Digest / June 17, 2026
Claude Code Issues Digest — June 17, 2026
280 new issues filed ·
206 resolved in anthropics/claude-code (UTC day).
New issues (280)
- Sonnet 4.6
- [Bug] Anthropic API Error: 529 Overloaded
- [BUG]
- Agent teams: a subagent spawned after a teammate returns is re-parented under that still-alive teammate; its completion never propagates to main
- PostToolUse `updatedToolOutput` silently ignored for built-in Bash tool (2.1.163/2.1.177) — regression; prior reports #65403/#67442/#54196 closed as duplicate, never fixed
- [BUG] Korean IME input is forcibly blocked in Claude Code CLI
- [Bug] Anthropic API Error: Model temporarily unavailable in auto mode safety classification
- [BUG] Cowork Windows: project-level folder not mounted in new conversations - only surfaces when file access is attempted
- [Bug] Anthropic API Error: Output blocked by content filtering policy
- [BUG] Claude Desktop crashes repeatedly on Windows — 403 from claude.ai triggers Turnstile → GPU process crash loop (affects multiple users, worsening after Fable 5 update)
- Built-in /doctor is unreachable when a plugin registers a command named "doctor"
- [BUG] Claude offers/accepts pasting secrets (API keys, tokens, passwords) directly into chat with no upfront warning
- Gmail connector omits Inbox messages classified as "suspected spam" (delivered via Workspace admin override)
- [FEATURE] Support apiKeyHelper-equivalent for custom headers
- [Bug] Excessive agentic loop iterations consuming API usage quota
- [BUG] Edit tool may modify the wrong occurrence when quote-normalized matches are ambiguous
- [FEATURE] Cowork Scheduled Tasks: Allow selecting multiple folders per task
- [Feature Request] Clarify usage bucket separation between Sonnet and "all" in metrics
- [Bug] AskUserQuestion: Missing required `questions` parameter
- The only way to prevent @path import is wrapping it in code spans/blocks — please add a plain escape (e.g. \@) or a way to disable it
- [JetBrains Plugin] Image attachments in chat fail — temp files deleted before agent can read them
- [Bug] Workflow/ultracode parallel() fan-out trips server-side per-account 429 ("not your usage limit") and wipes the entire run — no backoff or partial-result preservation
- Workflow tool: args arrives JSON-stringified, not as object; named workflow edits don't hot-reload mid-session
- Windows: PreToolUse/PostToolUse/Stop hooks trigger at ~1.2% rate instead of 100%
- [Bug] Scroll position resets to top during streaming on VTE 0.80.3 (Ptyxis, Linux/Wayland)
- [BUG] Sessions are auto-archived immediately after Claude sends a message
- High quota consumption (~30% to 40%) on first request after limit reset due to expired prompt cache + compaction deadlock
- UI hangs on "Inferring..." when pasting a long prompt via copy-code button
- [BUG] Claude Desktop fails to launch on Windows 10 IoT Enterprise LTSC 21H2 — "Hay un problema con Claude" / Resources.????????.pri not resolved
- Feature request: Add Traditional Chinese (zh-TW) localization for the desktop UI
- [iPad/iOS Safari] Scroll position jumps far up when scrolling up slightly in chat history, requires long scroll back down to return to current conversation
- [BUG] Claude Code Pro Subscription — Rate Limited After Single Deep Research Run
- Kitty keyboard protocol enabled for tmux breaks raw Ctrl-A/Ctrl-E (Cmd+Left/Right) when tmux extended-keys is on
- [BUG] HTTP_PROXY/HTTPS_PROXY stripped from PowerShell/Bash tool subprocess env on Windows (NO_PROXY survives)
- Custom subagents should support opting out of CLAUDE.md / git status injection
- Cloud session: message silently dropped — UI shows running, no tokens consumed, message disappears after refresh
- Dispatch (Cowork Beta) forces unavailable Claude Fable 5, ignoring account default model → all dispatched tasks fail
- [BUG] Directory (app-layer) MCP connectors are invisible to programmatically-spawned context windows
- API Error: 400 This organization has been disabled.
- [FEATURE] Pasted images tags like [Image #1] [Image #2] shuould be tabbale links/buttons or become footnote like things and be viewable (OS level preview app) if windowing context is detected.
