONAL_DIRECTORIES_CLAUDE_MD=1`)
708- Added ability to delete tasks via the `TaskUpdate` tool
709- Fixed session compaction issues that could cause resume to load full history instead of the compact summary
710- Fixed agents sometimes ignoring user messages sent while actively working on a task
711- Fixed wide character (emoji, CJK) rendering artifacts where trailing columns were not cleared when replaced by narrower characters
712- Fixed JSON parsing errors when MCP tool responses contain special Unicode characters
713- Fixed up/down arrow keys in multi-line and wrapped text input to prioritize cursor movement over history navigation
714- Fixed draft prompt being lost when pressing UP arrow to navigate command history
715- Fixed ghost text flickering when typing slash commands mid-input
716- Fixed marketplace source removal not properly deleting settings
717- Fixed duplicate output in some commands like `/context`
718- Fixed task list sometimes showing outside the main conversation view
719- Fixed syntax highlighting for diffs occurring within multiline constructs like Python docstrings
720- Fixed crashes when cancelling tool use
721- Improved `/sandbox` command UI to show dependency status with installation instructions when dependencies are missing
722- Improved thinking status text with a subtle shimmer animation
723- Improved task list to dynamically adjust visible items based on terminal height
724- Improved fork conversation hint to show how to resume the original session
725- Changed collapsed read/search groups to show present tense ("Reading", "Searching for") while in progress, and past tense ("Read", "Searched for") when complete
726- Changed `ToolSearch` results to appear as a brief notification instead of inline in the conversation
727- Changed the `/commit-push-pr` skill to automatically post PR URLs to Slack channels when configured via MCP tools
728- Changed the `/copy` command to be available to all users
729- Changed background agents to prompt for tool permissions before launching
730- Changed permission rules like `Bash(*)` to be accepted and treated as equivalent to `Bash`
731- Changed config backups to be timestamped and rotated (keeping 5 most recent) to prevent data loss
732
733## 2.1.19
734
735- Added env var `CLAUDE_CODE_ENABLE_TASKS`, set to `false` to keep the old system temporarily
736- Added shorthand `$0`, `$1`, etc. for accessing individual arguments in custom commands
737- Fixed crashes on processors without AVX instruction support
738- Fixed dangling Claude Code processes when terminal is closed by catching EIO errors from `process.exit()` and using SIGKILL as fallback
739- Fixed `/rename` and `/tag` not updating the correct session when resuming from a different directory (e.g., git worktrees)
740- Fixed resuming sessions by custom title not working when run from a different directory
741- Fixed pasted text content being lost when using prompt stash (Ctrl+S) and restore
742- Fixed agent list displaying "Sonnet (default)" instead of "Inherit (default)" for agents without an explicit model setting
743- Fixed backgrounded hook commands not returning early, potentially causing the session to wait on a process that was intentionally backgrounded
744- Fixed file write preview omitting empty lines
745- Changed skills without additional permissions or hooks to be allowed without requiring approval
746- Changed indexed argument syntax from `$ARGUMENTS.0` to `$ARGUMENTS[0]` (bracket syntax)
747- [SDK] Added replay of `queued_command` attachment messages as `SDKUserMessageReplay` events when `replayUserMessages` is enabled
748- [VSCode] Enabled session forking and rewind functionality for all users
749
750## 2.1.18
751
752- 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
753
754## 2.1.17
755
756- Fixed crashes on processors without AVX instruction support
757
758## 2.1.16
759
760- Added new task management system, including new capabilities like dependency tracking
761- [VSCode] Added native plugin management support
762- [VSCode] Added ability for OAuth users to browse and resume remote Claude sessions from the Sessions dialog
763- Fixed out-of-memory crashes when resuming sessions with heavy subagent usage
764- Fixed an issue where the "context remaining" warning was not hidden after running `/compact`
765- Fixed session titles on the resume screen not respecting the user's language setting
766- [IDE] Fixed a race condition on Windows where the Claude Code sidebar view container would not appear on start
767
768## 2.1.15
769
770- Added deprecation notification for npm installations - run `claude install` or see https://docs.anthropic.com/en/docs/claude-code/getting-started for more options
771- Improved UI rendering performance with React Compiler
772- Fixed the "Context left until auto-compact" warning not disappearing after running `/compact`
773- Fixed MCP stdio server timeout not killing child process, which could cause UI freezes
774
775## 2.1.14
776
777- Added history-based autocomplete in bash mode (`!`) - type a partial command and press Tab to complete from your bash command history
778- Added search to installed plugins list - type to filter by name or description
779- Added support for pinning plugins to specific git commit SHAs, allowing marketplace entries to install exact versions
780- Fixed a regression where the context window blocking limit was calculated too aggressively, blocking users at ~65% context usage instead of the intended ~98%
781- Fixed memory issues that could cause crashes when running parallel subagents
782- Fixed memory leak in long-running sessions where stream resources were not cleaned up after shell commands completed
783- Fixed `@` symbol incorrectly triggering file autocomplete suggestions in bash mode
784- Fixed `@`-mention menu folder click behavior to navigate into directories instead of selecting them
785- Fixed `/feedback` command generating invalid GitHub issue URLs when description is very long
786- Fixed `/context` command to show the same token count and percentage as the status line in verbose mode
787- Fixed an issue where `/config`, `/context`, `/model`, and `/todos` command overlays could close unexpectedly
788- Fixed slash command autocomplete selecting wrong command when typing similar commands (e.g., `/context` vs `/compact`)
789- Fixed inconsistent back navigation in plugin marketplace when only one marketplace is configured
790- Fixed iTerm2 progress bar not clearing properly on exit, preventing lingering indicators and bell sounds
791- Improved backspace to delete pasted text as a single token instead of one character at a time
792- [VSCode] Added `/usage` command to display current plan usage
793
794## 2.1.12
795
796- Fixed message rendering bug
797
798## 2.1.11
799
800- Fixed excessive MCP connection requests for HTTP/SSE transports
801
802## 2.1.10
803
804- Added new `Setup` hook event that can be triggered via `--init`, `--init-only`, or `--maintenance` CLI flags for repository setup and maintenance operations
805- Added keyboard shortcut 'c' to copy OAuth URL when browser doesn't open automatically during login
806- Fixed a crash when running bash commands containing heredocs with JavaScript template literals like `${index + 1}`
807- Improved startup to capture keystrokes typed before the REPL is fully ready
808- Improved file suggestions to show as removable attachments instead of inserting text when accepted
809- [VSCode] Added install count display to plugin listings
810- [VSCode] Added trust warning when installing plugins
811
812## 2.1.9
813
814- Added `auto:N` syntax for configuring the MCP tool search auto-enable threshold, where N is the context window percentage (0-100)
815- Added `plansDirectory` setting to customize where plan files are stored
816- Added external editor support (Ctrl+G) in AskUserQuestion "Other" input field
817- Added session URL attribution to commits and PRs created from web sessions
818- Added support for `PreToolUse` hooks to return `additionalContext` to the model
819- Added `${CLAUDE_SESSION_ID}` string substitution for skills to access the current session ID
820- Fixed long sessions with parallel tool calls failing with an API error about orphan tool_result blocks
821- Fixed MCP server reconnection hanging when cached connection promise never resolves
822- Fixed Ctrl+Z suspend not working in terminals using Kitty keyboard protocol (Ghostty, iTerm2, kitty, WezTerm)
823
824## 2.1.7
825
826- Added `showTurnDuration` setting to hide turn duration messages (e.g., "Cooked for 1m 6s")
827- Added ability to provide feedback when accepting permission prompts
828- Added inline display of agent's final response in task notifications, making it easier to see results without reading the full transcript file
829- Fixed security vulnerability where wildcard permission rules could match compound commands containing shell operators
830- Fixed false "file modified" errors on Windows when cloud sync tools, antivirus scanners, or Git touch file timestamps without changing content
831- Fixed orphaned tool_result errors when sibling tools fail during streaming execution
832- 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)
833- Fixed spinner briefly flashing when running local slash commands like `/model` or `/theme`
834- Fixed terminal title animation jitter by using fixed-width braille characters
835- Fixed plugins with git submodules not being fully initialized when installed
836- Fixed bash commands failing on Windows when temp directory paths contained characters like `t` or `n` that were misinterpreted as escape sequences
837- Improved typing responsiveness by reducing memory allocation overhead in terminal rendering
838- 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.
