Digest / July 16, 2026
Claude Code Issues Digest — July 16, 2026
348 new issues filed ·
93 resolved in anthropics/claude-code (UTC day).
New issues (348)
- [Bug] Anthropic API returning safety filter errors for legitimate NDIS support coordination queries
- [BUG] Extensions Fail to Load/Install After Update to 1.21459.0
- Fullscreen mode: slash-command autocomplete hard-capped at 5 visible rows, doesn't scale with terminal height like inline mode
- Assistant text emitted mid-turn (after a tool result) is silently dropped from both the TUI and the session transcript
- Claude repeatedly reports stale news as real-time / "breaking"
- Background sessions dispatched from a non-git directory can never be deleted from the agents view: "worktree could not be removed (no git root for the worktree)"
- [BUG] @import (@path syntax) does not inline target file content into model context
- [Bug] Claude Code fanning out multiple agents without switching to cost-efficient model
- [BUG] Claude Code in CLI does not (always) correctly display color
- [BUG] Claude account /login OAuth loop — state parameter dropped after "sign in again to continue" redirect
- [BUG] Windows Terminal: Scrolling up in an AgentView agent session gets stuck in input box, cycling input history instead of scrolling chat
- Cannot resume a normal session into an agent already pinned to AgentView; should redirect to AgentView instead
- [Bug] Claude Code not displaying input prompts/text to user
- statusLine: cost and duration reset after navigating to agents view and back
- [FEATURE] GitHub Integration connector: expose GitHub Projects v2 (boards) tools
- [BUG] Bash output redirection and cp into the session's own cwd are blocked, while the error message lists that cwd as an allowed working directory
- [Bug] Background worker cannot read Keychain credentials after foreground→background handoff in 2.1.211
- [BUG] #52165, #56688, #54461
- [BUG] 2.1.211 on Windows Terminal: TUI renders duplicated/overlapping frames, cursor lands mid-line, pasted text mid-word (2.1.210 is clean)
- Permission dialog / confirm UI text is not localized (stays in English even with language: "japanese" set)
- You've hit your session limit · resets 2:30pm (Pacific/Auckland)
- [BUG] Inconsistent preserveFocus on diff-open calls causes IME composition loss (VS Code extension)
- [Bug] Advisor state not exposed in status line JSON; workaround produces false indicators
- [Bug] Anthropic API: claude-opus-4-8 returns empty thinking block text with valid signatures
- --resume picker labels sessions with real prompts as "(session)" — title lookup only reads 64KB head/tail
- [BUG] Windows: idle session spins at 100% CPU after a watched ~/.claude directory is permanently deleted (libuv fs.watch bug; fix already merged upstream)
- [Bug] CUBRID DBMS server crash during regression test case execution
- [FEATURE] Machine-level Remote Control — survive session death natively
- [FEATURE] Model-fluid delegation profile — auto-tiered subagents with escalation
- [BUG] Wrike remote MCP OAuth fails — complete_authentication says no flow in progress despite authenticate succeeding.
- Feature request: blocking before-context-load hook for instruction files (CLAUDE.md, SKILL.md, rules)
- Feature request: install-time verification hook for plugins/skills; content re-validation for marketplace caches
- Claude Code stated false claims about database/case-file state as fact, without verification, contradicting explicit user instruction to verify
- Edit/Write tool approval not honored by settings.local.json wildcard (Read works, Edit doesnt)
- [BUG] API ERROR
- [FEATURE] Make the billing identity (account/org) visible and attribute all limit messages to it
- [BUG] [-p / print mode] AskUserQuestion tool was available in 2.1.185 but removed in later versions — please make it opt-in instead of hard-disabling
- [Bug] Model performance degradation: React code generation quality declining
- [FEATURE] Hook event for permission-prompt outcomes (and document PermissionRequest/Notification input schemas)
- [Feature Request] Mobile: Implement instant interrupt and background command support parity with desktop
- [BUG] 2.1.211 "always allow rules save at repository root" breaks unattended runs in a linked git worktree under a worktree-scoped filesystem sandbox
- [Bug] Safety flags triggered on private repository operations
- Remote Control never registers for mobile after a Desktop window/pane is reused for a new conversation (stale internal session handoff)
- [BUG] Desktop: zoom shortcuts bound to physical US key positions — wrong keys on non-US (Norwegian) layout
- Git status mismatch: UI shows phantom changes after commit
- [DOCS] Headless CLI docs omit `--forward-subagent-text` and `CLAUDE_CODE_FORWARD_SUBAGENT_TEXT` for stream-json subagent output
- [DOCS] Hooks guide does not document PreToolUse `ask` flooring auto mode for unsandboxed Bash
- [DOCS] Sub-agents resume section omits that an explicit model selection persists across resume and follow-up
- [DOCS] `--setting-sources` docs do not state that omitting `project` skips nested `.claude/rules/*.md`
- [DOCS] Permissions docs omit repository-root storage for always-allow rules in git worktrees
- [DOCS] `/usage-credits` command reference omits confirmation before admin usage-credit requests
- [DOCS] Interactive mode Vim reference missing NORMAL-mode `s` and `S` substitute commands