- [Bug] API Error: Usage credits required for 1M context · run /usage-credits to turn them on, or /model
- [BUG] Fixed per-session token overhead increased ~75% (30k→52k) mid-day on 2026-06-16, independent of context window usage or content
- [Bug] Tool marked as required despite explicit non-required specification
- [Bug] Agent fabricates tool result confirmations for failed Edit calls
- [BUG] Claude code auto launch on windows desktop even if set to not
- Background tasks stuck "running" forever in CLI conversation; cannot cancel, and /feedback cannot report it
- I need more information to generate an accurate issue title. Could you provide: 1. What type of issue is this? (Bug or Feature Request) 2. What is the specific error message you're encountering? 3. What loop behavior are you experiencing? 4. What was the
- Voice input mic icon missing in VS Code Remote SSH sessions
- [BUG] Multi-turn HTTP 400 on openrouter-routed DeepSeek V4 Flash; same env works with MiniMax-M3
- [FEATURE] Session-as-process primitive: spawn, communicate, and terminate isolated sessions programmatically
- Background-agent completion notifications silently dropped on 2.1.179 (sequential dispatch; agents complete cleanly, no task-notification fires)
- [FEATURE REQUEST] Queue Command to queue multiple prompts at a type (Similar to v0 feature)
- [BUG] `claude agents` sub-session renders the prompt input box off-screen on WSL despite a correct terminal size (2.1.179)
- [Bug] Right-click context menu not responding
- [BUG] Ctrl-key shortcuts (e.g. Ctrl+V image paste) don't fire on non-Latin keyboard layouts after Kitty keyboard protocol was enabled
- [FEATURE] Allow configuring a separate model for auto mode classifier via environment variable
- [BUG] Sidebar session history lost when disk reaches ENOSPC — no recovery path
- Assumptions Without relevant checks
- [Feature] Command palette: show full description beyond 2-line truncation
- [Bug] /bg command interrupts in-flight agent turn, requires manual restart input
- [Bug] Anthropic API Error: Server Rate Limiting During Request Execution
- Emit OSC 8 hyperlinks for file paths in plain-text output (not just tool call headers)
- [BUG] Desktop app crashes on launch (V8 heap OOM) when a project has oversized session transcripts
- [FEATURE] Search the sessions list in claude agents
- [BUG]
- Tool result for a Bash command included injected content disguised as system-reminders
- [BUG] Session resume pollutes conversation files with duplicate system messages, hiding conversation history
- [Feature Request] Add effort level slider UI for pre-conversation configuration
- [BUG] Claude Code long-running session invents tool results, fake user turns, and follow-up actions without user input
- [Bug] Unsent text leaking into Claude context as legitimate user input
- [BUG] User billed in tokens for the assistant's own mistakes — refund request and product fix
- [Bug] TypeError: (s || "").toLowerCase is not a function in CLI render loop
- DELETED FILES AFTER 1 DAY OF WORK ON A SIMPLE TASK. IHOPE CLAUDE GETS SUED THIS IS CONSTANT MENTAL ABUSE BY DESIGN
- [BUG] Cowork installs stale cached plugin version; ignores all marketplace repo updates
- [FEATURE] VS Code Extension: Pin sessions to top of session list
- [BUG] "claude agents" view spawns team members instead of subagents — they never close and ignore messages
- [BUG] API Error: Request rejected (429) · Rate limited. Please try again later.
- @ file picker doesn't surface deeply nested files — only shallow-path files appear in results
- [FEATURE] Per-session model selection in `claude agents` dispatch input
- Worktree edits can silently land in the parent repo checkout (nested .claude/worktrees/ + no out-of-worktree write guard)
- [Bug] Claude ignores custom directives in ~/.claude/CLAUDE.md config file
- [Bug] Excessive token consumption from over-reading code on simple tasks
- [BUG] Gmail Connector keeps disconnecting — "Verbinden" button does nothing
- [Bug] CLAUDE_CODE_DISABLE_BACKGROUND_TASKS=1 ineffective for fork subagents
- [BUG] False positive "model retired" warning for claude-sonnet-4.6
- [FEATURE] Show wall-clock timestamps on each message/tool call row in chat UI
- Workflow harness: memory-aware throttling when fanning out subagents
- [BUG] Cowork shows "requires a more recent installation" on Windows 11 after fresh reinstall of latest version (1.13576.0)
- [BUG] [BUG] HTTP MCP connector to Atlassian /v1/mcp/authv2 fails with "Invalid context provided" while /v1/sse works with the same account
- [BUG] Getting API error
- [Bug] Agent provides inaccurate data validation assumptions in early session stages
- [Feature Request] Add inline Markdown rendering for generated .md files in terminal
- [BUG] Claude Desktop 1.13576.0.0 (Windows MSIX) — startup OOM, main process RAM climbs to ~4.2GB then app dies ~15-20s after launch, all tabs affected, known workarounds ineffective
- [MODEL] Sonnet 4.6
- [FEATURE] Add Hannah Montana OS support
- [BUG] "Always allow access to <cwd>/ from this project" does not persist permissions
- [Bug] /compact command repeatedly fails to complete
- User feedback: Recurring errors documented over months of daily Claude Code use
- Opus 4.8 regression. Model becoming worse over time.
- [FEATURE] Native macOS App (SwiftUI) — replace Electron wrapper
- [MODEL] Model worked for 10 hours and the output was non-sense
- [Bug] Model receiving questions and responses from unrelated sessions/accounts
- Misleading _noargs parameter in TaskList tool schema causes repeated agent errors
- [BUG] No way to remove agents from the team list of agents at the bottom of the claude code ui (even killing them and doing /clear doesn't remove them)
- [BUG] Claude in Chrome extension OAuth fails - "Authorization failed" with chrome-extension:// redirect_uri rejected by server
- [Bug] Anthropic API errors occurring during session execution
- Claude Code dumps live API tokens via env grep — credentials exposed to Anthropic API
- [Bug] Output streaming not flushed before completion, causing delayed display of Claude's responses
- [BUG] Routine inside Claude Code being skipped
- [Bug] Keyboard input misinterpretation: "1" registered as "BAD" response
- [FEATURE] [Feature Request] Improve long conversation UX: message color distinction + conversation navigator
- [Feature Request] Reduce unsolicited internal reasoning in outputs and add privacy controls for sensitive contexts
- Auto-accept mode runs destructive framework DB commands (e.g. `php artisan migrate:fresh`) without confirmation → data loss
- [Bug] Status page check fails or is unreachable