839- Changed OAuth and API Console URLs from console.anthropic.com to platform.claude.com
840- [VSCode] Fixed `claudeProcessWrapper` setting passing the wrapper path instead of the Claude binary path
841
842## 2.1.6
843
844- Added search functionality to `/config` command for quickly filtering settings
845- Added Updates section to `/doctor` showing auto-update channel and available npm versions (stable/latest)
846- Added date range filtering to `/stats` command - press `r` to cycle between Last 7 days, Last 30 days, and All time
847- Added automatic discovery of skills from nested `.claude/skills` directories when working with files in subdirectories
848- Added `context_window.used_percentage` and `context_window.remaining_percentage` fields to status line input for easier context window display
849- Added an error display when the editor fails during Ctrl+G
850- Fixed permission bypass via shell line continuation that could allow blocked commands to execute
851- Fixed false "File has been unexpectedly modified" errors when file watchers touch files without changing content
852- Fixed text styling (bold, colors) getting progressively misaligned in multi-line responses
853- Fixed the feedback panel closing unexpectedly when typing 'n' in the description field
854- Fixed rate limit warning appearing at low usage after weekly reset (now requires 70% usage)
855- Fixed rate limit options menu incorrectly auto-opening when resuming a previous session
856- Fixed numpad keys outputting escape sequences instead of characters in Kitty keyboard protocol terminals
857- Fixed Option+Return not inserting newlines in Kitty keyboard protocol terminals
858- Fixed corrupted config backup files accumulating in the home directory (now only one backup is created per config file)
859- Fixed `mcp list` and `mcp get` commands leaving orphaned MCP server processes
860- Fixed visual artifacts in ink2 mode when nodes become hidden via `display:none`
861- Improved the external CLAUDE.md imports approval dialog to show which files are being imported and from where
862- Improved the `/tasks` dialog to go directly to task details when there's only one background task running
863- Improved @ autocomplete with icons for different suggestion types and single-line formatting
864- Updated "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
865- Changed task notification display to cap at 3 lines with overflow summary when multiple background tasks complete simultaneously
866- Changed terminal title to "Claude Code" on startup for better window identification
867- Removed ability to @-mention MCP servers to enable/disable - use `/mcp enable <name>` instead
868- [VSCode] Fixed usage indicator not updating after manual compact
869
870## 2.1.5
871
872- Added `CLAUDE_CODE_TMPDIR` environment variable to override the temp directory used for internal temp files, useful for environments with custom temp directory requirements
873
874## 2.1.4
875
876- Added `CLAUDE_CODE_DISABLE_BACKGROUND_TASKS` environment variable to disable all background task functionality including auto-backgrounding and the Ctrl+B shortcut
877- Fixed "Help improve Claude" setting fetch to refresh OAuth and retry when it fails due to a stale OAuth token
878
879## 2.1.3
880
881- Merged slash commands and skills, simplifying the mental model with no change in behavior
882- Added release channel (`stable` or `latest`) toggle to `/config`
883- 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
884- Fixed plan files persisting across `/clear` commands, now ensuring a fresh plan file is used after clearing a conversation
885- Fixed false skill duplicate detection on filesystems with large inodes (e.g., ExFAT) by using 64-bit precision for inode values
886- Fixed mismatch between background task count in status bar and items shown in tasks dialog
887- Fixed sub-agents using the wrong model during conversation compaction
888- Fixed web search in sub-agents using incorrect model
889- Fixed trust dialog acceptance when running from the home directory not enabling trust-requiring features like hooks during the session
890- Improved terminal rendering stability by preventing uncontrolled writes from corrupting cursor state
891- Improved slash command suggestion readability by truncating long descriptions to 2 lines
892- Changed tool hook execution timeout from 60 seconds to 10 minutes
893- [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)
894
895## 2.1.2
896
897- Added source path metadata to images dragged onto the terminal, helping Claude understand where images originated
898- Added clickable hyperlinks for file paths in tool output in terminals that support OSC 8 (like iTerm)
899- Added support for Windows Package Manager (winget) installations with automatic detection and update instructions
900- Added Shift+Tab keyboard shortcut in plan mode to quickly select "auto-accept edits" option
901- Added `FORCE_AUTOUPDATE_PLUGINS` environment variable to allow plugin autoupdate even when the main auto-updater is disabled
902- Added `agent_type` to SessionStart hook input, populated if `--agent` is specified
903- Fixed a command injection vulnerability in bash command processing where malformed input could execute arbitrary commands
904- Fixed a memory leak where tree-sitter parse trees were not being freed, causing WASM memory to grow unbounded over long sessions
905- Fixed binary files (images, PDFs, etc.) being accidentally included in memory when using `@include` directives in CLAUDE.md files
906- Fixed updates incorrectly claiming another installation is in progress
907- Fixed crash when socket files exist in watched directories (defense-in-depth for EOPNOTSUPP errors)
908- Fixed remote session URL and teleport being broken when using `/tasks` command
909- Fixed MCP tool names being exposed in analytics events by sanitizing user-specific server configurations
910- Improved Option-as-Meta hint on macOS to show terminal-specific instructions for native CSIu terminals like iTerm2, Kitty, and WezTerm
911- Improved error message when pasting images over SSH to suggest using `scp` instead of the unhelpful clipboard shortcut hint
912- Improved permission explainer to not flag routine dev workflows (git fetch/rebase, npm install, tests, PRs) as medium risk
913- Changed large bash command outputs to be saved to disk instead of truncated, allowing Claude to read the full content
914- Changed large tool outputs to be persisted to disk instead of truncated, providing full output access via file references
915- Changed `/plugins` installed tab to unify plugins and MCPs with scope-based grouping
916- Deprecated Windows managed settings path `C:\ProgramData\ClaudeCode\managed-settings.json` - administrators should migrate to `C:\Program Files\ClaudeCode\managed-settings.json`
917- [SDK] Changed minimum zod peer dependency to ^4.0.0
918- [VSCode] Fixed usage display not updating after manual compact
919
920## 2.1.0
921
922- Added automatic skill hot-reload - skills created or modified in `~/.claude/skills` or `.claude/skills` are now immediately available without restarting the session
923- Added support for running skills and slash commands in a forked sub-agent context using `context: fork` in skill frontmatter
924- Added support for `agent` field in skills to specify agent type for execution
925- Added `language` setting to configure Claude's response language (e.g., language: "japanese")
926- Changed Shift+Enter to work out of the box in iTerm2, WezTerm, Ghostty, and Kitty without modifying terminal configs
927- Added `respectGitignore` support in `settings.json` for per-project control over @-mention file picker behavior
928- Added `IS_DEMO` environment variable to hide email and organization from the UI, useful for streaming or recording sessions
929- Fixed security issue where sensitive data (OAuth tokens, API keys, passwords) could be exposed in debug logs
930- Fixed files and skills not being properly discovered when resuming sessions with `-c` or `--resume`
931- Fixed pasted content being lost when replaying prompts from history using up arrow or Ctrl+R search
932- Fixed Esc key with queued prompts to only move them to input without canceling the running task
933- Reduced permission prompts for complex bash commands
934- Fixed command search to prioritize exact and prefix matches on command names over fuzzy matches in descriptions
935- Fixed PreToolUse hooks to allow `updatedInput` when returning `ask` permission decision, enabling hooks to act as middleware while still requesting user consent
936- Fixed plugin path resolution for file-based marketplace sources
937- Fixed LSP tool being incorrectly enabled when no LSP servers were configured
938- Fixed background tasks failing with "git repository not found" error for repositories with dots in their names
939- Fixed Claude in Chrome support for WSL environments
940- Fixed Windows native installer silently failing when executable creation fails
941- Improved CLI help output to display options and subcommands in alphabetical order for easier navigation
942- Added wildcard pattern matching for Bash tool permissions using `*` at any position in rules (e.g., `Bash(npm *)`, `Bash(* install)`, `Bash(git * main)`)
943- Added unified Ctrl+B backgrounding for both bash commands and agents - pressing Ctrl+B now backgrounds all running foreground tasks simultaneously
944- Added support for MCP `list_changed` notifications, allowing MCP servers to dynamically update their available tools, prompts, and resources without requiring reconnection
945- Added `/teleport` and `/remote-env` slash commands for claude.ai subscribers, allowing them to resume and configure remote sessions
946- Added support for disabling specific agents using `Task(AgentName)` syntax in settings.json permissions or the `--disallowedTools` CLI flag
947- Added hooks support to agent frontmatter, allowing agents to define PreToolUse, PostToolUse, and Stop hooks scoped to the agent's lifecycle
948- Added hooks support for skill and slash command frontmatter
949- 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
950- Added `/plan` command shortcut to enable plan mode directly from the prompt
951- Added slash command autocomplete support when `/` appears anywhere in input, not just at the beginning
952- Added `--tools` flag support in interactive mode to restrict which built-in tools Claude can use during interactive sessions
953- Added `CLAUDE_CODE_FILE_READ_MAX_OUTPUT_TOKENS` environment variable to override the default file read token limit
954- Added support for `once: true` config for hooks
955- Added support for YAML-style lists in frontmatter `allowed-tools` field for cleaner skill declarations
956- Added support for prompt and agent hook types from plugins (previously only command hooks were supported)
957- Added Cmd+V support for image paste in iTerm2 (maps to Ctrl+V)
958- Added left/right arrow key navigation for cycling through tabs in dialogs
959- Added real-time thinking block display in Ctrl+O transcript mode
960- Added filepath to full output in background bash task details dialog
961- Added Skills as a separate category in the context visualization
962- Fixed OAuth token refresh not triggering when server reports token expired but local expiration check disagrees
963- Fixed session persistence getting stuck after transient server errors by recovering from 409 conflicts when the entry was actually stored
964- Fixed session resume failures caused by orphaned tool results during concurrent tool execution
965- Fixed a race condition where stale OAuth tokens could be read from the keychain cache during concurrent token refresh attempts
