Digest / June 13, 2026
Claude Code Issues Digest — June 13, 2026
224 new issues filed ·
326 resolved in anthropics/claude-code (UTC day).
New issues (224)
- [MODEL] Fable 5
- Startup blocks blank for 15s on "Grove settings" fetch (/api/oauth/account/settings) when network is slow
- I asked a question. I didn't see an answer: Any thoughts for a process and or tools for a better design?
- [Bug] Model downgraded to Opus during security checks blocking security vulnerability analysis
- Remote Control: built-in slash commands (/clear, /context) not intercepted — delivered to the model as plain text
- Statusline OSC 8 links with custom URI schemes (e.g. vscode://) aren't dispatched on click in `claude agents` / FleetView
- Feature request: `claude agents` should support repo-scoped filtering that includes worktrees
- [DOCS] New `footerLinksRegexes` setting from v2.1.176 is undocumented
- Scheduled tasks hang forever: Bash commands with shell expansion get no permission suggestions, can never auto-approve, and exhaust active-session limits
- [DOCS] Auto mode classifier Fable 5 fallback added in v2.1.176 is not described in `auto-mode-config.md`
- [DOCS] `claude daemon status` version-skew explanation added in v2.1.176 is not in `agent-view.md` or `cli-reference.md`
- [Bug] Claude Code Ultra review crashes with rate-limit error and consumes quota
- General-purpose sub-agents recursively spawn unbounded child agents, causing exponential fan-out and massive token burn
- [Bug] Anthropic API Error: Server Rate Limiting During Normal Usage
- Chat pane minimum width is fixed and cannot be resized
- [BUG] Desktop app (macOS): 'Previous session' shortcuts (Ctrl+Shift+Tab / Cmd+Shift+[) do nothing, while 'Next session' (Ctrl+Tab) works
- [Bug] Model unexpectedly switches from Claude 3.5 Fable to Opus during emulator execution
- [BUG] Cowork scheduled tasks silently dropping fires — June 2026 recurrence (references #55378)
- [BUG] "Restore to GitHub" prompt misinterpreted as push instead of pull
- Background tasks: dead agents/shells render as Running; lost shell-completion wake-ups freeze agents; app lifecycle events recorded as 'interrupted by user'
- [DOCS] Symlinked skill directories under ~/.claude/skills/ work — please document the support level (cf. #41451)
- Desktop app: Chat pane minimum width too wide, cannot reduce
- [Bug] Anthropic API Error: Model Not Found During Active Session
- [Bug] Anthropic API Error: Invalid or Inaccessible Model claude-fable-5
- [Bug] Locked out of Fable 5 after transcript revision with setup issues
- [Bug] Anthropic API Error: Invalid Model Name for Fable
- [BUG] claude-fable-5 is unavailable
- [Feature Request] Display model name for spawned subagents in CLI output
- [Bug] Anthropic API Error: Invalid or Inaccessible Model Configuration
- [Feature Request]: Auto-clear / context-checkpoint hook for long agent-loop runs
- [BUG] There's an issue with the selected model (claude-fable-5). It may not exist or you may not have access to it. Run /model to pick a different model.