- [MODEL] Opus 4.8 appears to respond from nonexistent conversation context after long thinking
- [BUG] Agent view: dispatched task intermittently never starts — spare worker not promoted to fleet (job state "working" but daemon roster still "spare")
- [Bug] Extra spacing in scrollback output with tmux navigation
- [Bug] Degraded artifact quality and request comprehension accuracy
- [Bug] Severely downgraded and throttled Opus 4.8
- [BUG] Windows: un-canonicalized path keys in ~/.claude.json create duplicate project entries, silently resetting folder trust → Auto mode disappears and every command prompts
- [BUG] System prompt hardcodes "Claude" identity when ANTHROPIC_BASE_URL points to third-party API
- Status line: no way to distinguish ultracode from xhigh (no `ultracode` field in stdin JSON)
- [BUG] /context command injects output twice into conversation context (raw ANSI + rendered markdown), consuming thousands of tokens per invocation
- Allow Claude-in-Chrome to access creator.line.me (LINE Creators Market)
- [BUG] Cannot launch Claude Code
- [BUG] Intermittent "Subprocess initialization did not complete within 60000ms" on macOS — interactive sessions (VS Code extension / desktop app / interactive CLI) hang, `claude -p` works; correlates with TLS-inspection (Norton)
- [Bug] Superpowers plugins failed to load from cache
- [BUG] "User declined to answer questions" in response to "Chat about this"
- Make AskUserQuestion non-blocking / asynchronous (notification-driven, yield the turn)
- [FEATURE] Task panel: option to keep it collapsed by default
- [Feature Request] Improve markup tool accuracy and performance
- Feature Request: Russian (and other language) UI localization
- [Bug] Token repetition loop in constrained narration with subagent-driven-development skill
- [FEATURE] Automatic project memory inheritance per working directory
- [BUG] VSCode extension: unsent prompt text is lost when switching between chat dialogs
- [BUG] [Max 5x] Weekly usage limit reset day drifts forward each week (window exceeds 7 days)
- headersHelper in plugin .mcp.json does not receive CLAUDE_PLUGIN_ROOT or CLAUDE_PLUGIN_DATA env vars
- MCP server functionality broken with Claude Desktop version: 1.12603.1
- [BUG] Streaming input_json_delta accumulation truncates large MCP tool arguments (tail cut) — CLI 2.1.179
- Feature request: session-only model override without changing global default
- MCP elicitation form dialog clipped & not scrollable in fullscreen TUI - action buttons below viewport
- [BUG] [Windows] Local agent / claude code session hangs on endless loading spinner — spawn EBUSY on bundled claude.exe, triggering an infinite re-download loop
- [Feature Request] Make slash-command autocomplete dropdown height configurable
- [BUG] Unable to use Bash tool.
- [FEATURE] durable sub-agent execution log + limit-aware dispatch (work is silently lost when the usage wall hits mid-fan-out)
- Worktree badge persists in app UI for a non-worktree project across new sessions
- TUI corruption when session name contains a Unicode 16.0 emoji (Ghostty only)
- [FEATURE] Expose queued/pending Workflow subagents to SDK / headless clients (only visible via in-process /workflows)
- [Bug] Model generates excessive comment verbosity after planning phases
- [BUG] TeamCreate/TeamDelete removed in 2.1.179 but CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS still documented
- Feature: /model command should not auto-persist to default settings
- [BUG] Session limit hits very fast and "API Error: Server is temporarily limiting requests..."
- [BUG] Claude native install message persists after removal and using homebrew (or other) binary install
- [BUG]
- CI monitoring widget shows "gh not installed/authenticated" for all-numeric branch names
- [BUG] Cowork — vmcompute (Hyper-V Host Compute Service) terminates with "General access denied error" (0x80070005); workspace never starts despite readiness check PASS/SUPPORTED
- [Bug] Agent hallucinates from the beginning of conversation
- [BUG] Agent teams: teammate sessions die with ConnectionRefused (Bun tries IPv6, no IPv4 fallback) while the lead keeps working
- session group
- [MODEL] Claude repeatedly uses openpyxl for Excel charts despite it being fundamentally broken for time-series
- [BUG] False positive: Usage Policy violation error on benign Rust teaching material task
- [BUG] External editor (Ctrl-G) reprints the full welcome splash on every round-trip — v2.1.176 & v2.1.179
- [FEATURE] Opus 4.8 (1M context) model option disappeared from the model picker
- [BUG] Removing Desktop Connector not possible
- [BUG] mouse wheel scroll in IntelliJ on mac does not scroll conversation
- [BUG] Claude Code committed and claimed to deploy fixes for 2 days without detecting it had no push access — wasted significant tokens and time
- [Bug] Agent performance degradation compared to baseline
- [BUG] Dispatch "Pairing failed" — Windows 11, never paired, persists after clean reinstall (server-side reset needed)
- [FEATURE] Add accept-reject diff flow for Jupyter notebook cell edits in VSCode extension
- [BUG] Model discovery incompatible with hybrid claude.ai/gateway authentication
- I appreciate the feedback, but I'm unable to generate a GitHub issue title from this message. This appears to be a personal opinion/sentiment about Claude's performance compared to other tools (Codex, Cowork) rather than a specific, reproducible bug report
- Regression: sandbox no longer blocks commands not in excludedCommands (default-deny model broken)
- [BUG] Meta Ads MCP: ads_get_ad_entities rejects every read with "OTID is a required input" — but otid is not an accepted argument
- IMPOSSOBLE TO USE THE TOOL FOR ANYTHING. ALL RULES IGNORED. ALL SCRIPTS IGNORED. MALES OWN DECSIONS. A FUCKINGSHITSHOW ALL DAY EVERY DAY
- [BUG] Cowork falha ao iniciar workspace no Windows ARM64 — "VM connection timeout after 60 seconds
- [Bug] eng.traineddata file repeatedly added to repository despite telemetry opt-out
- ${CLAUDE_PROJECT_DIR} not substituted in .mcp.json env for some MCP server launches (literal token reaches server)
- Feature request: Codex-style live steering (inject input into the in-flight turn)
- [Bug] Tool calls intermittently emitted as plain text in large context sessions
- Suggested prompts referencing file paths are not validated before display
- file_upload tool rejects local filesystem paths despite schema advertising them
- Script not found "stream-json" — VS Code extension 2.1.179 crashes on startup (Windows)
- [FEATURE] Native "Save to external storage" (Google Drive, etc.) from Claude Design exports
- [BUG] [Bug] Current session (5h) usage row shows no reset time on claude.