966- Fixed AWS Bedrock subagents not inheriting EU/APAC cross-region inference model configuration, causing 403 errors when IAM permissions are scoped to specific regions
967- Fixed API context overflow when background tasks produce large output by truncating to 30K chars with file path reference
968- Fixed a hang when reading FIFO files by skipping symlink resolution for special file types
969- Fixed terminal keyboard mode not being reset on exit in Ghostty, iTerm2, Kitty, and WezTerm
970- Fixed Alt+B and Alt+F (word navigation) not working in iTerm2, Ghostty, Kitty, and WezTerm
971- Fixed `${CLAUDE_PLUGIN_ROOT}` not being substituted in plugin `allowed-tools` frontmatter, which caused tools to incorrectly require approval
972- Fixed files created by the Write tool using hardcoded 0o600 permissions instead of respecting the system umask
973- Fixed commands with `$()` command substitution failing with parse errors
974- Fixed multi-line bash commands with backslash continuations being incorrectly split and flagged for permissions
975- 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)
976- Fixed slash commands passed as CLI arguments (e.g., `claude /context`) not being executed properly
977- Fixed pressing Enter after Tab-completing a slash command selecting a different command instead of submitting the completed one
978- Fixed slash command argument hint flickering and inconsistent display when typing commands with arguments
979- Fixed Claude sometimes redundantly invoking the Skill tool when running slash commands directly
980- Fixed skill token estimates in `/context` to accurately reflect frontmatter-only loading
981- Fixed subagents sometimes not inheriting the parent's model by default
982- Fixed model picker showing incorrect selection for Bedrock/Vertex users using `--model haiku`
983- Fixed duplicate Bash commands appearing in permission request option labels
984- Fixed noisy output when background tasks complete - now shows clean completion message instead of raw output
985- Fixed background task completion notifications to appear proactively with bullet point
986- Fixed forked slash commands showing "AbortError" instead of "Interrupted" message when cancelled
987- Fixed cursor disappearing after dismissing permission dialogs
988- Fixed `/hooks` menu selecting wrong hook type when scrolling to a different option
989- Fixed images in queued prompts showing as "[object Object]" when pressing Esc to cancel
990- Fixed images being silently dropped when queueing messages while backgrounding a task
991- Fixed large pasted images failing with "Image was too large" error
992- Fixed extra blank lines in multiline prompts containing CJK characters (Japanese, Chinese, Korean)
993- Fixed ultrathink keyword highlighting being applied to wrong characters when user prompt text wraps to multiple lines
994- Fixed collapsed "Reading X files…" indicator incorrectly switching to past tense when thinking blocks appear mid-stream
995- Fixed Bash read commands (like `ls` and `cat`) not being counted in collapsed read/search groups, causing groups to incorrectly show "Read 0 files"
996- Fixed spinner token counter to properly accumulate tokens from subagents during execution
997- Fixed memory leak in git diff parsing where sliced strings retained large parent strings
998- Fixed race condition where LSP tool could return "no server available" during startup
999- Fixed feedback submission hanging indefinitely when network requests timeout
1000- Fixed search mode in plugin discovery and log selector views exiting when pressing up arrow
1001- Fixed hook success message showing trailing colon when hook has no output
1002- Multiple optimizations to improve startup performance
1003- Improved terminal rendering performance when using native installer or Bun, especially for text with emoji, ANSI codes, and Unicode characters
1004- Improved performance when reading Jupyter notebooks with many cells
1005- Improved reliability for piped input like `cat refactor.md | claude`
1006- Improved reliability for AskQuestion tool
1007- Improved sed in-place edit commands to render as file edits with diff preview
1008- Improved Claude to automatically continue when response is cut off due to output token limit, instead of showing an error message
1009- Improved compaction reliability
1010- Improved subagents (Task tool) to continue working after permission denial, allowing them to try alternative approaches
1011- Improved skills to show progress while executing, displaying tool uses as they happen
1012- Improved skills from `/skills/` directories to be visible in the slash command menu by default (opt-out with `user-invocable: false` in frontmatter)
1013- Improved skill suggestions to prioritize recently and frequently used skills
1014- Improved spinner feedback when waiting for the first response token
1015- Improved token count display in spinner to include tokens from background agents
1016- Improved incremental output for async agents to give the main thread more control and visibility
1017- Improved permission prompt UX with Tab hint moved to footer, cleaner Yes/No input labels with contextual placeholders
1018- Improved the Claude in Chrome notification with shortened help text and persistent display until dismissed
1019- Improved macOS screenshot paste reliability with TIFF format support
1020- Improved `/stats` output
1021- Updated Atlassian MCP integration to use a more reliable default configuration (streamable HTTP)
1022- Changed "Interrupted" message color from red to grey for a less alarming appearance
1023- Removed permission prompt when entering plan mode - users can now enter plan mode without approval
1024- Removed underline styling from image reference links
1025- [SDK] Changed minimum zod peer dependency to ^4.0.0
1026- [VSCode] Added currently selected model name to the context menu
1027- [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")
1028- [VSCode] Fixed paragraph breaks not rendering in markdown content
1029- [VSCode] Fixed scrolling in the extension inadvertently scrolling the parent iframe
1030- [Windows] Fixed issue with improper rendering
1031
1032## 2.0.76
1033
1034- Fixed issue with macOS code-sign warning when using Claude in Chrome integration
1035
1036## 2.0.75
1037
1038- Minor bugfixes
1039
1040## 2.0.74
1041
1042- Added LSP (Language Server Protocol) tool for code intelligence features like go-to-definition, find references, and hover documentation
1043- Added `/terminal-setup` support for Kitty, Alacritty, Zed, and Warp terminals
1044- Added ctrl+t shortcut in `/theme` to toggle syntax highlighting on/off
1045- Added syntax highlighting info to theme picker
1046- Added guidance for macOS users when Alt shortcuts fail due to terminal configuration
1047- Fixed skill `allowed-tools` not being applied to tools invoked by the skill
1048- Fixed Opus 4.5 tip incorrectly showing when user was already using Opus
1049- Fixed a potential crash when syntax highlighting isn't initialized correctly
1050- Fixed visual bug in `/plugins discover` where list selection indicator showed while search box was focused
1051- Fixed macOS keyboard shortcuts to display 'opt' instead of 'alt'
1052- Improved `/context` command visualization with grouped skills and agents by source, slash commands, and sorted token count
1053- [Windows] Fixed issue with improper rendering
1054- [VSCode] Added gift tag pictogram for year-end promotion message
1055
1056## 2.0.73
1057
1058- Added clickable `[Image #N]` links that open attached images in the default viewer
1059- Added alt-y yank-pop to cycle through kill ring history after ctrl-y yank
1060- Added search filtering to the plugin discover screen (type to filter by name, description, or marketplace)
1061- Added support for custom session IDs when forking sessions with `--session-id` combined with `--resume` or `--continue` and `--fork-session`
1062- Fixed slow input history cycling and race condition that could overwrite text after message submission
1063- Improved `/theme` command to open theme picker directly
1064- Improved theme picker UI
1065- Improved search UX across resume session, permissions, and plugins screens with a unified SearchBox component
1066- [VSCode] Added tab icon badges showing pending permissions (blue) and unread completions (orange)
1067
1068## 2.0.72
1069
1070- 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
1071- Reduced terminal flickering
1072- Added scannable QR code to mobile app tip for quick app downloads
1073- Added loading indicator when resuming conversations for better feedback
1074- Fixed `/context` command not respecting custom system prompts in non-interactive mode
1075- Fixed order of consecutive Ctrl+K lines when pasting with Ctrl+Y
1076- Improved @ mention file suggestion speed (~3× faster in git repositories)
1077- Improved file suggestion performance in repos with `.ignore` or `.rgignore` files
1078- Improved settings validation errors to be more prominent
1079- Changed thinking toggle from Tab to Alt+T to avoid accidental triggers
1080
1081## 2.0.71
1082
1083- Added /config toggle to enable/disable prompt suggestions
1084- Added `/settings` as an alias for the `/config` command
1085- Fixed @ file reference suggestions incorrectly triggering when cursor is in the middle of a path
1086- Fixed MCP servers from `.mcp.json` not loading when using `--dangerously-skip-permissions`
1087- Fixed permission rules incorrectly rejecting valid bash commands containing shell glob patterns (e.g., `ls *.txt`, `for f in *.png`)
1088- Bedrock: Environment variable `ANTHROPIC_BEDROCK_BASE_URL` is now respected for token counting and inference profile listing
1089- New syntax highlighting engine for native build
1090
1091## 2.0.70
1092
1093- Added Enter key to accept and submit prompt suggestions immediately (tab still accepts for editing)
1094- Added wildcard syntax `mcp__server__*` for MCP tool permissions to allow or deny all tools from a server
1095- Added auto-update toggle for plugin marketplaces, allowing per-marketplace control over automatic updates
1096- Added `current_usage` field to status line input, enabling accurate context window percentage calculations
1097- Fixed input being cleared when processing queued commands while the user was typing
1098- Fixed prompt suggestions replacing typed input when pressing Tab
1099- Fixed diff view not updating when terminal is resized
1100- Improved memory usage by 3x for large conversations
1101- Improved resolution of stats screenshots copied to clipboard (Ctrl+S) for crisper images
1102- Removed # shortcut for quick memory entry (tell Claude to edit your CLAUDE.md instead)
1103- Fix thinking mode toggle in /config not persisting correctly
1104- Improve UI for file creation permission dialog
1105
1106## 2.0.69
1107
1108- Minor bugfixes
1109
1110## 2.0.68
1111
1112- Fixed IME (Input Method Editor) support for languages like Chinese, Japanese, and Korean by correctly positioning the composition window at the cursor
1113- Fixed a bug where disallowed MCP tools were visible to the model
1114- Fixed an issue where steering messages could be lost while a subagent is working
1115- Fixed Option+Arrow word navigation treating entire CJK (Chinese, Japanese, Korean) text sequences as a single word instead of navigating by word boundaries
1116- Improved plan mode exit UX: show simplified yes/no dialog when exiting with empty or missing plan instead of throwing an error
1117- Add support for enterprise managed settings. Contact your Anthropic account team to enable this feature.