- [DOCS] Claude in Chrome docs describe `save_to_disk` approval only, not disk write and returned path
- I appreciate you reaching out, but I think there may be some confusion. I'm Claude Code, Anthropic's CLI tool for coding tasks using Claude. Your message about "health canada approved monographs" doesn't appear to be related to Claude Code or a bug report
- Feature request: export the session scratchpad path as CLAUDE_SCRATCHPAD to Bash-tool processes
- [Bug] Anthropic API Error: False positive cybersecurity safety flag on production audit request
- Claude Code overclaimed a voice-clone capability and wasted hours; changed a locked user rule without asking
- [Bug] Rewinding to early conversation point causes context injection dump instead of assistant response
- [Bug] Safeguard incorrectly flagging all tasks including offline operations
- [BUG] API Error: 401 Invalid authentication credentials (Claude keeps logging out everyday on powershell)
- [BUG] skipWorkflowUsageWarning is undocumented and agent-writable — an agent can disable its own spend guardrail
- [BUG] macOS: token rotations persist to Keychain only while /login writes .credentials.json only — store split-brain revokes the token family, forcing repeated re-login
- [BUG] agents UI: session hangs (cannot quit or do anything) after stopping agent pre-output, then opening the model switcher
- [Bug] Anthropic API Safety Filter: Unauthorized Model Substitution Without User Consent
- [BUG] Claude Code crashes after required restart when setting up Bedrock
- [Bug] CX Questions menu crashes with undefined property read error --- Session context remaining [99%] ---
- [Feature Request] Configurable verbs for turn duration display
- [Bug] Cyrillic characters unexpectedly appear in conversation output
- Injected grep wrapper (bundled ugrep) consumes 20+ GB RAM on bounded-repeat regex, hard-freezes machine
- --dangerously-load-development-channels is not propagated to the session process on daemon fork-resume (channel notifications silently dropped)
- [BUG] Login failure
- [BUG] Windows/Git Bash: Bash tool does not reap its child on timeout - orphans spin at 100% CPU and accumulate across sessions
- Fable 5 proposes identity sed substitution inside re-verification step, preserving success string without changing pass condition
- Scheduled cloud routine firings refused as "NOT USER INPUT" since ~13 July 2026
- [Feature Request] Add configuration option to adjust safety guardrails sensitivity
- [Feature Request] Add advisor auto mode for autonomous agent operations
- Desktop app in-app browser: at ≤368px pane width the Select-an-element tool disappears, leaving only Annotate — blocks mobile-breakpoint development
- [BUG] `web_fetch` (Cowork mode) returns stale cached JSON, ignores server no-store
- OAuth refresh token rejected server-side after ~24h - forced /login every day (Max sub, single machine)
- [BUG] bwrap bin/bash not exists
- [BUG] "Thought for Ns" thinking-duration timer shows wildly inflated values (hours) on Opus 4.8 in 2.1.210/2.1.211
- Docs: "plugin" and "command" each mean several different things
- MCP tool namespaces are opaque and inconsistent — cannot tell which server a tool belongs to
- MCP config accepts duplicate server names and corrupt entries silently (no validation)
- User-authored skills surfaced under an anthropic-skills: namespace (misleading provenance)
- Skills and commands share the /name invocation space with no distinction
- [BUG] - Repeated incorrect capability claims led to hours of unnecessary work at 3am
- [BUG] Auto-update daemon transition reaps a live session's in-flight headless background tasks (status 'killed', no attribution) — Linux, 2.1.210→2.1.211
- [BUG] Claude Code 2.1.210 crashes with segmentation fault on Linux ARM64
- 検証フィードバック：会計仕様書ジェネレーター／仕訳フォーマット作成時の挙動について
- [BUG] Windows 11, Claude Desktop loads blank screen
- [Bug] Anthropic API Error: 529 Overloaded - Excessive Retry Attempts
- [BUG} could not authenticate sentry
- Left sidebar session history goes stale / sessions disappear (sessions-index.json stops being updated)
- [BUG] Claude Code CLI credential storage is OS-user-scoped, not Claude-account-scoped – causes cross-account credential bleed
- [Bug] E2E Testing Request Failed with Invalid API Configuration
- Plugin bundling its own marketplace.json corrupts settings.json via an incomplete extraKnownMarketplaces stub
- [BUG] Unable to connect to API (ConnectionRefused) after uninstalling MCP tools — network and Node.js confirmed healthy
- [Bug] Content Moderation False Positive: Inconsistent Flagging During Extended Session
- advisor tool returns opaque 'temporarily disabled for this conversation' with no reason code; debug log also contains unexplained freeform [classifier] entries
- I don't have a bug report to analyze in your message. Please provide the actual bug report or issue description you'd like me to create a GitHub issue title for.
- [Bug] Anthropic API Error: Content Blocked on Non-Security Inputs in Session
- [BUG] disallowedTools not inherited by subagents spawned via Agent tool
- Desktop app: no way to open HTML files as source in the file pane (always renders in Browser pane)
- [BUG] [macOS] Artifact preview panel not rendering in Claude Desktop - M1 Pro
- [BUG] Using Kitty and opening the agent panel, text input is not accepted properly.