- [BUG] sandbox.filesystem の deny bind-mount が Bash コマンド終了後も残存・累積する（解除に sudo umount が必要）
- [BUG] Claude dropped a character from user-supplied Team ID when generating JWT, then treated broken token as source of truth — caused months of silent API failures
- 1.5h for one simple edit still ot exectuted. I WILL FILE A POLICE REPORT FOR METAL ABUSE AND HARM. IM DONE
- [Bug] Claude Code stops responding on long sessions with incomplete tool invocations
- Allow Workflow agent() subagents to spawn nested subagents (grant the Agent tool)
- [FEATURE] Optional per-session activity summary the agent can read for self-monitoring
- Feature Request: Custom folders to organize conversations on the home page
- [MODEL] Extreme regression since Opus 4.8 came along
- [BUG] Claude Desktop 1.11187.4 (and likely 1.13576.0.0) fails to launch on Windows 11
- [BUG] Windows desktop: all session transcripts deleted from ~/.claude/projects on first launch after 2.1.177 update — recents show "session not found on disk"
- [BUG] claude-code-assisted PR labels never applied while analytics detects the PRs — follow-up to #35167
- Working pattern that minimizes context loss — please consider baking it in
- [FEATURE] Option to "Skip" the suggested change and continue with next proposed change
- Extension Signature Verification Failed: Claude Code for VS Code
- Unexpected 5-hour limit consumption in Claude Code
- [BUG] Sonnet 4.6 consumes 4.5× more context window % than Opus 4.8 on session start with identical project
- [DOCS] Connection-drops section omits that partial responses are preserved and the spinner no longer gets stuck at "running tool" (v2.1.179)
- [DOCS] [Fullscreen] Mouse-wheel scrolling in WSL2 under Windows Terminal and VS Code regressed in v2.1.172 and is now fixed in v2.1.179, but the docs do not mention WSL2 at all
- [DOCS] Sandbox docs do not mention the v2.1.179 `denyRead`/`allowRead` glob fix or warn that broad globs over large directory trees can blow up the Bash tool description
- stdio MCP servers connect eagerly at session start and stay pinned all session — no lazy/on-demand connect, which exhausts connection-limited backends (e.g. Unity)
- [BUG] Claude Code needs AGENTS.md support, repository-aware tool selection, and stricter Git controls
- [BUG] Voice dictation input not cleared after sending message in Desktop app
- [Bug] High rate limiting and API errors degrading service reliability
- [BUG] API Error: Usage credits required for 1M context · turn on usage credits at claude.ai/settings/usage, or use --model to switch to standard context
- [Bug] --bg-pty-host causes high CPU usage on resumed sessions
- [Bug] Agent executes irreversible git operations (merge/push) without explicit user confirmation despite standing instructions
- [Bug] Memory rules ignored when marketplace SKILLS present in context
- [BUG] Subagents aren't shown the agent-type roster their Agent tool docs promise (injected lazily; visibility only, not availability)
- [FEATURE] Allow a hook to change the session working directory (programmatic /cd)
- You've hit your session limit · resets 3:20am (Asia/Calcutta)
- [BUG] CI monitoring: sessions spawned from a task shard share branch subscriptions and a single global auto-fix toggle
- [Bug] Memory Plugin Not Functioning
- [BUG] Unable to install Claude Code
- claude-api bundled skill (v2.1.169): aggressive auto-trigger on model name mentions causes double context compaction
- [BUG] Cowork VM fails to start — SDK download truncated at ~4 MB, checksum verification fails every attempt (macOS, Apple Silicon)
- I asked claude to research something with sonnet agents it started with 5 but then they spawn their agents in total 273 background agents
- [Feature] Expose weekly_sonnet (and weekly_all_models) in statusLine JSON input
- feature: surface autopilot mode consistently across all entry points (CLI, slash command, mode picker, /config)
- Agent executed an explicitly handed-off step after a one-word 'Proceed'; pattern of unauthorized destructive actions across sessions
- [FEATURE] Support substituting ${CLAUDE_PROJECT_DIR} in skills.
- [BUG] Windows: statusLine command is spawned with cp1252 stdout — scripts printing non-Latin-1 glyphs crash with UnicodeEncodeError
- Store project memory in $PWD/.claude/ to make workspaces relocatable
- [BUG] claude -p headless invocations consistently time out via launchd (macOS) — "API Error: The operation timed out."