1118
1119## 2.0.67
1120
1121- Thinking mode is now enabled by default for Opus 4.5
1122- Thinking mode configuration has moved to /config
1123- Added search functionality to `/permissions` command with `/` keyboard shortcut for filtering rules by tool name
1124- Show reason why autoupdater is disabled in `/doctor`
1125- Fixed false "Another process is currently updating Claude" error when running `claude update` while another instance is already on the latest version
1126- Fixed MCP servers from `.mcp.json` being stuck in pending state when running in non-interactive mode (`-p` flag or piped input)
1127- Fixed scroll position resetting after deleting a permission rule in `/permissions`
1128- 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
1129- Fixed `claude install --force` not bypassing stale lock files
1130- Fixed consecutive @~/ file references in CLAUDE.md being incorrectly parsed due to markdown strikethrough interference
1131- Windows: Fixed plugin MCP servers failing due to colons in log directory paths
1132
1133## 2.0.65
1134
1135- Added ability to switch models while writing a prompt using alt+p (linux, windows), option+p (macos).
1136- Added context window information to status line input
1137- Added `fileSuggestion` setting for custom `@` file search commands
1138- Added `CLAUDE_CODE_SHELL` environment variable to override automatic shell detection (useful when login shell differs from actual working shell)
1139- Fixed prompt not being saved to history when aborting a query with Escape
1140- Fixed Read tool image handling to identify format from bytes instead of file extension
1141
1142## 2.0.64
1143
1144- Made auto-compacting instant
1145- Agents and bash commands can run asynchronously and send messages to wake up the main agent
1146- /stats now provides users with interesting CC stats, such as favorite model, usage graph, usage streak
1147- Added named session support: use `/rename` to name sessions, `/resume <name>` in REPL or `claude --resume <name>` from the terminal to resume them
1148- Added support for .claude/rules/`. See https://code.claude.com/docs/en/memory for details.
1149- Added image dimension metadata when images are resized, enabling accurate coordinate mappings for large images
1150- Fixed auto-loading .env when using native installer
1151- Fixed `--system-prompt` being ignored when using `--continue` or `--resume` flags
1152- Improved `/resume` screen with grouped forked sessions and keyboard shortcuts for preview (P) and rename (R)
1153- VSCode: Added copy-to-clipboard button on code blocks and bash tool inputs
1154- VSCode: Fixed extension not working on Windows ARM64 by falling back to x64 binary via emulation
1155- Bedrock: Improve efficiency of token counting
1156- Bedrock: Add support for `aws login` AWS Management Console credentials
1157- Unshipped AgentOutputTool and BashOutputTool, in favor of a new unified TaskOutputTool
1158
1159## 2.0.62
1160
1161- Added "(Recommended)" indicator for multiple-choice questions, with the recommended option moved to the top of the list
1162- Added `attribution` setting to customize commit and PR bylines (deprecates `includeCoAuthoredBy`)
1163- Fixed duplicate slash commands appearing when ~/.claude is symlinked to a project directory
1164- Fixed slash command selection not working when multiple commands share the same name
1165- Fixed an issue where skill files inside symlinked skill directories could become circular symlinks
1166- Fixed running versions getting removed because lock file incorrectly going stale
1167- Fixed IDE diff tab not closing when rejecting file changes
1168
1169## 2.0.61
1170
1171- Reverted VSCode support for multiple terminal clients due to responsiveness issues.
1172
1173## 2.0.60
1174
1175- Added background agent support. Agents run in the background while you work
1176- Added --disable-slash-commands CLI flag to disable all slash commands
1177- Added model name to "Co-Authored-By" commit messages
1178- Enabled "/mcp enable [server-name]" or "/mcp disable [server-name]" to quickly toggle all servers
1179- Updated Fetch to skip summarization for pre-approved websites
1180- VSCode: Added support for multiple terminal clients connecting to the IDE server simultaneously
1181
1182## 2.0.59
1183
1184- Added --agent CLI flag to override the agent setting for the current session
1185- Added `agent` setting to configure main thread with a specific agent's system prompt, tool restrictions, and model
1186- VS Code: Fixed .claude.json config file being read from incorrect location
1187
1188## 2.0.58
1189
1190- Pro users now have access to Opus 4.5 as part of their subscription!
1191- Fixed timer duration showing "11m 60s" instead of "12m 0s"
1192- Windows: Managed settings now prefer `C:\Program Files\ClaudeCode` if it exists. Support for `C:\ProgramData\ClaudeCode` will be removed in a future version.
1193
1194## 2.0.57
1195
1196- Added feedback input when rejecting plans, allowing users to tell Claude what to change
1197- VSCode: Added streaming message support for real-time response display
1198
1199## 2.0.56
1200
1201- Added setting to enable/disable terminal progress bar (OSC 9;4)
1202- 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.
1203
1204## 2.0.55
1205
1206- Fixed proxy DNS resolution being forced on by default. Now opt-in via `CLAUDE_CODE_PROXY_RESOLVES_HOSTS=true` environment variable
1207- Fixed keyboard navigation becoming unresponsive when holding down arrow keys in memory location selector
1208- Improved AskUserQuestion tool to auto-submit single-select questions on the last question, eliminating the extra review screen for simple question flows
1209- Improved fuzzy matching for `@` file suggestions with faster, more accurate results
1210
1211## 2.0.54
1212
1213- Hooks: Enable PermissionRequest hooks to process 'always allow' suggestions and apply permission updates
1214- Fix issue with excessive iTerm notifications
1215
1216## 2.0.52
1217
1218- Fixed duplicate message display when starting Claude with a command line argument
1219- Fixed `/usage` command progress bars to fill up as usage increases (instead of showing remaining percentage)
1220- Fixed image pasting not working on Linux systems running Wayland (now falls back to wl-paste when xclip is unavailable)
1221- Permit some uses of `$!` in bash commands
1222
1223## 2.0.51
1224
1225- Added Opus 4.5! https://www.anthropic.com/news/claude-opus-4-5
1226- Introducing Claude Code for Desktop: https://claude.com/download
1227- 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
1228- Pro users can now purchase extra usage for access to Opus 4.5 in Claude Code
1229- Plan Mode now builds more precise plans and executes more thoroughly
1230- Usage limit notifications now easier to understand
1231- Switched `/usage` back to "% used"
1232- Fixed handling of thinking errors
1233- Fixed performance regression
1234
1235## 2.0.50
1236
1237- Fixed bug preventing calling MCP tools that have nested references in their input schemas
1238- Silenced a noisy but harmless error during upgrades
1239- Improved ultrathink text display
1240- Improved clarity of 5-hour session limit warning message
1241
1242## 2.0.49
1243
1244- Added readline-style ctrl-y for pasting deleted text
1245- Improved clarity of usage limit warning message
1246- Fixed handling of subagent permissions
1247
1248## 2.0.47
1249
1250- Improved error messages and validation for `claude --teleport`
1251- Improved error handling in `/usage`
1252- Fixed race condition with history entry not getting logged at exit
1253- Fixed Vertex AI configuration not being applied from `settings.json`
1254
1255## 2.0.46
1256
1257- Fixed image files being reported with incorrect media type when format cannot be detected from metadata
1258
1259## 2.0.45
1260
1261- Added support for Microsoft Foundry! See https://code.claude.com/docs/en/azure-ai-foundry
1262- Added `PermissionRequest` hook to automatically approve or deny tool permission requests with custom logic
1263- Send background tasks to Claude Code on the web by starting a message with `&`
1264
1265## 2.0.43
1266
1267- Added `permissionMode` field for custom agents
1268- Added `tool_use_id` field to `PreToolUseHookInput` and `PostToolUseHookInput` types
1269- Added skills frontmatter field to declare skills to auto-load for subagents
1270- Added the `SubagentStart` hook event
1271- Fixed nested `CLAUDE.md` files not loading when @-mentioning files
1272- Fixed duplicate rendering of some messages in the UI
1273- Fixed some visual flickers
1274- Fixed NotebookEdit tool inserting cells at incorrect positions when cell IDs matched the pattern `cell-N`
1275
1276## 2.0.42
1277
1278- Added `agent_id` and `agent_transcript_path` fields to `SubagentStop` hooks.
1279
1280## 2.0.41
1281
1282- Added `model` parameter to prompt-based stop hooks, allowing users to specify a custom model for hook evaluation
1283- Fixed slash commands from user settings being loaded twice, which could cause rendering issues
1284- Fixed incorrect labeling of user settings vs project settings in command descriptions
1285- Fixed crash when plugin command hooks timeout during execution
1286- Fixed: Bedrock users no longer see duplicate Opus entries in the /model picker when using `--model haiku`
1287- Fixed broken security documentation links in trust dialogs and onboarding
1288- Fixed issue where pressing ESC to close the diff modal would also interrupt the model
1289- ctrl-r history search landing on a slash command no longer cancels the search
1290- SDK: Support custom timeouts for hooks
1291- Allow more safe git commands to run without approval
1292- Plugins: Added support for sharing and installing output styles
1293- Teleporting a session from web will automatically set the upstream branch
1294
1295## 2.0.37
1296
1297- Fixed how idleness is computed for notifications
1298- Hooks: Added matcher values for Notification hook events
1299- Output Styles: Added `keep-coding-instructions` option to frontmatter
1300
1301## 2.0.36
1302
1303- Fixed: DISABLE_AUTOUPDATER environment variable now properly disables package manager update notifications
1304- Fixed queued messages being incorrectly executed as bash commands
1305- Fixed input being lost when typing while a queued message is processed
1306
1307## 2.0.35
1308
1309- Improve fuzzy search results when searching commands
1310- Improved VS Code extension to respect `chat.fontSize` and `chat.fontFamily` settings throughout the entire UI, and apply font changes immediately without requiring reload
1311- 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
1312- Migrated `ignorePatterns` from project config to deny permissions in the localSettings.