- [BUG] Fable is not available
- [Bug] Anthropic API Error: Unexpected model downgrade to Opus with content policy violation on legitimate infrastructure task
- [Bug] Model access lost without changes: claude-fable-5 unavailable on max plan
- [Bug] Local development login flagged as offensive content by content policy
- [Bug] Rate limiting or model access restriction prevents continued work session after extended usage
- [Bug] Skill listing truncated with descriptions dropped due to context budget constraints
- [Bug] Skill listing truncated with descriptions dropped due to context budget constraints
- [BUG] Windows 11 Terminal and Claude Code: Ctrl-V not working
- [Bug] Anthropic API Error: Invalid or inaccessible model claude-fable-5
- [FEATURE] Open local file links (e.g. PDF) in the macOS default app instead of the in-app viewer
- False positive: Fable 5 auto-switched model for Thai royal condolence popup (HTML/JS)
- [BUG] TodoWrite re-stamps a front-of-context block on every update, invalidating the entire prompt cache (full ~900k-token rewrite per todo update)
- [Bug] Claude Code unavailable after network reconnection
- [Bug] Model unavailable after usage limit reached in session
- [Feature Request] Auto-attach file & line context when pasting code snippets from editor (like Copilot)
- [Bug] CLI fails to run tests when 2FA authentication is required
- [Bug] Auto-update error despite passing doctor checks
- [BUG] [Linux] Transient daemon respawns/displaces itself every ~52s while `claude agents` is open (storm of ~1 `claude daemon run`/sec); drops claude.ai bridge + MCP each cycle
- Subagent model override silently dropped after a continuation boundary (SendMessage follow-up / post-compaction resume)
- [Feature Request] Support for Fable5 model
- [BUG] Linux: files under a sandbox-reachable directory get enumerated ~one rule per file into the `bash` tool schema, overflowing `/context` "System tools"
- Model-unavailable error on startup causes permanent input freeze (not recovered by /model switch)
- [Bug] Generated text contains random repetitions on v4.8
- Bash tool dies (exit 1, no output) when a bound NFS/autofs mountpoint hangs — bubblewrap namespace setup blocks; poisoned until restart
- I don't have a bug report to generate a title from. Please provide the bug report details so I can create an appropriate GitHub issue title.
- Model systematically misrepresents test coverage: claims comprehensive testing while leaving entire subsystems untested
- [Bug] Session continuation fails with mismatched tool_use_id in advisor_tool_result blocks
- Bug: Tool call intermittently emitted in a malformed envelope ("court" + bare <invoke>), so it never executes
- You've hit your monthly spend limit · raise it at claude.ai/settings/usage
- [BUG] Plugin subagent doesn't persist MCP permissions
- [BUG] [Opus 4.8] Generating non-existent information as user input during "invisible turns" and attempting to proceed based on that.
- [FEATURE] Avoid shell quoting issues when passing markdown to CLI tools like gh
- [Bug] Claude Code agent skipping skill file reads despite instructions to read them
- You've hit your session limit · resets 1:50pm (Asia/Damascus)
- [BUG] `/context` is silently sent as a chat message (not executed) in Claude Code on the web — `/usage` works fine
- [BUG]
- [FEATURE] Per-message model selection in Claude Code CLI
- [BUG] At the beginning of the session, the essential documents for the workflow are summarized internally.
- /resume picker reads the entire ~/.claude/projects into a ~23 GB heap at open (no session index since v2.1.30)
- panic(main thread): Segmentation fault at address 0x1269432E492
- [BUG] Claude is not available - Money or compensation!
- [BUG] Slash commands (/compact) sent as plain text instead of executing — regression in v2.1.177 (worked in 2.1.173), affects local terminal AND Remote Control
- Allow grouping sessions into folders/collections (Projects-style organization)
- Write/Edit tool hangs 2–4 min (ep_poll, no network/CPU) when a stdio MCP server is connected
- [BUG] vscode extension model selector update `~/.claude/settings.json` but won't read from it
- [BUG] Claude Code should not refuse to edit ".a" files
- [BUG] plansDirectory does not expand tilde (~) in path
- [Feature] Drag-to-reorder GROUPS in the left sidebar (same as sessions)
- Feature request: Support localization/language setting for permission prompt UI dialogs
- Telegram plugin: quoted/reply messages lose reply_to context
- [BUG] Subscription not detected in interactive CLI, while headless (--print) and Desktop/Web work — org-managed account
- Mobile app: '+' create-session flow fails with 'The request was invalid' on every environment; web works
- [BUG] Regression: Bash tool again escapes "!" → "\!" (regression of #32864, fixed in 2.1.87)
- [Bug] Agent stuck in repetitive loop on multi-criteria optimization tasks
- Agent gives confident but false accounts of its own actions earlier in the same session
- [BUG] Remote Control bridge fails silently on Desktop app
- Web UI: Files/Terminal/Preview greyed out, Diff never shows changes, base mislabeled as 'main'
- [FEATURE] Clearer user-facing communication for access restrictions and restoration status
- [Bug] TDZ error masks actual initialization failure in agent-view sessions
- [Bug Report] Unable to process - insufficient information provided
- [BUG] ⏺ There's an issue with the selected model (claude-fable-5[1m]). It may not exist or you may not have access to it. Run /model to pick a different model.