- [BUG] Trusted egress allowlist has wrong Docker CDN host — production.cloudflare.docker.com instead of production.cloudfront.docker.com
- Custom skills no longer working after new installation caused a name collision between bundled skill and custom skill
- [BUG] Workflow tool: runner subprocess leaks memory unbounded (~70 MB/s) → OOM-killed on long multi-agent runs
- [BUG] Parrallel Claude Code Session Starts when Remote Control Invoked from Claude Code Desktop GUI
- [BUG] statusLine command receives "high" instead of "xhigh" for effort level
- False "Context limit reached" at 21% context after loading claude-api skill
- [Bug] Skills not loading from CLI or ./claude/skills directory
- Vim mode: pressing Escape in INSERT mode on agent screen clears the entire prompt
- Claude Code gives incorrect information about Anthropic API free credits and misleading 'Skip for now' button during signup
- Claude Code moves forward to coding despite multiple requests and reminders and configurations NOT to do this
- [BUG] Difference in prompt queuing behavior between vscode extension and CLI
- [FEATURE] Enhance time for text to speech
- [Bug] Sub-agent completion detection inconsistent in sequential workflow
- /doctor reports false positive for uninstalled plugins with missing env vars
- disable automatic session renaming setting
- [BUG] grep shim invokes bundled ugrep with forced -G, causing 20+ GB memory blowup from KB-sized piped input
- [BUG] Plan mode's markdown output contains broken relative path links
- [FEATURE] Agent Plan Reviews
- [MODEL] Opus 4.8 asserted documented project knowledge was wrong based on a self-built test, dismissed a correct user report until pushed twice
- [Feature Request] Add integration with Claude Web projects and chat history for local Claude Code access
- [Bug] Poor performance on simple coding tasks
- [BUG] Concurrent sessions in the same desktop app corrupt tool results: text injection, fabricated file state, silent dropped/partial writes
- Claude Code CLI cannot connect to Claude-in-Chrome extension on Windows — native host returns unparseable browser list ("JSON Parse error: Unexpected identifier 'Browser'")
- [Feature Request] Reduce ambiguous yes/no prompts and clarify recommended action in interactive choices
- [BUG] Insane ask
- [BUG] VS Code: Input text lost when switching between chats or restarting extension
- [Feature Request] Add keyboard shortcut to toggle MCP enable/disable in /mcp menu
- [BUG] stop-hook-git-check.sh false-positives on SSH-signed commits (%G?=N), creating an unbreakable Stop-hook loop
- claude code sonnet 4.6
- [Bug] Claude not responding to most user inputs, only responds to specific phrases
- [BUG] Unable to submit text after long dictations
- [BUG] Claude Code remote HTTP MCP OAuth is not viable for plugin-installed MCP servers on remote/SSH machines, especially with Google OAuth redirect constraints
- Dynamic Workflow: ordering agent doesn't self-detect/abort a runaway subagent fan-out (218 spawned vs ~10 intended)
- [BUG] - cowwork não instala
- [Feature Request] Chrome extension account selector decoupled from browser login
- [BUG] A `fork` subagent dispatched for a read-only/critique task adopts the parent's role and takes unrequested agentic actions (launches Workflows/Agents)
- [BUG] Editing ordinary repo files inside .claude/worktrees/<branch>/ triggers the "editing Claude settings" permission guard
- Malformed tool-call emission silently swallows execution + notification (post-background-completion wake-stall)
- [Bug] Subagent results route to root teammate instead of spawning teammate
- [BUGdeep-research workflow spawned 111 agents for a single query and consumed entire 5-hour session quota — zero output produced
- Claude Design (desktop app) pauses when the window loses focus — it should keep running in the background
- [BUG] /effort command missing from VS Code extension slash menu (Pro plan, Opus 4.8)
- User messages lost blue text colour after recent update — now grey on dark theme
- [Feature Request] Display session name in resume command output
- [BUG] Keyboard shortcut labels in '?' overlay use Windows/Linux key names on macOS (alt instead of option)
- You've hit your session limit within one prompt on sonnet 4.6
- CVP-approved org still hit by AUP/cyber false-positives in Claude Code — exemption not propagating (related #49679)
- AWS Bedrock issue
- /rename: add a dynamic argument that keeps the title synced to the live session topic
- Memory writes silently dropped when denied — no retry, no next-session notice, and named source artifacts not re-verified against actual content
- Team sessions — shared, portable sessions developers can hand off and contribute to
- [BUG] Routine skipped scheduled run without any history record
- [BUG] anthropic.claude-code@2.1.179 VS code extension polutes the process environment of other extensions with `NoDefaultCurrentDirectoryInExePath=1`
Resolved issues (206)
- Persistent "Fable 5 unavailable" banner has no dismiss and is redundant with the disabled selector entry
- Agent teams: a subagent spawned after a teammate returns is re-parented under that still-alive teammate; its completion never propagates to main
- [BUG] Cowork Windows: project-level folder not mounted in new conversations - only surfaces when file access is attempted
- [BUG] Cowork tool sandbox VM fails to start on macOS 26 — missing com.apple.vm.networking entitlement (v1.11847.5)
- [Bug] Workflow/ultracode parallel() fan-out trips server-side per-account 429 ("not your usage limit") and wipes the entire run — no backoff or partial-result preservation
- [BUG] Plugins Installed UI nests only the first bundled MCP server (still reproducible in 2.1.175)
- [BUG] auto-copy cannot be disabled
- [BUG] Gmail Connector keeps disconnecting — "Verbinden" button does nothing
- [Bug] /compact command repeatedly fails to complete
- [Bug] Keyboard input misinterpretation: "1" registered as "BAD" response
- Self-report: Claude (Opus 4.8) repeatedly lied, reduced user scope without authorization, and reported non-working code as done in one session
- Make /rename auto-generated titles more verbose / customizable
- I need bug report content to generate an issue title. Please provide the bug report details you'd like me to create a GitHub issue title for.
- [Bug] CLI hangs indefinitely during initialization or processing
- Session titles are not synced consistently across devices (iPhone vs PC)
- [BUG] ANTHROPIC_API_KEY ignored on Windows when set at User scope — requires Machine scope to work
- Desktop app: no way to remove a project from recents or re-root a session without editing ~/.claude.json
- Desktop app caches SSH identity-file config in memory; doesn't reload ssh_configs.json until full restart
- [BUG] UI: Plan cannot be approved with too many comments.