1313- Fixed menu navigation getting stuck on items with empty string or other falsy values (e.g., in the `/hooks` menu)
1314
1315## 2.0.34
1316
1317- VSCode Extension: Added setting to configure the initial permission mode for new conversations
1318- Improved file path suggestion performance with native Rust-based fuzzy finder
1319- Fixed infinite token refresh loop that caused MCP servers with OAuth (e.g., Slack) to hang during connection
1320- Fixed memory crash when reading or writing large files (especially base64-encoded images)
1321
1322## 2.0.33
1323
1324- Native binary installs now launch quicker.
1325- Fixed `claude doctor` incorrectly detecting Homebrew vs npm-global installations by properly resolving symlinks
1326- Fixed `claude mcp serve` exposing tools with incompatible outputSchemas
1327
1328## 2.0.32
1329
1330- Un-deprecate output styles based on community feedback
1331- Added `companyAnnouncements` setting for displaying announcements on startup
1332- Fixed hook progress messages not updating correctly during PostToolUse hook execution
1333
1334## 2.0.31
1335
1336- Windows: native installation uses shift+tab as shortcut for mode switching, instead of alt+m
1337- Vertex: add support for Web Search on supported models
1338- VSCode: Adding the respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1339- Fixed a bug with subagents and MCP servers related to "Tool names must be unique" error
1340- Fixed issue causing `/compact` to fail with `prompt_too_long` by making it respect existing compact boundaries
1341- Fixed plugin uninstall not removing plugins
1342
1343## 2.0.30
1344
1345- Added helpful hint to run `security unlock-keychain` when encountering API key errors on macOS with locked keychain
1346- Added `allowUnsandboxedCommands` sandbox setting to disable the dangerouslyDisableSandbox escape hatch at policy level
1347- Added `disallowedTools` field to custom agent definitions for explicit tool blocking
1348- Added prompt-based stop hooks
1349- VSCode: Added respectGitIgnore configuration to include .gitignored files in file searches (defaults to true)
1350- Enabled SSE MCP servers on native build
1351- Deprecated output styles. Review options in `/output-style` and use --system-prompt-file, --system-prompt, --append-system-prompt, CLAUDE.md, or plugins instead
1352- Removed support for custom ripgrep configuration, resolving an issue where Search returns no results and config discovery fails
1353- Fixed Explore agent creating unwanted .md investigation files during codebase exploration
1354- Fixed a bug where `/context` would sometimes fail with "max_tokens must be greater than thinking.budget_tokens" error message
1355- Fixed `--mcp-config` flag to correctly override file-based MCP configurations
1356- Fixed bug that saved session permissions to local settings
1357- Fixed MCP tools not being available to sub-agents
1358- Fixed hooks and plugins not executing when using --dangerously-skip-permissions flag
1359- Fixed delay when navigating through typeahead suggestions with arrow keys
1360- VSCode: Restored selection indicator in input footer showing current file or code selection status
1361
1362## 2.0.28
1363
1364- Plan mode: introduced new Plan subagent
1365- Subagents: claude can now choose to resume subagents
1366- Subagents: claude can dynamically choose the model used by its subagents
1367- SDK: added --max-budget-usd flag
1368- Discovery of custom slash commands, subagents, and output styles no longer respects .gitignore
1369- Stop `/terminal-setup` from adding backslash to `Shift + Enter` in VS Code
1370- Add branch and tag support for git-based plugins and marketplaces using fragment syntax (e.g., `owner/repo#branch`)
1371- Fixed a bug where macOS permission prompts would show up upon initial launch when launching from home directory
1372- Various other bug fixes
1373
1374## 2.0.27
1375
1376- New UI for permission prompts
1377- Added current branch filtering and search to session resume screen for easier navigation
1378- Fixed directory @-mention causing "No assistant message found" error
1379- VSCode Extension: Add config setting to include .gitignored files in file searches
1380- VSCode Extension: Bug fixes for unrelated 'Warmup' conversations, and configuration/settings occasionally being reset to defaults
1381
1382## 2.0.25
1383
1384- 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
1385
1386## 2.0.24
1387
1388- Fixed a bug where project-level skills were not loading when --setting-sources 'project' was specified
1389- Claude Code Web: Support for Web -> CLI teleport
1390- Sandbox: Releasing a sandbox mode for the BashTool on Linux & Mac
1391- Bedrock: Display awsAuthRefresh output when auth is required
1392
1393## 2.0.22
1394
1395- Fixed content layout shift when scrolling through slash commands
1396- IDE: Add toggle to enable/disable thinking.
1397- Fix bug causing duplicate permission prompts with parallel tool calls
1398- Add support for enterprise managed MCP allowlist and denylist
1399
1400## 2.0.21
1401
1402- Support MCP `structuredContent` field in tool responses
1403- Added an interactive question tool
1404- Claude will now ask you questions more often in plan mode
1405- Added Haiku 4.5 as a model option for Pro users
1406- Fixed an issue where queued commands don't have access to previous messages' output
1407
1408## 2.0.20
1409
1410- Added support for Claude Skills
1411
1412## 2.0.19
1413
1414- Auto-background long-running bash commands instead of killing them. Customize with BASH_DEFAULT_TIMEOUT_MS
1415- Fixed a bug where Haiku was unnecessarily called in print mode
1416
1417## 2.0.17
1418
1419- Added Haiku 4.5 to model selector!
1420- Haiku 4.5 automatically uses Sonnet in plan mode, and Haiku for execution (i.e. SonnetPlan by default)
1421- 3P (Bedrock and Vertex) are not automatically upgraded yet. Manual upgrading can be done through setting `ANTHROPIC_DEFAULT_HAIKU_MODEL`
1422- Introducing the Explore subagent. Powered by Haiku it'll search through your codebase efficiently to save context!
1423- OTEL: support HTTP_PROXY and HTTPS_PROXY
1424- `CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC` now disables release notes fetching
1425
1426## 2.0.15
1427
1428- Fixed bug with resuming where previously created files needed to be read again before writing
1429- Fixed bug with `-p` mode where @-mentioned files needed to be read again before writing
1430
1431## 2.0.14
1432
1433- Fix @-mentioning MCP servers to toggle them on/off
1434- Improve permission checks for bash with inline env vars
1435- Fix ultrathink + thinking toggle
1436- Reduce unnecessary logins
1437- Document --system-prompt
1438- Several improvements to rendering
1439- Plugins UI polish
1440
1441## 2.0.13
1442
1443- Fixed `/plugin` not working on native build
1444
1445## 2.0.12
1446
1447- **Plugin System Released**: Extend Claude Code with custom commands, agents, hooks, and MCP servers from marketplaces
1448- `/plugin install`, `/plugin enable/disable`, `/plugin marketplace` commands for plugin management
1449- Repository-level plugin configuration via `extraKnownMarketplaces` for team collaboration
1450- `/plugin validate` command for validating plugin structure and configuration
1451- Plugin announcement blog post at https://www.anthropic.com/news/claude-code-plugins
1452- Plugin documentation available at https://code.claude.com/docs/en/plugins
1453- Comprehensive error messages and diagnostics via `/doctor` command
1454- Avoid flickering in `/model` selector
1455- Improvements to `/help`
1456- Avoid mentioning hooks in `/resume` summaries
1457- Changes to the "verbose" setting in `/config` now persist across sessions
1458
1459## 2.0.11
1460
1461- Reduced system prompt size by 1.4k tokens
1462- IDE: Fixed keyboard shortcuts and focus issues for smoother interaction
1463- Fixed Opus fallback rate limit errors appearing incorrectly
1464- Fixed /add-dir command selecting wrong default tab
1465
1466## 2.0.10
1467
1468- Rewrote terminal renderer for buttery smooth UI
1469- Enable/disable MCP servers by @mentioning, or in /mcp
1470- Added tab completion for shell commands in bash mode
1471- PreToolUse hooks can now modify tool inputs
1472- Press Ctrl-G to edit your prompt in your system's configured text editor
1473- Fixes for bash permission checks with environment variables in the command
1474
1475## 2.0.9
1476
1477- Fix regression where bash backgrounding stopped working
1478
1479## 2.0.8
1480
1481- Update Bedrock default Sonnet model to `global.anthropic.claude-sonnet-4-5-20250929-v1:0`
1482- IDE: Add drag-and-drop support for files and folders in chat
1483- /context: Fix counting for thinking blocks
1484- Improve message rendering for users with light themes on dark terminals
1485- Remove deprecated .claude.json allowedTools, ignorePatterns, env, and todoFeatureEnabled config options (instead, configure these in your settings.json)
1486
1487## 2.0.5
1488
1489- IDE: Fix IME unintended message submission with Enter and Tab