- [Feature Request] Improve code review and error detection accuracy in agentic mode
- [BUG] Claude.AI desktop version causes Task Manager and Start Up applications to crash
- [Bug] Prompt Injection Detected in Tool Output Display Layer
- [MODEL] Opus 4.8 convinced me I was under a prompt injection attack .It startled me. After repeated verification, it turned out to be an illusion. It was just that the process was so realistic.
- Agent tool `isolation: "worktree"` is silently ignored when spawning teammates with `team_name`
- macOS Keychain dialog appears repeatedly for every new Claude Code session
- [BUG] Claude Code Crash
- Voice mode: configurable silence timeout (re-filing #33792)
- [loop] Stale ScheduleWakeup fires after early manual success — no cancellation mechanism
- [Bug Report] Unable to determine issue from empty bug report
- [BUG] [Fable 5 model] is grayed out although I have a max account (individual)
- WSL2 falsely detected as having conhost viewport yank bug — streaming text unnecessarily disabled
- [BUG] Cowork "yukonSilver not supported" on Windows 11 25H2 (Build 26200) — stable GA channel
- [BUG] I cant log in to my Claude
- Background-job sessions unrecoverable via --resume despite intact transcripts (no signal they exist)
- [BUG] Desktop app black screen on Mac
- [Bug] Unable to process empty bug report
- "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)
- Feature request: Add Hebrew to voice dictation + RTL support in the terminal input line
- [Bug] Disabling Extended Thinking via button shows raw <thinking> tags in chat (VSCode extension)
- Launch preview webview blocks clipboard write (NotAllowedError) even with focus + user gesture
- [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] Personal Max subscription blocked in Claude Code — "Your organization has disabled Claude subscription access"
- [BUG] claude-opus-4-8 fabricates user messages inside thinking blocks and replies to them (~26 min self-conversation; session restart does not help, only model rollback does)
- Git-repo detection cached at session start — agent worktree isolation permanently fails after in-session git init
- [Bug] Anthropic API Error: Invalid Model in Tool Definition (tools.10.model)
- [BUG] Extension install downloads but never completes/registers — silent failure on Windows
- [FEATURE] Show usage stats in the main UI, not just in Settings
- Expired model trial bricks session — no auto-fallback, no remote recovery
- [BUG] Claude-spawned child shells on Windows can't complete new outbound TLS handshakes to non-Anthropic endpoints (breaks Perforce)
- Telegram plugin: message text not XML-escaped, enables prompt injection via crafted messages
- [BUG] Windows: npx-based MCP/plugin servers fail to connect (spawn EINVAL from .cmd shim)
- Non-CLI clients (Android + claude.ai web) prepend a <system-reminder>Message sent at …</system-reminder> block to message bodies, breaking slash-command parsing (regression)
- [MODEL] Claude Code ignores stop button and continues unauthorized operations
- [FEATURE] "Show all projects" toggle in the history panel
- [BUG] Tool-use blocks intermittently rendered as plain text instead of executing
- [BUG] Pro: "Usage credits required for 1M context" despite credits enabled — Sonnet 4.6 [1m] forced to 200K after Fable 5 suspension
- I'm ready to help generate a GitHub issue title for Claude Code. However, I don't see any bug report content in your message. Please provide the bug report details, error messages, or feature request description, and I'll generate an appropriate technica
- AN OPEN COMPLAINT TO ANTHROPIC, INC. On the Machine That Softens What Women Made Sharp
- Plugin MCP relative headersHelper resolved against session CWD, not plugin root