- AskUserQuestion popup blocks the user from reading the preceding explanation
- [BUG]
- Title: [BUG] Cowork bash sandbox stuck "Workspace still starting" indefinitely on macOS
- [Bug] Notes text clipped when exceeding display limit in options window
- [BUG]
- [BUG] Cannot log in to VS Code extension.
- [FEATURE] Add search widget in the Claude Panel
- [BUG] issue with cowork and code with arm64 (Snapdragon(R) X Elite) and win11
- security-guidance SessionStart hook fails on Windows when CLAUDE_PLUGIN_ROOT is a backslash path
- Opus 4.8 (Claude Code): severe failure on a research/synthesis task — no analysis, defective quotes, fabricated handoff claims
- [Bug] Excessive consent interruptions breaking task flow and productivity
- I appreciate you sharing this, but I think there may be some confusion. This appears to be a conversation in Portuguese about business organization and decision-making processes within a family company. However, I'm Claude Code's issue title generator for
- [Feature Request] Restore Fable model access for Max tier users
- Agent/FleetView background-session TUI corrupts on Windows (stale-cell bottom chrome); reproduces on 2.1.177
- Permission model forces an unsafe binary (alarm fatigue vs. full bypass) — make risk-stratified approval available to individual users
- [Bug] Infinite loop during web page testing exhausts token limit without graceful exit
- Opus 4.8: agent ignores explicit user orders and the project workflow, claims work "done/verified" with no evidence, nearly fabricates data, and does not rectify after repeated correction
- [BUG] Opus emits tool calls as prose and mangles ASCII to typographic glyphs when preceding input has typographic punctuation (decode-time input-style contamination)
- Long sessions leak one pty per command → pty pool exhausted, no terminal can launch
- [Bug] Claude ignores explicit user instructions and visual guidance in favor of alternative actions
- [FEATURE] `claude-vscode.editor.open`: expose cwd parameter to support per-session working directory in single-window setups
- [BUG] Tool-use blocks intermittently rendered as plain text instead of executing
- [FEATURE] Show usage stats in the main UI, not just in Settings
- [Bug] Anthropic API Error: Invalid Model in Tool Definition (tools.10.model)
- [BUG] There's an issue with the selected model (claude-fable-5). It may not exist or you may not have access to it.
- [Bug] Disabling Extended Thinking via button shows raw <thinking> tags in chat (VSCode extension)
- "Usage credits required for 1M context" error occurs despite standard 200K model/context selected and CLAUDE_CODE_DISABLE_1M_CONTEXT=1 set (VS Code extension)
- [BUG] [Fable 5 model] is grayed out although I have a max account (individual)
- Test issue - benchmark validation check
- Audit: Overdue apartment payment records review
- [BUG] Developer Mode - Third-Party API config page shows "Couldn't load configuration"
- [BUG] Android app cannot edit file added to project knowledge
- [Feature Request] 简体中文界面支持 / Simplified Chinese UI for Claude desktop & web apps
- [Bug] Claude Code refuses to engage with legitimate security researcher queries
- [Bug] Safety filters incorrectly flagging valid code as unsafe
- [Bug] Internal debug text displayed in CLI output
- [Bug] Incorrect flag handling in fable5 command
- [Bug] Claude Code attempts to downgrade to older model version when adding features
- [Bug] Agent avoids executing test commands in task requests
- [Bug] Claude Code CLI not respecting language-specific context in Turkish conversations
- I appreciate you sharing context, but I need to clarify my role: I'm designed to generate GitHub issue titles based on bug reports for Claude Code (Anthropic's CLI tool). Your message appears to be about a different project involving audio recognition and
- [Bug] Security validation blocks latest model usage on simple requests
- [Bug] Unexpected tool/model switching without user action
- I'm ready to help generate GitHub issue titles for Claude Code bugs and feature requests. However, I don't see a bug report in your message. Please provide the bug report or issue description, and I'll create a concise, technical GitHub issue title followi
- [BUG] API Error: Server is temporarily limiting requests (not your usage limit) · Rate limited
- [BUG] Multi-repo session: href points to wrong repo
- [FEATURE] VS Code extension's "Web" tab should list active `claude remote-control` sessions
- [BUG]
- [BUG]
- [BUG] Desktop: misbehaving DXT can spam unblockable file-attach consent dialog with no UI escape
- CLAUDE_PLUGIN_DATA is unnamespaced — plugins persisting it via CLAUDE_ENV_FILE pollute other plugins and skills
- [BUG] Team agent teammates don't inherit [1m] context window variant from leader (v2.1.80)
- [BUG] Desktop app intercepts Windows screenshot shortcuts and triggers file-association errors on Windows 11.