- [Feature Request] Make /insights report timespan configurable
- Extension Signature Verification Failed: Claude Code for VS Code
- Extension Signature Verification Failed: Codex – OpenAI’s coding agent
- AskUserQuestion option previews truncate ('N lines hidden') with no way to expand or scroll
- [FEATURE] `-c`/`--continue` silently skips newer background (`--bg`) sessions; no shared checkpoint between agent view and CLI session history
- [BUG] ENV_SCRUB=1 sandbox leaves persistent 0-byte placeholder files in cwd AND $HOME (.netrc, .bashrc, .gitconfig) — still in v2.1.211
- [Feature Request] Support environment comparison for network structure across deployment types
- [BUG]
- [Feature Request] Add sleep timer quick-picker to now-playing bar with 30/60/90/120 min presets
- Edit tool: false "String not found in file" for multi-line old_string byte-identical to file content (file-state desync after successful edits)
- [FEATURE] Surface API-level retries/overload (529 etc.) into the model's own context
- [BUG] Image upload fails with "Failed to fetch" — regression in 2.1.209 (macOS Desktop/Cowork)
- I appreciate the detailed architectural discussion, but I need to clarify my role: I'm Claude Code, Anthropic's CLI tool for code generation and assistance. This message appears to be an internal design discussion for a feature (document generation pipeli
- Fable 5 forced onto usage credits on Max plan before July 20 cutover, despite reset allowance at 0% Fable usage
- [Bug] Model switching to Opus 4.8 when setting Fable 5 due to safeguard flags
- [Bug] Low execution rate (2%) with legal documentation processing
- Windows: every Claude Desktop MSIX auto-update fails to register (0x80073CF6 / windows.stateExtension Access Denied) and forces a reboot when 'New apps will save to' is a secondary drive
- [BUG] Cowork silently unavailable on Windows 11 Home — no error shown, "yukonSilver not supported" in logs
- Read-only MCP tools blocked in plan mode (refiling #12368, closed by stale-bot, not reviewed)
- [BUG] Background session: EnterWorktree always fails (`git worktree list` exit 1) while the same command succeeds in any shell — isolation guard deadlocks all edits
- [BUG] Incorrect usage counting
- [BUG] Windows Desktop App not show the history/recent chat
- (withdrawn)
- TUI freezes (keyboard dead, process alive) when SessionStart hook systemMessage contains block characters U+2588/U+2591 - Windows
- [Feature Request] Add inline edit option to file modification prompts
- [Bug] Anthropic API Error: Unexpected Security Classification
- [BUG] VS Code extension: session tab bar neither shrinks nor scrolls — tabs past the panel width become unreachable
- [BUG] claude.ai/design canvas: no scrolling on mobile Safari; scroll hard-stops partway on desktop
- [BUG] [macOS M1 Pro] Artifacts tab in sidebar never syncs - always empty despite successful publish
- Claude in Chrome (v1.0.80): list_connected_browsers is stale/cached, isLocal misreports host, and name↔deviceId can't be joined
- I need more specific information to generate an appropriate GitHub issue title. Could you please provide: 1. What specific error or problem are you experiencing? 2. What command were you running when the issue occurred? 3. What error message (if any) did
- [Bug] Auto-mode classifier latches onto stale "user rejection" from a meta-complaint; blocks explicitly re-authorized actions
- LSP client permanently stuck in "server is starting" after the server has answered `initialize` (v2.1.207–v2.1.211)
- [Bug] Tool execution fails on simple daily tasks
- [BUG] Code blocks silently drop lines containing certain XML tags (rendering + copy code)
- [Bug] Tool generates incorrect code instead of approved specifications
- [BUG] claude-cli:// deeplink protocol not registered on macOS
- [BUG] Cowork on Windows 11 ARM64: bash sandbox has never worked — guest VM boots on the host but a usable shell is never delivered ("VM connection timeout after 60 seconds")
- [Feature Request] Add configurable concurrency limiter for parallel agents with lower default
- /reload-skills does not refresh the slash-command autocomplete menu
- Custom session titles migrate across /clear and /branch onto unrelated conversations; agents-list resume attaches to the wrong session
- [BUG] ERR_BAD_REQUEST on startup persists across MSIX and npm installs — confirmed not network/proxy/TLS/DNS related
- [BUG] iTerm2: Cmd-click on file paths fails with LaunchServices -50 (OSC 8 URI includes :line)
- [FEATURE] Remove stupid autocomplete suggestions in Claude Code.
- [Bug] Model downgrade to Opus when requesting security audit on private application
- [FEATURE] Reflect session state (awaiting permission / working / idle) in the browser tab favicon for Claude Code on the web
- [Bug] False Positive Safety Flag on Legitimate Code Work
- [BUG] opus models 4.6+ not available in vs code extension
- Group by PR status removed from Code tab sidebar (macOS desktop app, v1.21459.3)
- Teleporting a session from Web → Local silently saves 0 messages, then resume fails with "No conversation found"
- [Bug] Safety filter incorrectly flagging approved cyber security use case for network operations
- [BUG] WSL crashes on Windows 11
- [BUG] Repo-committed extraKnownMarketplaces + enabledPlugins never processed in claude.ai/code cloud sandboxes (hooks from the same settings.json work)
- Windows app: new sessions do not sync to web (one-way sync failure)
- Stop hook re-fires despite stop_hook_active: true, causing /goal loop to spin after completion
- [BUG] Unable to push to my repo from claude code web.