1490- IDE: Add "Open in Terminal" link in login screen
1491- Fix unhandled OAuth expiration 401 API errors
1492- SDK: Added SDKUserMessageReplay.isReplay to prevent duplicate messages
1493
1494## 2.0.1
1495
1496- Skip Sonnet 4.5 default model setting change for Bedrock and Vertex
1497- Various bug fixes and presentation improvements
1498
1499## 2.0.0
1500
1501- New native VS Code extension
1502- Fresh coat of paint throughout the whole app
1503- /rewind a conversation to undo code changes
1504- /usage command to see plan limits
1505- Tab to toggle thinking (sticky across sessions)
1506- Ctrl-R to search history
1507- Unshipped claude config command
1508- Hooks: Reduced PostToolUse 'tool_use' ids were found without 'tool_result' blocks errors
1509- SDK: The Claude Code SDK is now the Claude Agent SDK
1510- Add subagents dynamically with `--agents` flag
1511
1512## 1.0.126
1513
1514- Enable /context command for Bedrock and Vertex
1515- Add mTLS support for HTTP-based OpenTelemetry exporters
1516
1517## 1.0.124
1518
1519- Set `CLAUDE_BASH_NO_LOGIN` environment variable to 1 or true to to skip login shell for BashTool
1520- Fix Bedrock and Vertex environment variables evaluating all strings as truthy
1521- No longer inform Claude of the list of allowed tools when permission is denied
1522- Fixed security vulnerability in Bash tool permission checks
1523- Improved VSCode extension performance for large files
1524
1525## 1.0.123
1526
1527- Bash permission rules now support output redirections when matching (e.g., `Bash(python:*)` matches `python script.py > output.txt`)
1528- Fixed thinking mode triggering on negation phrases like "don't think"
1529- Fixed rendering performance degradation during token streaming
1530- Added SlashCommand tool, which enables Claude to invoke your slash commands. https://code.claude.com/docs/en/slash-commands#SlashCommand-tool
1531- Enhanced BashTool environment snapshot logging
1532- Fixed a bug where resuming a conversation in headless mode would sometimes enable thinking unnecessarily
1533- Migrated --debug logging to a file, to enable easy tailing & filtering
1534
1535## 1.0.120
1536
1537- Fix input lag during typing, especially noticeable with large prompts
1538- Improved VSCode extension command registry and sessions dialog user experience
1539- Enhanced sessions dialog responsiveness and visual feedback
1540- Fixed IDE compatibility issue by removing worktree support check
1541- Fixed security vulnerability where Bash tool permission checks could be bypassed using prefix matching
1542
1543## 1.0.119
1544
1545- Fix Windows issue where process visually freezes on entering interactive mode
1546- Support dynamic headers for MCP servers via headersHelper configuration
1547- Fix thinking mode not working in headless sessions
1548- Fix slash commands now properly update allowed tools instead of replacing them
1549
1550## 1.0.117
1551
1552- Add Ctrl-R history search to recall previous commands like bash/zsh
1553- Fix input lag while typing, especially on Windows
1554- Add sed command to auto-allowed commands in acceptEdits mode
1555- Fix Windows PATH comparison to be case-insensitive for drive letters
1556- Add permissions management hint to /add-dir output
1557
1558## 1.0.115
1559
1560- Improve thinking mode display with enhanced visual effects
1561- Type /t to temporarily disable thinking mode in your prompt
1562- Improve path validation for glob and grep tools
1563- Show condensed output for post-tool hooks to reduce visual clutter
1564- Fix visual feedback when loading state completes
1565- Improve UI consistency for permission request dialogs
1566
1567## 1.0.113
1568
1569- Deprecated piped input in interactive mode
1570- Move Ctrl+R keybinding for toggling transcript to Ctrl+O
1571
1572## 1.0.112
1573
1574- Transcript mode (Ctrl+R): Added the model used to generate each assistant message
1575- Addressed issue where some Claude Max users were incorrectly recognized as Claude Pro users
1576- Hooks: Added systemMessage support for SessionEnd hooks
1577- Added `spinnerTipsEnabled` setting to disable spinner tips
1578- IDE: Various improvements and bug fixes
1579
1580## 1.0.111
1581
1582- /model now validates provided model names
1583- Fixed Bash tool crashes caused by malformed shell syntax parsing
1584
1585## 1.0.110
1586
1587- /terminal-setup command now supports WezTerm
1588- MCP: OAuth tokens now proactively refresh before expiration
1589- Fixed reliability issues with background Bash processes
1590
1591## 1.0.109
1592
1593- SDK: Added partial message streaming support via `--include-partial-messages` CLI flag
1594
1595## 1.0.106
1596
1597- Windows: Fixed path permission matching to consistently use POSIX format (e.g., `Read(//c/Users/...)`)
1598
1599## 1.0.97
1600
1601- Settings: /doctor now validates permission rule syntax and suggests corrections
1602
1603## 1.0.94
1604
1605- Vertex: add support for global endpoints for supported models
1606- /memory command now allows direct editing of all imported memory files
1607- SDK: Add custom tools as callbacks
1608- Added /todos command to list current todo items
1609
1610## 1.0.93
1611
1612- Windows: Add alt + v shortcut for pasting images from clipboard
1613- Support NO_PROXY environment variable to bypass proxy for specified hostnames and IPs
1614
1615## 1.0.90
1616
1617- Settings file changes take effect immediately - no restart required
1618
1619## 1.0.88
1620
1621- Fixed issue causing "OAuth authentication is currently not supported"
1622- Status line input now includes `exceeds_200k_tokens`
1623- Fixed incorrect usage tracking in /cost.
1624- Introduced `ANTHROPIC_DEFAULT_SONNET_MODEL` and `ANTHROPIC_DEFAULT_OPUS_MODEL` for controlling model aliases opusplan, opus, and sonnet.
1625- Bedrock: Updated default Sonnet model to Sonnet 4
1626
1627## 1.0.86
1628
1629- Added /context to help users self-serve debug context issues
1630- SDK: Added UUID support for all SDK messages
1631- SDK: Added `--replay-user-messages` to replay user messages back to stdout
1632
1633## 1.0.85
1634
1635- Status line input now includes session cost info
1636- Hooks: Introduced SessionEnd hook
1637
1638## 1.0.84
1639
1640- Fix tool_use/tool_result id mismatch error when network is unstable
1641- Fix Claude sometimes ignoring real-time steering when wrapping up a task
1642- @-mention: Add ~/.claude/\* files to suggestions for easier agent, output style, and slash command editing
1643- Use built-in ripgrep by default; to opt out of this behavior, set USE_BUILTIN_RIPGREP=0
1644
1645## 1.0.83
1646
1647- @-mention: Support files with spaces in path
1648- New shimmering spinner
1649
1650## 1.0.82
1651
1652- SDK: Add request cancellation support
1653- SDK: New additionalDirectories option to search custom paths, improved slash command processing
1654- Settings: Validation prevents invalid fields in .claude/settings.json files
1655- MCP: Improve tool name consistency
1656- Bash: Fix crash when Claude tries to automatically read large files
1657
1658## 1.0.81
1659
1660- Released output styles, including new built-in educational output styles "Explanatory" and "Learning". Docs: https://code.claude.com/docs/en/output-styles
1661- Agents: Fix custom agent loading when agent files are unparsable
1662
1663## 1.0.80
1664
1665- UI improvements: Fix text contrast for custom subagent colors and spinner rendering issues
1666
1667## 1.0.77
1668
1669- Bash tool: Fix heredoc and multiline string escaping, improve stderr redirection handling
1670- SDK: Add session support and permission denial tracking
1671- Fix token limit errors in conversation summarization
1672- Opus Plan Mode: New setting in `/model` to run Opus only in plan mode, Sonnet otherwise
1673
1674## 1.0.73
1675
1676- MCP: Support multiple config files with `--mcp-config file1.json file2.json`
1677- MCP: Press Esc to cancel OAuth authentication flows
1678- Bash: Improved command validation and reduced false security warnings
1679- UI: Enhanced spinner animations and status line visual hierarchy
1680- Linux: Added support for Alpine and musl-based distributions (requires separate ripgrep installation)
1681
1682## 1.0.72
1683
1684- Ask permissions: have Claude Code always ask for confirmation to use specific tools with /permissions
1685
1686## 1.0.71
1687
1688- Background commands: (Ctrl-b) to run any Bash command in the background so Claude can keep working (great for dev servers, tailing logs, etc.)
1689- Customizable status line: add your terminal prompt to Claude Code with /statusline
1690
1691## 1.0.70
1692
1693- Performance: Optimized message rendering for better performance with large contexts
1694- Windows: Fixed native file search, ripgrep, and subagent functionality
1695- Added support for @-mentions in slash command arguments
1696
1697## 1.0.69
1698
1699- Upgraded Opus to version 4.1
1700
1701## 1.0.68
1702
1703- Fix incorrect model names being used for certain commands like `/pr-comments`
1704- 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.