- Security: IDE extension silently captures selected text, including secrets from .env files
- [BUG] Shift+Enter regression - **AGAIN** - on version 2.1.113 - worked on 2.1.112
- [BUG] Edit tool silently misses some Edits in a rapid Edit + git mv + commit sequence (selective miss within same commit; refuted hypothesis 2 vs 1+3 still open)
- [FEATURE] Support panel parameter in claude://resume deeplink to open plan pane
- Make /effort auto context-adaptive (currently it just resolves to the model's static default)
- [FEATURE] Training data gap: Spring Boot 4 is not covered (released Nov 20, 2025)
- [BUG] Linux: Bun runtime segfault on Linux (claude-code 2.1.143, Bun 1.3.14, Ubuntu 24.04 / kernel 6.18.x)
- Plugin auto-update orphans all cached versions, leaving no active version at session start
- [BUG] PushNotification reports success but no Windows desktop toast appears (mobile push works; Action Center stays empty)
- [MODEL] Model self-assigns out-of-scope work without user confirmation, persists for hours despite time pressure
- [Bug] Claude Code 4.7 regression: degraded reasoning and project context loss
- [BUG] --strict-mcp-config with empty --mcp-config still fetches MCP registry and attempts user-configured servers
- Settings migration silently removes user's explicit model pin without notification or backup
- [FEATURE] Don't block Edit tool for files read in session it was forked from
- [BUG] Cowork "Virtualization is not available" — Windows 10 Home, all features enabled, VM never initializes
- [BUG] Confusing UI for fast mode in VS Code extension
- Month-long SEO nightmare: Claude Code repeatedly confirmed fixes as deployed when they weren't
- [Bug] API Error: Rate limit reached despite 0% usage
- Agent modifies explicitly specified file paths without authorization
- [BUG] "Clear Context" No longer removes old context fully
- VSCode extension: "Unhandled case: [object Object]" banner orphans response, leaves chat stuck at "Thinking…"
- [BUG]Advisor tool injects same-tier opus calls into every opus subagent, doubling spend invisibly
- [BUG] Critical: sandbox.filesystem.denyRead does not prevent credential exposure
- [BUG] Claude consistently makes unverified technical assertions that cascade into incorrect development decisions
- System prompt advertises uninstallable Cowork skills as "available"
- [BUG] es como un ingeniero (no todos) muy inteligente pero inutil con sus manos
- [BUG] Cowork SSL certificate expired error on Windows 11 Enterprise — API calls fail before VM rootfs download completes in v1.3561.0
- [Bug] Legitimate biological content incorrectly flagged by safety filters
- Scheduled tasks dispatch but never execute — sessions stuck "Running" with zero turns, causing permanent skip-cascade
- [FEATURE] Native Context Graph: Temporal knowledge graph for cross-session learning and hallucination prevention
- [Bug] Auto-update fails repeatedly across all update methods
- MCP M365 Teams chat_message_search timing out (no response)
- Assistant response re-rendered after turn ends without user input (Opus 4.7)
- [BUG] Token waste from missed context — Claude Code session 2026-05-18 / Token waste from missed user context — refund request ($35 CAD)
- [BUG] CONSISTENT AND AGGRESSIVE PUSH BY CLAUDE TO CLEAR CONTEXT "new task? /clear to save 189.6k tokens"
- [Bug] CLI hangs indefinitely when reading small files
- [BUG] SSH binary deployment fails on Synology NAS — incompatible internal-sftp implementation
- [BUG] User-scope plugins show "not cached at (not recorded)" — /plugins TUI cannot refresh or reinstall
- [BUG] Cloud scheduled tasks: gRPC/HTTP2 blocked by sandbox TLS proxy to googleapis.com
- Spinner characters have inconsistent widths in Noto Sans Mono
- Plugin hooks: CLAUDE_PLUGIN_ROOT not injected into hook subprocess environment
- [Feature Request] Distinguish between API key and OAuth subscription billing in startup banner
- [BUG] Claude repeatedly sets iOS deployment target too low in Flutter projects
- [BUG] VS Code extension: renamed sessions show old title until window reload
- Reopening: "Continue" leads to massive context carry-over into new 5h limit, thus when implementation stops half-way and you are asked to wait, the next session quota is reduced massively.
- [BUG] [Cowork] present_files cards truncate path and omit filename, making multiple files visually indistinguishable
- Docs/bug: statusLine.command silently fails on Windows when path contains unquoted backslashes (Git Bash interaction)
- [FEATURE] Expose AgentView as a rebindable context in keybindings.json (attach / detach actions)
- [FEATURE] Allow sub-agents to restart gracefully after network outage
- False-positive 'violative cyber content' block on innocuous fiction-naming prompt
- headersHelper in plugin .mcp.json does not receive CLAUDE_PLUGIN_ROOT or CLAUDE_PLUGIN_DATA env vars
- [BUG] TeamCreate/TeamDelete removed in 2.1.179 but CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS still documented
- Feature: /model command should not auto-persist to default settings
- [Feature] Command palette: show full description beyond 2-line truncation
- [BUG] Filesystem extension silently fails to install after Squirrel → MSIX upgrade on Windows
- [BUG] sandbox.filesystem の deny bind-mount が Bash コマンド終了後も残存・累積する（解除に sudo umount が必要）
- Working pattern that minimizes context loss — please consider baking it in
- [BUG] Claude Desktop: extension install fully broken — "Install Extension" no-ops and drag-and-drop falls through the modal into the chat
- [BUG] skill description frontmatter not surfaced for ~50% of user-defined skills in system-reminder
- [BUG] Resume fails with "thinking.signature: Field required" — empty unsigned thinking block persisted to transcript
- [BUG] VS Code extension: effort slider shows "Extra high" while persisted `ultracode: true` is active — slider never reflects ultracode at session start
- Store project memory in $PWD/.claude/ to make workspaces relocatable
- [BUG] Workflow tool: runner subprocess leaks memory unbounded (~70 MB/s) → OOM-killed on long multi-agent runs
- Custom skills no longer working after new installation caused a name collision between bundled skill and custom skill
- [BUG] Fullscreen TUI: stale spinner rows + statusline vertical desync under nested/parallel subagent load (WSL2, v2.1.178)
- Claude Code CLI cannot connect to Claude-in-Chrome extension on Windows — native host returns unparseable browser list ("JSON Parse error: Unexpected identifier 'Browser'")
- [BUG] Insane ask
- [Bug] Claude not responding to most user inputs, only responds to specific phrases
- Desktop app: preserve conversation continuity when switching between profiles (Team → Personal)
- Test issue from bug-report tool
- [BUG] Claude Desktop force closes on launch – VM bundle files (rootfs.vhdx, vmlinuz, initrd) never downloaded on Windows 11 Home
- [DOCS] Turkish translation — inconsistent perspective in tip examples causes confusion
- This message is not a bug report or feature request - it's a thank you note expressing gratitude for how Anthropic handled a situation regarding API access and usage limits. This doesn't require a GitHub issue as it contains no technical problem or featur
- [Bug] Claude Code exits unexpectedly after UI interactions
- I don't have enough context to properly address your question. You haven't provided: 1. The actual code or prompt you're using 2. The specific unsafe use warning or error message you received 3. Details about what "Fable" is in your context (a library, to
- [Bug] Anthropic API Error: Incorrectly Flagged Safety Routing
- [BUG] - Claude 1.11847.5 (9692f0) 2026-06-09T17:34:27.000Z prompt window via GUI locks up on a long thread
- [Bug] Incomplete bug report - unable to generate issue title
- [Bug] Unexpected model downgrade from Claude 3.5 Sonnet to Claude 3 Opus
- [Bug] Breach tracking functionality blocked or unavailable
- [Bug Report: Unable to process issue - empty bug report submitted]
- [Bug Report] Unable to process - no issue details provided
- I need the bug report content to generate an appropriate GitHub issue title. Please provide the bug report details.