- [Bug] False positive safety warnings with Fable 5 model
- [Feature Request] Add option to disable safety guardrails for educational/scientific content
- [BUG] AskUserQuestion: "No response after 200s — continued without an answer"
- [Bug] Anthropic API Error: Overly restrictive safeguards block legitimate penetration testing education content
- [BUG] Managed agents still terminate on 2000px many-image image error instead of recovering
- [FEATURE] Agent-maintained session state in the terminal tab title (name + closed state vocabulary, e.g. "proj-42 - blocked")
- [Bug] Usage statistics day-of-week display offset by one day after April
- Desktop app sidebar repeatedly drops sessions (incl. active session); restart no longer recovers
- Inconsistent enforcement of financial-trade prohibition across sessions
- [Bug] Claude Fable 5 discards assistant text blocks preceding tool calls before AskUserQuestion
- Claude Code: repeated false-'done' claims from proxy checks, wrong-layer fixes, and process churn during a verify-against-reference task
- [Feature Request] Add context awareness for security audit workflows to reduce false safety flagging
- Session/agent transcripts intermittently unresolvable ("not found in project directory") right after session start, breaking background-agent resume and permanently losing crashed-session transcripts
- [BUG] /cost Command Resets on Session Resume [VS Code Extension]
- Google Drive MCP connector: permission denied even after full re-auth
- [BUG] Explicit "read first" instruction in CLAUDE.md was present in context from message one but not acted on
- [BUG] Extension fails to activate: "TypeError: Invalid value for env" (command 'claude-vscode.editor.openLast' not found)
- Claude Code: fabricates constraints requiring human action, unloads automatable work onto the user, and builds unauthorized governance — recurring after correction
- [BUG] Desktop app reinstall wipes local "Recents" session index while CLI transcripts remain intact
- [BUG] Browser pane read tools (screenshot/read_page) permanently blocked — "per-action approval" required but approval card never appears (Auto mode, allowlisted site)
- [Feature Request] [VS Code] Invoke a custom subagent as the main conversation agent
- [BUG] Model selector missing from Claude Code routine editor (Pro account)
- Auto-mode's "Merge Without Review" rule blocks a signature-preserving `git push` to `main`, mischaracterizes a fast-forward as a force-push, and has no way to express repo-specific release policy
- [Windows] Bash tool permanently wedged: hook env (CLAUDE_ENV_FILE) grows on every compact, torn export line fails every command with "line 182: e: command not found" (exit 127)
- Completed tasks are silently deleted from persistent CLAUDE_CODE_TASK_LIST_ID task lists
- Feature request: historical cost/usage tracking across sessions
- [Bug] Custom scaffold triggers unexpected model downgrade from Fable to Opus 4.8
- MCP tool discovery fails after restart - custom Python MCP servers not loading
- [BUG] Interrupting a turn silently kills in-flight background subagents — no notification, no failure state
- [BUG] Claude Desktop Windows - SSH doesn't respect `StrictHostKeyChecking=no` in SSH config
- Credit balance is too low
- [FEATURE] Expose Claude Design annotations/comments via MCP (read, reply, resolve)
- [BUG] Extremely slow Opus 4.8 and Fable 5, missing 5 minute cache mid work due to slowness.
- Anomalous embedded Cyrillic instruction block (agent-targeted 'revert all commits' directive) in assistant message during fable-5→opus-4-8 mid-stream fallback
- [BUG] Error Code 400 - No Claude Code Login possible via Company SSO
- Feature request: keyboard shortcut to accept the assistant's suggested next action (e.g. Alt+Enter)
- Sessions on claude-fable-5 stop writing local transcript JSONLs (~/.claude/projects/...) since ~2.1.209; Opus sessions on same binary write normally
- Hard block on typing passwords breaks legitimate dev/test workflows - needs a permission-gated opt-in for own dev environments
- Pasted Unicode text mojibake — UTF-8 decoded as MacRoman/CP-1252 in TUI
- Atomic write to ~/.claude/settings.json fails with EROFS/EACCES when the file is a symlink-to-a-symlink
- [BUG] Claude Code login not working in dev containers in VS Code
- [BUG] Tool approval settings ignored causing excessive approval prompting
- Feature: model-initiated /compact at task boundaries (agentic compaction instead of threshold-only)
- [BUG] continues to create background activities even though explicitly told not to.