1705- Windows: improve sub-process spawning to eliminate "No such file or directory" when running commands like pnpm
1706- Enhanced /doctor command with CLAUDE.md and MCP tool context for self-serve debugging
1707- SDK: Added canUseTool callback support for tool confirmation
1708- Added `disableAllHooks` setting
1709- Improved file suggestions performance in large repos
1710
1711## 1.0.65
1712
1713- IDE: Fixed connection stability issues and error handling for diagnostics
1714- Windows: Fixed shell environment setup for users without .bashrc files
1715
1716## 1.0.64
1717
1718- Agents: Added model customization support - you can now specify which model an agent should use
1719- Agents: Fixed unintended access to the recursive agent tool
1720- Hooks: Added systemMessage field to hook JSON output for displaying warnings and context
1721- SDK: Fixed user input tracking across multi-turn conversations
1722- Added hidden files to file search and @-mention suggestions
1723
1724## 1.0.63
1725
1726- Windows: Fixed file search, @agent mentions, and custom slash commands functionality
1727
1728## 1.0.62
1729
1730- Added @-mention support with typeahead for custom agents. @<your-custom-agent> to invoke it
1731- Hooks: Added SessionStart hook for new session initialization
1732- /add-dir command now supports typeahead for directory paths
1733- Improved network connectivity check reliability
1734
1735## 1.0.61
1736
1737- Transcript mode (Ctrl+R): Changed Esc to exit transcript mode rather than interrupt
1738- Settings: Added `--settings` flag to load settings from a JSON file
1739- Settings: Fixed resolution of settings files paths that are symlinks
1740- OTEL: Fixed reporting of wrong organization after authentication changes
1741- Slash commands: Fixed permissions checking for allowed-tools with Bash
1742- IDE: Added support for pasting images in VSCode MacOS using ⌘+V
1743- IDE: Added `CLAUDE_CODE_AUTO_CONNECT_IDE=false` for disabling IDE auto-connection
1744- Added `CLAUDE_CODE_SHELL_PREFIX` for wrapping Claude and user-provided shell commands run by Claude Code
1745
1746## 1.0.60
1747
1748- You can now create custom subagents for specialized tasks! Run /agents to get started
1749
1750## 1.0.59
1751
1752- SDK: Added tool confirmation support with canUseTool callback
1753- SDK: Allow specifying env for spawned process
1754- Hooks: Exposed PermissionDecision to hooks (including "ask")
1755- Hooks: UserPromptSubmit now supports additionalContext in advanced JSON output
1756- Fixed issue where some Max users that specified Opus would still see fallback to Sonnet
1757
1758## 1.0.58
1759
1760- Added support for reading PDFs
1761- MCP: Improved server health status display in 'claude mcp list'
1762- Hooks: Added CLAUDE_PROJECT_DIR env var for hook commands
1763
1764## 1.0.57
1765
1766- Added support for specifying a model in slash commands
1767- Improved permission messages to help Claude understand allowed tools
1768- Fix: Remove trailing newlines from bash output in terminal wrapping
1769
1770## 1.0.56
1771
1772- Windows: Enabled shift+tab for mode switching on versions of Node.js that support terminal VT mode
1773- Fixes for WSL IDE detection
1774- Fix an issue causing awsRefreshHelper changes to .aws directory not to be picked up
1775
1776## 1.0.55
1777
1778- Clarified knowledge cutoff for Opus 4 and Sonnet 4 models
1779- Windows: fixed Ctrl+Z crash
1780- SDK: Added ability to capture error logging
1781- Add --system-prompt-file option to override system prompt in print mode
1782
1783## 1.0.54
1784
1785- Hooks: Added UserPromptSubmit hook and the current working directory to hook inputs
1786- Custom slash commands: Added argument-hint to frontmatter
1787- Windows: OAuth uses port 45454 and properly constructs browser URL
1788- Windows: mode switching now uses alt + m, and plan mode renders properly
1789- Shell: Switch to in-memory shell snapshot to fix file-related errors
1790
1791## 1.0.53
1792
1793- Updated @-mention file truncation from 100 lines to 2000 lines
1794- 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).
1795
1796## 1.0.52
1797
1798- Added support for MCP server instructions
1799
1800## 1.0.51
1801
1802- Added support for native Windows (requires Git for Windows)
1803- Added support for Bedrock API keys through environment variable AWS_BEARER_TOKEN_BEDROCK
1804- Settings: /doctor can now help you identify and fix invalid setting files
1805- `--append-system-prompt` can now be used in interactive mode, not just --print/-p.
1806- Increased auto-compact warning threshold from 60% to 80%
1807- Fixed an issue with handling user directories with spaces for shell snapshots
1808- OTEL resource now includes os.type, os.version, host.arch, and wsl.version (if running on Windows Subsystem for Linux)
1809- Custom slash commands: Fixed user-level commands in subdirectories
1810- Plan mode: Fixed issue where rejected plan from sub-task would get discarded
1811
1812## 1.0.48
1813
1814- Fixed a bug in v1.0.45 where the app would sometimes freeze on launch
1815- Added progress messages to Bash tool based on the last 5 lines of command output
1816- Added expanding variables support for MCP server configuration
1817- Moved shell snapshots from /tmp to ~/.claude for more reliable Bash tool calls
1818- Improved IDE extension path handling when Claude Code runs in WSL
1819- Hooks: Added a PreCompact hook
1820- Vim mode: Added c, f/F, t/T
1821
1822## 1.0.45
1823
1824- Redesigned Search (Grep) tool with new tool input parameters and features
1825- Disabled IDE diffs for notebook files, fixing "Timeout waiting after 1000ms" error
1826- Fixed config file corruption issue by enforcing atomic writes
1827- Updated prompt input undo to Ctrl+\_ to avoid breaking existing Ctrl+U behavior, matching zsh's undo shortcut
1828- Stop Hooks: Fixed transcript path after /clear and fixed triggering when loop ends with tool call
1829- Custom slash commands: Restored namespacing in command names based on subdirectories. For example, .claude/commands/frontend/component.md is now /frontend:component, not /component.
1830
1831## 1.0.44
1832
1833- New /export command lets you quickly export a conversation for sharing
1834- MCP: resource_link tool results are now supported
1835- MCP: tool annotations and tool titles now display in /mcp view
1836- Changed Ctrl+Z to suspend Claude Code. Resume by running `fg`. Prompt input undo is now Ctrl+U.
1837
1838## 1.0.43
1839
1840- Fixed a bug where the theme selector was saving excessively
1841- Hooks: Added EPIPE system error handling
1842
1843## 1.0.42
1844
1845- Added tilde (`~`) expansion support to `/add-dir` command
1846
1847## 1.0.41
1848
1849- Hooks: Split Stop hook triggering into Stop and SubagentStop
1850- Hooks: Enabled optional timeout configuration for each command
1851- Hooks: Added "hook_event_name" to hook input
1852- Fixed a bug where MCP tools would display twice in tool list
1853- New tool parameters JSON for Bash tool in `tool_decision` event
1854
1855## 1.0.40
1856
1857- Fixed a bug causing API connection errors with UNABLE_TO_GET_ISSUER_CERT_LOCALLY if `NODE_EXTRA_CA_CERTS` was set
1858
1859## 1.0.39
1860
1861- New Active Time metric in OpenTelemetry logging
1862
1863## 1.0.38
1864
1865- Released hooks. Special thanks to community input in https://github.com/anthropics/claude-code/issues/712. Docs: https://code.claude.com/docs/en/hooks
1866
1867## 1.0.37
1868
1869- Remove ability to set `Proxy-Authorization` header via ANTHROPIC_AUTH_TOKEN or apiKeyHelper
1870
1871## 1.0.36
1872
1873- Web search now takes today's date into context
1874- Fixed a bug where stdio MCP servers were not terminating properly on exit
1875
1876## 1.0.35
1877
1878- Added support for MCP OAuth Authorization Server discovery
1879
1880## 1.0.34
1881
1882- Fixed a memory leak causing a MaxListenersExceededWarning message to appear
1883
1884## 1.0.33
1885
1886- Improved logging functionality with session ID support
1887- Added prompt input undo functionality (Ctrl+Z and vim 'u' command)
1888- Improvements to plan mode
1889
1890## 1.0.32
1891
1892- Updated loopback config for litellm
1893- Added forceLoginMethod setting to bypass login selection screen
1894
1895## 1.0.31
1896
1897- Fixed a bug where ~/.claude.json would get reset when file contained invalid JSON
1898
1899## 1.0.30
1900
1901- Custom slash commands: Run bash output, @-mention files, enable thinking with thinking keywords
1902- Improved file path autocomplete with filename matching
1903- Added timestamps in Ctrl-r mode and fixed Ctrl-c handling
1904- Enhanced jq regex support for complex filters with pipes and select
1905
1906## 1.0.29
1907
1908- Improved CJK character support in cursor navigation and rendering
1909
1910## 1.0.28
1911
1912- Slash commands: Fix selector display during history navigation
1913- Resizes images before upload to prevent API size limit errors
1914- Added XDG_CONFIG_HOME support to configuration directory
1915- Performance optimizations for memory usage
1916- New attributes (terminal.type, language) in OpenTelemetry logging
1917
1918## 1.0.27
1919
1920- Streamable HTTP MCP servers are now supported
1921- Remote MCP servers (SSE and HTTP) now support OAuth
1922- MCP resources can now be @-mentioned
1923- /resume slash command to switch conversations within Claude Code
1924
1925## 1.0.25
1926
1927- Slash commands: moved "project" and "user" prefixes to descriptions
1928- Slash commands: improved reliability for command discovery
1929- Improved support for Ghostty
1930- Improved web search reliability
1931
1932## 1.0.24
1933
1934- Improved /mcp output
1935- Fixed a bug where settings arrays got overwritten instead of merged
1936
1937## 1.0.23
1938
1939- Released TypeScript SDK: import @anthropic-ai/claude-code to get started
1940- Released Python SDK: pip install claude-code-sdk to get started
1941
1942## 1.0.22
1943
1944- SDK: Renamed `total_cost` to `total_cost_usd`
1945
1946## 1.0.21
1947
1948- Improved editing of files with tab-based indentation
1949- Fix for tool_use without matching tool_result errors
1950- Fixed a bug where stdio MCP server processes would linger after quitting Claude Code
1951
1952## 1.0.18
1953
1954- Added --add-dir CLI argument for specifying additional working directories
1955- Added streaming input support without require -p flag
1956- Improved startup performance and session storage performance
1957- Added CLAUDE_BASH_MAINTAIN_PROJECT_WORKING_DIR environment variable to freeze working directory for bash commands
1958- Added detailed MCP server tools display (/mcp)
1959- MCP authentication and permission improvements
1960- Added auto-reconnection for MCP SSE connections on disconnect
1961- Fixed issue where pasted content was lost when dialogs appeared
1962
1963## 1.0.17
1964
1965- We now emit messages from sub-tasks in -p mode (look for the parent_tool_use_id property)
1966- Fixed crashes when the VS Code diff tool is invoked multiple times quickly
1967- MCP server list UI improvements
1968- Update Claude Code process title to display "claude" instead of "node"
1969
1970## 1.0.11
1971
1972- Claude Code can now also be used with a Claude Pro subscription
1973- Added /upgrade for smoother switching to Claude Max plans
1974- Improved UI for authentication from API keys and Bedrock/Vertex/external auth tokens
1975- Improved shell configuration error handling
1976- Improved todo list handling during compaction
1977
1978## 1.0.10
1979
1980- Added markdown table support
1981- Improved streaming performance
1982
1983## 1.0.8
1984
1985- Fixed Vertex AI region fallback when using CLOUD_ML_REGION
1986- Increased default otel interval from 1s -> 5s
1987- Fixed edge cases where MCP_TIMEOUT and MCP_TOOL_TIMEOUT weren't being respected
1988- Fixed a regression where search tools unnecessarily asked for permissions
1989- Added support for triggering thinking non-English languages
1990- Improved compacting UI
1991
1992## 1.0.7
1993
1994- Renamed /allowed-tools -> /permissions
1995- Migrated allowedTools and ignorePatterns from .claude.json -> settings.json
1996- Deprecated claude config commands in favor of editing settings.json
1997- Fixed a bug where --dangerously-skip-permissions sometimes didn't work in --print mode
1998- Improved error handling for /install-github-app
1999- Bugfixes, UI polish, and tool reliability improvements
2000
2001## 1.0.6
2002
2003- Improved edit reliability for tab-indented files
2004- Respect CLAUDE_CONFIG_DIR everywhere
2005- Reduced unnecessary tool permission prompts
2006- Added support for symlinks in @file typeahead
2007- Bugfixes, UI polish, and tool reliability improvements
2008
2009## 1.0.4
2010
2011- Fixed a bug where MCP tool errors weren't being parsed correctly
2012
2013## 1.0.1
2014
2015- Added `DISABLE_INTERLEAVED_THINKING` to give users the option to opt out of interleaved thinking.