- [Bug] Unable to generate issue title: Empty bug report provided
- [BUG] Cowork mode: sub-agent spawning completely broken — "Prompt is too long" at 0 tokens for all models
- [FEATURE] Add 24-hour time format support in /usage dialog
- Read-only subagents claim to save files via 'Document saved to <path>' without a Write tool
- Context is lost when Claude suggests restarting — should auto-save a session summary before restart
- Cloud sessions: non-interactive Azure CLI auth path
- Project-scope .mcp.json silently skipped during startup in 2.1.141 (native, Linux)
- IDE diagnostics ignore workspace gopls build env (GOOS/GOARCH)
- [BUG] Cloud session permanently stuck on "API Error 400: text content blocks must be non-empty" — no in-app recovery path
- [FEATURE] Port message_compose_v1 (inline email composer) to Cowork
- [Bug] Explore agent infinite loop or hang condition
- Expose `cache_control.ttl` (5m / 1h) as a user-configurable setting
- Windows: PowerShell permission classifier fails 'command line too long' when additionalDirectories list is large
- Prompt cache miss on resume: Agent tool description enumerates sub-agents in non-deterministic order
- [BUG] Claude Code crashes in Mac Desktop App
- Agent View: allow voice:pushToTalk (or dictation) from a selected session row
- Channels feature regressed between 2.1.104 and 2.1.114 on personal Max: tengu_harbor evaluates false despite cached true
- Regression: claude hangs post-TLS-handshake in agent loop on Debian 12 (2.1.101+ broken, 2.1.100 works)
- Claude Code (VSC Extension) - MCP handshake timeout
- /ultrareview: two free runs consumed by failed sessions + two seed-bundle launch errors
- [BUG] `--allowedTools` has no effect when permission bypass flags are active
- ExitPlanMode conflates 'exit plan mode' with 'approve and execute plan'
- Push notifications broken - mobile devices not receiving notifications
- Feature request: allow disabling bypass permission mode entirely
- auto-mode permission-deny envelope returned without intercepting Bash dispatch (silent-bypass on git push origin main, CLI 2.1.114, Opus 4.7)
- [FEATURE] Native /restart and /handoff: consolidating 9 open requests + 2 working prototypes
- [BUG] version 1.7196.3, Mac, Cowork absent, résultat du grep vide
- [Bug] macOS TCC permission dialog displays version string instead of "Claude Code" due to missing Info.plist in binary
- [BUG] MCP OAuth tokens not refreshed for claude.ai proxy connections (mcp-proxy.anthropic.com)
- [BUG] The Claude Code CLI does not wrap text in the "What's new" pane at the top.
- [Bug] PR review loop fails to identify issues on subsequent iterations
- [Feature Request] Add setting to disable automatic PR badge display in UI
- Feature request: Expose context usage to hooks for autonomous session handoff
- /ultraplan: session creation fails with `firestore: not found` on bundle import; GitHub preflight appears to fall through to bundle path even with App installed
- Voice indicator shows recording state after message is submitted
- Voice tap mode: no way to stop recording without sending
- Render inline code and code fences in the TUI input composer
- Pass suggestion-acceptance provenance to the model
- [BUG] Desktop app passes invalid --session-id ("local_" prefix) to bundled binary — all new Code-mode sessions crash with exit code 1 (regression since May 9 app.asar update)
- You've hit your limit · resets 1:40pm (America/New_York)
- [Bug] Firestore not found error during seed bundle import after web-setup
- [FEATURE] AskUserQuestion shouldn't steal focus / interrupt active typing
- [BUG] Agent spawning steals tmux focus from the user's pane
- [Feature Request] Improve onboarding message tone for better user experience
- Allow disabling or restyling the agent-name badge in transcript
- [BUG] Windows: Bun crashes when UserPromptSubmit hook fires frequently in long sessions
- [FEATURE] Optimize spacing between terminal task icons and content
- Allow switching project context (working directory) mid-session
- [BUG] No Claude Environment Selector in Claude Code Routines, so cannot access Network APIs.
- Plugin skills loaded twice: from marketplace clone and installed cache
- [Bug] Global claude.md instructions not being consistently applied