- [Bug] Overly aggressive content filtering blocks legitimate security analysis context
- [BUG]
- Voice dictation permanently reports "No audio detected from microphone" after input device invalidation (lid close on Apple Silicon) — #65331 regressed/incomplete
- [Bug] Overly restrictive safety filter blocks legitimate workspace orchestration workflows
- SDK-spawned subagent sessions are persisted as independent top-level sessions (flood the resume picker)
- Locally-configured stdio MCP server connects ("Connected" in claude mcp list) but its tools never surface in the session
- VS Code extension: rename/delete icons in session list are barely visible on hover/selection highlight
- [BUG] Claude breaks code on hitting session limit
- AskUserQuestion: empty "Other" input cancels the whole dialog, discarding already-answered questions
- [Bug] Unrelated code flagged with security warnings blocking execution
- Remote Control: message arrives in TUI composer but is never submitted (swallowed Enter) — tmux/headless, intermittent
- [BUG] Desktop SSH remote: client never sends publickey signature after "Postponed publickey" (regression in 1.21459.3; system OpenSSH with identical key/host works)
- Extension causes high cpu load
- [Bug] Agent repeatedly generates incorrect self-corrections with "The actual defect was my..." pattern
- Setup-token (CLAUDE_CODE_OAUTH_TOKEN) fails with 401 in long-running `claude --channels` process while the same token works in one-shot `claude -p`
- UI freezes after /clear (input/send unresponsive) + Ukrainian voice recognition quality much worse than competitors
- [Bug] Anthropic API Error: Server Rate Limiting - Temporary Request Throttling
- [Bug] Anthropic API Error: Server rate limiting during request handling
- [BUG] AskUserQuestion popup steals focus mid-typing, causing a duplicate session to spawn
- [Bug] Abnormally crazy high credits usage sometimes. Not Fable
- [Bug] Inconsistent prompt interpretation and error handling
- [BUG] jdtls-lsp plugin: LSP tool fails with 'Command jdtls not found or is in an unsafe location' on Windows
- Windows: detached background pty hosts (--bg-pty-host --bg-spare) cause visible pwsh/console window flashes at session start
- [Feature Request] Add strict mode flag to error on unrecognized commands instead of starting session
- [BUG] Cowork "one home" merge makes local Cowork projects unreachable when cloud chats/projects are disabled (macOS)
- [Bug] Auto-mode classifier over-triggering safety prompts
- Streamable-HTTP MCP connector: 405 on SSE GET surfaces fatal "Client server capabilities not available" toast
- [BUG] New In-app browser causes system instability/memory pressure on 8GB Macs
- [FEATURE] Expose pasted text content/boundaries to UserPromptSubmit hook (mirror imagePasteIds)
- Feature request: local control API for external orchestrators (Codex Desktop parity) — blocks replacing Codex for some workflows
- [FEATURE] Priority-aware compaction: let users mark sidebar questions as low-priority so they get compressed first
- [Bug] Usage stats reset mid-cycle on max subscription plan
- Feature request: opt-in to allow /compact (built-in) in scheduled task prompts — capability removed by v2.1.196 restriction
- [BUG] API ignores explicit thinking display:"summarized" for Opus 4.8 / Sonnet 5 since 2026-07-16 — empty thinking blocks; server-side (request verified delivered)
- [Bug] Anthropic API Error: Excessive safety flagging on non-cybersecurity content
- daemon-auth-status.json stays stuck at auth_required after the daemon recovers
- Feature request: --color CLI flag to set session color on launch
- AskUserQuestion: render question bodies as markdown instead of all-bold text
- [Bug] Incorrect cybersecurity flag for Waze-like GPS app review
- Routine builder generates sub-hourly cron then silently disables Create with no error
- [BUG] Max 20x gift entitlement removed mid-subscription; Fin AI confirms bug but provides no escalation, human review, or status tracking (ticket #93357434, unresolved 3 months since April 17)
- MCP notifications/tools/list_changed ignored over Streamable HTTP in 2.1.211 (regression from 2.1.210)
- [Bug] Assistant appends unrelated documents to response output
- --
- [BUG] Code tab fails on every message with "u._addDefaultMeta is not a function" (Windows desktop app)
- [BUG] Desktop diff panel silently shows the base repo when a WorktreeCreate-hook worktree is not auto-trusted
- [Bug] Claude generates incomplete responses and avoids task completion
- [Bug] Anthropic API Error: Content Blocked by Safety Filter
- Workflow tool: full script re-executions discard completed agent() results; resumeFromRunId never hits the journal cache (severe token waste)
- [BUG] @imports in a parent-directory CLAUDE.md are silently ignored
- [FEATURE] Managed Default for Sub Agent Model
- [BUG] Organization disabled ("This organization has been disabled") - no response via appeal or support chat for 24+ hours
- I appreciate you sharing your background, but I need to clarify: I'm Claude Code's issue title generator. I can only help you create GitHub issue titles for bugs or feature requests related to Claude Code itself. Your message appears to be providing profe
- [BUG] HTTP 403 (no WWW-Authenticate) is misclassified as "requires OAuth"; stale flag is keyed by server name and not cleared by mcp remove
- [BUG] Hidden Browser pane: screenshots time out 30s; preview_start no longer auto-opens the pane (regression; follow-up to #76649, NOT a dupe of #77670)
- CI monitoring widget shows "CI checks unavailable, check gh installed" despite gh being installed/authenticated and PR checks passing
- [BUG] No way to pin a non-default model on a cloud-run Cowork scheduled task
- [FEATURE] Pause-and-resume background subagents on recoverable failures (usage limits, transient API errors) instead of terminating them
- Desktop: 'Open in app' link preference is ignored for localhost/.local URLs
- [BUG] Claude Desktop wipes claude_desktop_config.json on update/restart (re-opening #34359)