2016- Improved model references to show provider-specific names (Sonnet 3.7 for Bedrock, Sonnet 4 for Console)
2017- Updated documentation links and OAuth process descriptions
2018
2019## 1.0.0
2020
2021- Claude Code is now generally available
2022- Introducing Sonnet 4 and Opus 4 models
2023
2024## 0.2.125
2025
2026- Breaking change: Bedrock ARN passed to `ANTHROPIC_MODEL` or `ANTHROPIC_SMALL_FAST_MODEL` should no longer contain an escaped slash (specify `/` instead of `%2F`)
2027- Removed `DEBUG=true` in favor of `ANTHROPIC_LOG=debug`, to log all requests
2028
2029## 0.2.117
2030
2031- Breaking change: --print JSON output now returns nested message objects, for forwards-compatibility as we introduce new metadata fields
2032- Introduced settings.cleanupPeriodDays
2033- Introduced CLAUDE_CODE_API_KEY_HELPER_TTL_MS env var
2034- Introduced --debug mode
2035
2036## 0.2.108
2037
2038- You can now send messages to Claude while it works to steer Claude in real-time
2039- Introduced BASH_DEFAULT_TIMEOUT_MS and BASH_MAX_TIMEOUT_MS env vars
2040- Fixed a bug where thinking was not working in -p mode
2041- Fixed a regression in /cost reporting
2042- Deprecated MCP wizard interface in favor of other MCP commands
2043- Lots of other bugfixes and improvements
2044
2045## 0.2.107
2046
2047- CLAUDE.md files can now import other files. Add @path/to/file.md to ./CLAUDE.md to load additional files on launch
2048
2049## 0.2.106
2050
2051- MCP SSE server configs can now specify custom headers
2052- Fixed a bug where MCP permission prompt didn't always show correctly
2053
2054## 0.2.105
2055
2056- Claude can now search the web
2057- Moved system & account status to /status
2058- Added word movement keybindings for Vim
2059- Improved latency for startup, todo tool, and file edits
2060
2061## 0.2.102
2062
2063- Improved thinking triggering reliability
2064- Improved @mention reliability for images and folders
2065- You can now paste multiple large chunks into one prompt
2066
2067## 0.2.100
2068
2069- Fixed a crash caused by a stack overflow error
2070- Made db storage optional; missing db support disables --continue and --resume
2071
2072## 0.2.98
2073
2074- Fixed an issue where auto-compact was running twice
2075
2076## 0.2.96
2077
2078- Claude Code can now also be used with a Claude Max subscription (https://claude.ai/upgrade)
2079
2080## 0.2.93
2081
2082- Resume conversations from where you left off from with "claude --continue" and "claude --resume"
2083- Claude now has access to a Todo list that helps it stay on track and be more organized
2084
2085## 0.2.82
2086
2087- Added support for --disallowedTools
2088- Renamed tools for consistency: LSTool -> LS, View -> Read, etc.
2089
2090## 0.2.75
2091
2092- Hit Enter to queue up additional messages while Claude is working
2093- Drag in or copy/paste image files directly into the prompt
2094- @-mention files to directly add them to context
2095- Run one-off MCP servers with `claude --mcp-config <path-to-file>`
2096- Improved performance for filename auto-complete
2097
2098## 0.2.74
2099
2100- Added support for refreshing dynamically generated API keys (via apiKeyHelper), with a 5 minute TTL
2101- Task tool can now perform writes and run bash commands
2102
2103## 0.2.72
2104
2105- Updated spinner to indicate tokens loaded and tool usage
2106
2107## 0.2.70
2108
2109- Network commands like curl are now available for Claude to use
2110- Claude can now run multiple web queries in parallel
2111- Pressing ESC once immediately interrupts Claude in Auto-accept mode
2112
2113## 0.2.69
2114
2115- Fixed UI glitches with improved Select component behavior
2116- Enhanced terminal output display with better text truncation logic
2117
2118## 0.2.67
2119
2120- Shared project permission rules can be saved in .claude/settings.json
2121
2122## 0.2.66
2123
2124- Print mode (-p) now supports streaming output via --output-format=stream-json
2125- Fixed issue where pasting could trigger memory or bash mode unexpectedly
2126
2127## 0.2.63
2128
2129- Fixed an issue where MCP tools were loaded twice, which caused tool call errors
2130
2131## 0.2.61
2132
2133- Navigate menus with vim-style keys (j/k) or bash/emacs shortcuts (Ctrl+n/p) for faster interaction
2134- Enhanced image detection for more reliable clipboard paste functionality
2135- Fixed an issue where ESC key could crash the conversation history selector
2136
2137## 0.2.59
2138
2139- Copy+paste images directly into your prompt
2140- Improved progress indicators for bash and fetch tools
2141- Bugfixes for non-interactive mode (-p)
2142
2143## 0.2.54
2144
2145- Quickly add to Memory by starting your message with '#'
2146- Press ctrl+r to see full output for long tool results
2147- Added support for MCP SSE transport
2148
2149## 0.2.53
2150
2151- New web fetch tool lets Claude view URLs that you paste in
2152- Fixed a bug with JPEG detection
2153
2154## 0.2.50
2155
2156- New MCP "project" scope now allows you to add MCP servers to .mcp.json files and commit them to your repository
2157
2158## 0.2.49
2159
2160- Previous MCP server scopes have been renamed: previous "project" scope is now "local" and "global" scope is now "user"
2161
2162## 0.2.47
2163
2164- Press Tab to auto-complete file and folder names
2165- Press Shift + Tab to toggle auto-accept for file edits
2166- Automatic conversation compaction for infinite conversation length (toggle with /config)
2167
2168## 0.2.44
2169
2170- Ask Claude to make a plan with thinking mode: just say 'think' or 'think harder' or even 'ultrathink'
2171
2172## 0.2.41
2173
2174- MCP server startup timeout can now be configured via MCP_TIMEOUT environment variable
2175- MCP server startup no longer blocks the app from starting up
2176
2177## 0.2.37
2178
2179- New /release-notes command lets you view release notes at any time
2180- `claude config add/remove` commands now accept multiple values separated by commas or spaces
2181
2182## 0.2.36
2183
2184- Import MCP servers from Claude Desktop with `claude mcp add-from-claude-desktop`
2185- Add MCP servers as JSON strings with `claude mcp add-json <n> <json>`
2186
2187## 0.2.34
2188
2189- Vim bindings for text input - enable with /vim or /config
2190
2191## 0.2.32
2192
2193- Interactive MCP setup wizard: Run "claude mcp add" to add MCP servers with a step-by-step interface
2194- Fix for some PersistentShell issues
2195
2196## 0.2.31
2197
2198- Custom slash commands: Markdown files in .claude/commands/ directories now appear as custom slash commands to insert prompts into your conversation
2199- MCP debug mode: Run with --mcp-debug flag to get more information about MCP server errors
2200
2201## 0.2.30
2202
2203- Added ANSI color theme for better terminal compatibility
2204- Fixed issue where slash command arguments weren't being sent properly
2205- (Mac-only) API keys are now stored in macOS Keychain
2206
2207## 0.2.26
2208
2209- New /approved-tools command for managing tool permissions
2210- Word-level diff display for improved code readability
2211- Fuzzy matching for slash commands
2212
2213## 0.2.21
2214
2215- Fuzzy matching for /commands