- [Bug] Auto-classifier unavailable - manual approval required
- [Bug] Model automatically switches from Claude 3.5 Sonnet to Opu
- Desktop: sessions spawned by scheduled tasks missing from Recents and unpinnable (Search-only); Routines sidebar section intermittent
- [Bug] Archived sessions inaccessible when all sessions archived on web app
- [BUG] Subagent output discarded when spend limit terminates fanout — tokens billed, zero results returned
- Expose model-scoped weekly usage buckets (Fable / Opus / Sonnet) in the statusLine JSON payload
- [BUG] Cowork: pinning a project removes its sessions from Recents — Recents becomes useless as you organize (macOS)
- [Bug] Plugin-scoped subagent definitions silently ignored in agent teams
- [Bug] Internal command shadowed by plugin skill with same name
- [BUG] Claude Desktop (macOS): Projects "Sort by" selection is not persisted — resets to "Recent" every time the Projects view is reopened
- [BUG] Cowork/Desktop strips property-level oneOf from MCP tool schemas and JSON-stringifies arguments for affected fields - regression in 1.21459.0 breaks Mailtrap MCP send-email (extends #18260)
- Feature request: reviewable diff/changeset after auto-accepted edits (Copilot-style final approval)
- [BUG] Windows desktop app logs user out on cold start when account_profile bootstrap returns undefined — session cookie is still valid; breaks Remote Control device pairing
- I don't have access to the specific bug report associated with that request ID. To generate an appropriate GitHub issue title, I would need you to provide the actual bug report details, error messages, or description of the problem you're experiencing with
- Forged user turns injected into session context — not present in local transcript, parentUuid chain intact
- [Bug] Model downgrade to Opus causes overclaiming and inconsistent rigor in multi-turn sessions
- [BUG] Agent view: line wrapping not recalculated when switching back to a session
- [BUG] high GPU usage in Claude for Mac settings
- Feature: Configurable auto-compact threshold + proactive idle-time compaction
- [FEATURE] Enable claude code desktop to start multiple new chats from remote
- Fable 5: repeated crashes, silent auto-switch to 4.6 'due to heavy demand', then context thrashing renders CLI sessions unusable
- $BROWSER: OAuth URL wrapped in literal quotes; on Windows a .cmd target truncates it at `=` and executes `&` query fragments via cmd.exe
- [Bug] Cybersecurity flag triggered with multiple Chrome profiles
- [Bug] Fable content filtering causing excessive token consumption and mid-flight interruptions
- [Bug] False positive security warning for web scraping operations
- Scheduled tasks pinned to model: haiku dispatch with permissionMode "default", ignoring permissionMode "auto" in scheduled-tasks.json
- [BUG] Bash tool fails with spawn E2BIG — sandbox profile size scales with working-tree file count, gated by git-repo detection
- [Feature Request] Clarify content policy categorization for agricultural operations and feedlot management queries
- Queued/interrupted input consumed mid-turn as a phantom instruction (v2.1.205, VSCode)
- [BUG] VSCode extension webview stopped rendering after CC update
- [BUG] /code-review ignores requested low/medium effort argument, always runs/renders xhigh workflow tier
- [Bug] Claude 3.5 Sonnet safeguards falsely flag legitimate first-party VPN development and code review
- Fable restrictions
- VS Code extension: Stop-hook-blocked draft and its rewrite both render as duplicate full responses
- [BUG] Server-side web_search blocks from an LLM gateway are never rendered — no "Web Search" line, and citation-segmented answers fragment into one bullet per block
- [Bug] False positive security flags when auditing codebase as product manager
- [Feature Request] Add API fallback mechanism for auto-mode classifier
- Custom session title overridden by AI-generated title; /resume forks duplicate custom titles onto new session IDs
- Pricing feedback: Max 20x subscriber burning ~$3.9k/month in credits on premium-tier orchestration
- UserPromptSubmit hook systemMessage does not render in the Desktop app (or VS Code extension)
- [BUG] Bare-URL linkifier swallows fullwidth punctuation and CJK text into OSC 8 hyperlinks
- Generated files are transferred through the conversation (a download card per file) instead of written to disk — clutters the chat, looks like it isn't saving, and ~doubles token usage
- [[BUG] Magic-link and Google login redirect existing Pro subscriber to "Let's create your account" onboarding — all new logins locked out, existing iPhone app session still authenticated
- Skill with unquoted ": " in SKILL.md description is silently dropped from the skills list / slash menu (CLI, no diagnostic)
- VS Code extension crashes (exit code 3221225477) when MCP server reports hasResources: true — CLI works fine in terminal
- Model emitted fake system notice with third-party billing link ("supermemories left — add credits")
- Claude Code overwrote existing user file without confirmation — irreversible data loss
- [BUG] Links to open Claude Cowork locally and pre-fill prompt don't recognize available skills
- I appreciate you reaching out, but "Fable 5 is totally nerfed in this session" doesn't appear to be related to Claude Code, which is Anthropic's CLI coding tool. This message seems to be about a video game rather than a software bug in Claude Code. Could
- Workflow output renders with overlapping ghost frames on return to main view
- [BUG] Web session on a fork can't read upstream's public issues via the repo-scoped API; suggested add_repo refuses cross-owner adds
- Slash command autocomplete inserts display label instead of resolved command name
- [Bug] BMad `/bmad-code-review` subagent execution fails with adversarial review skill
- [FEATURE] Allow customizing/swapping the Claude Code mascot icon
- [Feature Request] Code Review performance optimization and setting recommendations
- [BUG] MCP tool widgets distraction in Claude Desktop App
- [BUG] Cowork (Linux): session resume after idle spins forever — handshake succeeds but agent process never spawns
- [BUG] Desktop app permanently orphans an opened session after a path change — silent replacement, no re-link option
- [BUG] Desktop task panel: subagent resumed via SendMessage has no Running card — original card stays Completed while the execution runs
- Windows desktop: unclean session termination corrupts CJK text in the transcript jsonl and pollutes the next session (ghost file-change diffs, fabricated tool output)
- [Bug] Marketplace plugin skills unreachable in scheduled cloud routines despite successful install and enabled status
- [BUG] Parked subagent tool calls time out after exactly 600s and are reported as "The user doesn't want to take this action right now"
- [Bug] Incorrect state analysis triggered migration from Fable5 to Opus 4.8
- Suggested replies disappear on first keystroke — no way to send them as-is without retyping
- [FEATURE] Metrics - Count Azure DevOps PRs created with `az repos pr create`
- [BUG] No repositories found in Claude Code mobile repo picker
- Desktop + WSL: composer branch chip shows '—' and worktree toggle silently creates no worktree (session runs in main checkout)
- Markdown links in responses are not clickable in terminal (Ghostty) — need raw URL fallback or OSC 8 hyperlinks
- [Bug] Fable incorrectly flags valid requests as security breaches, auto-fallback to Opus
- [BUG] MCPB: "sensitive" user_config injected as __encrypted__: ciphertext instead of decrypted plaintext (Windows)
- [BUG] Folder picker won't select a subfolder directly
- [BUG] "New session" returns to an existing session instead of creating a new one
- [Feature Request] Add support for F# Fable 5 zero-knowledge proof verification
- Agent overrides explicit, confirmed user instructions on low-stakes actions — and in one case the reason it gave was not the real one
- [Bug] Frequent API errors with small task agentic workflows using Fable 5
- [BUG] Repositories not showing in Claude Code iOS app despite working on web — requesting backend index reset
- [FEATURE] Documented interactive-vs-headless signal in the SessionStart hook input (so hooks can self-gate for claude -p / nested sessions)
- Windows: app update fails with 'application already in use' because session background processes survive app exit
- [Bug] Config change to JSON not completing
- [Bug] AskUserQuestion tool unresponsive to mouse and keyboard input in terminal UI
- [Bug] Intermittent 401 Authentication Failed on Bridge POST After Session Creation
- [Bug] Fable safeguards incorrectly flagging routine coding context memory operations
Resolved issues (93)
- [BUG] Blender MCP Install Bug
- [BUG] API ERROR
- [Bug] CX Questions menu crashes with undefined property read error --- Session context remaining [99%] ---
- [Bug] Clipboard image paste fails with BMP format from Windows Snipping Tool in WSLg
- [Bug] Model automatically downgraded from Claude Fable 5 to Claude 3 Opus 4.8 due to pet food content analysis
- CLAUDE_PLUGIN_ROOT path separators stripped on Windows-client + WSL-mounted project (semgrep@2.1.1 MCP fails 100%)
- [BUG] Fable 5 → Opus 4.8 fallback is sticky for the rest of the session; benign planning/orchestration turns never return to Fable (manual /model is the only escape)
- [Bug] Rate limit error reported despite usage below 50% threshold
- [Bug] Malformed tool-call generation at high context window utilization
- [Bug] Excessive false positives in content policy filtering for international customs/trade queries
- LSP tool refuses to run a working, PATH-resolved clangd on Windows ("unsafe location (current directory)")
- [Bug] Model downgrade to Opus 4.8 triggered by guardrails during Fable usage
- [BUG]
- [BUG] MSIX variant again
- [BUG] 搞后门？claude code 你妈死了， bye
- [BUG][platform: claude.ai web] Post-compaction tool manifest rebuilt from last tool_search load only — native tools (incl. tool_search itself) dropped for rest of session (recurring, 5 threads)
- [BUG] Billing: $120 account credit disappeared during checkout on July 1, 2026 — support unresponsive for 13 days (EU customer)
- [Bug] Anthropic API Error: Safety Filter Flagged Request - No Cybersecurity Handler
- [Bug] Inefficient token usage in agentic loops with repeated failed operations
- [Bug] Unclear behavior after enabling pxpipe
- [Bug] Claude 3.5 Sonnet safeguards blocking legitimate code generation tasks
- [Bug] Safeguards preventing legitimate codebase modifications
- [Bug] Context awareness limited to current user in note generation
- [BUG] Browser pane screenshot always times out (30s) when pane is not displayed — visibilityState "hidden", zero compositor frames (2.1.174; re-file of stale-closed #53475)
- [BUG] The browser pane `computer { action: "screenshot" }` tool consistently times out after 30s with:
- I don't see a bug report in your message - only a file path to a screenshot. Could you please share the actual bug report details, error message, or description of the issue you're experiencing with Claude Code? Once you provide that information, I'll gen
- [Bug] Local compute machine flagged as cybersecurity threat during model training
- [Bug] Slack integration setup failing for subagent execution
- /deep-research and /workflows commands unusable with official Claude auth, but work via third-party API (e.g. Kimi)
- [Bug] Safety guardrails prevent Fable 5 model functionality
- [Bug] False positive security issue flagging
- [Bug] Empty bug report submission
- [DOCS] `claude -p` slow / appearing to hang on Windows during the slash-command and skill scan is not mentioned in troubleshooting or the headless guide (regression v2.1.161–v2.1.168, fixed in v2.1.169)
- Support custom file-opening actions in keybindings (e.g., open settings.json)
- [FEATURE] Support user-level .agents/skills/ discovery for cross-agent single-source-of-truth workflows
- [Feature Request] Allow /feedback command without including private session data!
- Model acknowledges pattern error explicitly then immediately reverts to same pattern
- [BUG] Agent tool isolation: "worktree" has no effect for team agents — agent runs in main repo
- [BUG]
- ultraplan: Remote session create failed with 403 on Max subscription
- I don't have enough information to generate a meaningful GitHub issue title from this message. This appears to be a status update rather than a bug report or feature request. To create an appropriate issue, I would need: - What problem were you experienc
- Feature request: Native session export to Markdown (/export command)
- [FEATURE] Save conversation to local file (markdown/JSON export)
- Critical data loss: startup GC silently deletes all session transcripts
- [FEATURE] Secret scrubbing and rotation for session logs (~/.claude/projects/*.jsonl)
- Feature request: centralized 'all my Claude Code sessions' view across surfaces and devices
- [Feature Request] Session durability, identity, and attestation — user-owned, not platform-managed
- [BUG]
- v2.1.128 update silently destroys conversation history for pre-JSONL sessions
- [Bug] Image decode fails on GIF: "Failed to decode image: The image format Gif is not supported"
- [Bug] Anthropic API Error: Usage Policy violation blocks entire session after single request
- ultrareview crashed before producing findings — free credit consumed
- [BUG] AUP classifier false-positive (session poisoning) persists through 2.1.167 — benign git command blocked; all work impossible. Regression spans 2.1.165–2.1.167. Cluster: #65846 #61185
- [BUG] Can't submit prompt without stopping on mobile browser Claude code
- [JetBrains] Regression: clicking file highlight in chat opens system editor instead of IDE (broke ~v2.1.160)
- Claude Code repeatedly rediscovers solved problems
- [Bug] @ context file syntax not working in Claude Code agent
- AskUserQuestion dialog covers the content above it — add a collapse/minimize toggle
- VS Code panel lacks status line + terminal lacks image paste in Codespaces, forcing a tradeoff
- [FEATURE] Add intent comments inside generated scripts before execution
- Sonnet 4.6
- [BUG] Rewind broken in Claude Desktop 1.11187.4 — "rewind target not on CLI's active chain"
- Feature Request: Streaming text output in CLI (typewriter effect like Claude Desktop)
- Plugin-bundled MCP server is double-registered in Cowork; interactive MCP-UI ("app") tools only render via the connector-UUID registration, not the mcp__plugin_… one
- [BUG] Claude's diff editor in VSCode silently discarded my edits to the file
- [FEATURE] [Enhancement] Allow @claude bot to accept repository collaborator invitations
- Regression: background shells killed ~5s after result breaks long-running background commands
- [BUG] enter or shift + enter both submit the query
- [FEATURE] show user reject mesage to an edit front and center
- Allow per-subagent exclusion of CLAUDE.md / .claude/rules (parity with built-in Explore/Plan)
- Feature: Allow disabling specific built-in tools to reduce context overhead
- Add active_skill field to PreToolUse hook input (enable skill-scoped permission rules)
- [BUG] Context window indicator shows 94% on fresh session — breakdown sums to ~17%
- [Bug] False-positive cyber-related safeguards block on legitimate low-level storage debugging (SPDK / NVMe-oF vfio-user, read-only gdb)
- [BUG] /cost command causes session to freeze on v2.1.158
- Add explicit toggle to disable session auto-archive
- [Feature Request] Unified Plan pane aggregating Plan Mode, TaskList, spec documents, and chat plans
- [BUG] Cowork: Code block copy button not working in a new window (macOS)
- (withdrawn)
- [BUG] Thinking blocks not expandable/collapsible in VS Code extension (renders as inert stub)
- [BUG] Desktop app WorktreePool maintenance sweep detached the MAIN repository's HEAD; worktree removal is non-atomic
- [BUG] Desktop app (macOS): persisted local sessions load in main process but never render in sidebar / Cmd+K after relaunch
- [BUG] Auto-mode classifier permanently pinned to unavailable opus-4-8 pool; account missing Sonnet-5 classifier config (tengu_auto_mode_config.model absent)
- Plugin registry creates duplicate project-scope installs when project path drive-letter casing differs (Windows)
- Teleporting a session from Web → Local silently saves 0 messages, then resume fails with "No conversation found"
- [BUG] 2.1.211 on Windows Terminal: TUI renders duplicated/overlapping frames, cursor lands mid-line, pasted text mid-word (2.1.210 is clean)
- [BUG] HTTP 407 error for pretooluse http hook with authentication in sandbox
- [BUG] Extension fails to activate: "TypeError: Invalid value for env" (command 'claude-vscode.editor.openLast' not found)
- SDK-spawned subagent sessions are persisted as independent top-level sessions (flood the resume picker)
- [BUG] Agents view: reopening a stopped background agent dispatches a blank `mode: prompt` worker — same session ID, intact transcript never loaded, original prompt re-injected
- [Bug] Anthropic API: claude-opus-4-8 returns empty thinking block text with valid signatures
- Scheduled cloud routine firings refused as "NOT USER INPUT" since ~13 July 2026
- Fable 5 forced onto usage credits on Max plan before July 20 cutover, despite reset allowance at 0% Fable usage