- [BUG] account ban without refund and review request response, don't buy max plan they ban you
- [BUG] Read tool can inject huge partial file output into context, causing repeated ECONNRESET on follow-up turns
- [FEATURE] `claude-vscode.editor.open`: expose cwd parameter to support per-session working directory in single-window setups
- Fork-Subagent recursion guard false-positives on substring match of `<fork-boilerplate` in user messages (v2.1.177)
- forkpty: Device not configured — Claude.app leaks pty masters until kern.tty.ptmx_max is exhausted (macOS)
- AskUserQuestion (multiSelect): typed free-text answer is silently dropped, recorded as empty string
- gitStatus reports branch as ".invalid" in reftable-backend repos (reads .git/HEAD directly)
- [Bug] Claude ignores explicit user instructions and visual guidance in favor of alternative actions
- [Bug] Performance degradation on current date
- [BUG] .mcpb install is a silent no-op on Claude Desktop 1.12603.1 (macOS): "Handling DXT/MCPB file" logs with no preview/install dialog and no error; identical binary works via manual claude_desktop_config.json
- [BUG] `CLAUDE_CODE_OAUTH_TOKEN` is shadowed by a stale `~/.claude/.credentials.json` (precedence reversed in 2.1.x) → `/login`
- Remote Control: built-in slash commands fail because a "Message sent at" system-reminder is prepended ahead of the leading slash
- Long sessions leak one pty per command → pty pool exhausted, no terminal can launch
- [BUG] Opus emits tool calls as prose and mangles ASCII to typographic glyphs when preceding input has typographic punctuation (decode-time input-style contamination)
- [BUG]
- 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] Infinite loop during web page testing exhausts token limit without graceful exit
- [FEATURE]Durable cron jobs should survive session close, not just app restart
- Permission model forces an unsafe binary (alarm fatigue vs. full bypass) — make risk-stratified approval available to individual users
- ccRemoteControlDefaultEnabled: true not honored on restart — Remote Control resets to OFF every time
- [BUG] Max 5x plan incorrectly downgraded to Free after Max 20x refund (Fable 5 government removal, June 13 2026)
- Remote Control: slash commands (/clear, /compact, /context, /bug) route to model as plain text instead of executing
- [BUG] /doctor PATH check is a false positive on Debian/Ubuntu systems with standard .bashrc guard
- Benchmark test: Reykjavik summer solstice astronomy data verification
- [Bug] Claude Code generates hallucinated information instead of verifying facts - THIS IS SYTEMATIC- METAL ABUSE
- [BUG] Hot mic after closing flip phone
- [BUG] Claude Max subscription silently downgraded to Free plan 2 days after payment — invoice TRW50WXN-0015
- [BUG] Slack connector not seen by /schedule
- [BUG] Parser skipped input between top-level statements
- Reykjavik does NOT experience a true midnight sun — only near-24-hour daylight (21h 09m on summer solstice)
- [Bug Report: Unable to process - no bug details provided] COMPLETELY IGNORES ALL INSTRUCTIONS - MENTAL ABUSE DAY IN ALL DAY EVERY DAY
- [FEATURE] Subfolders / nested organization within a Claude.ai Project (web, desktop, mobile)
- Cultural Heritage Coverage Report: European Paintings in Met Museum
- [BUG] Cloud Routines: declared plugins never load — SKIP_PLUGIN_MARKETPLACE=true force-set, startup marketplace resolution skipped
- Benchmark test: weather station coverage verification for San Diego area
- [BUG] MCP OAuth tokens not persisted to disk — re-authentication required every session
- HITL prompt open: submitting a new prompt shows no feedback (indicator doesn't animate), then answers later
- Agent/FleetView background-session TUI corrupts on Windows (stale-cell bottom chrome); reproduces on 2.1.177
- [FEATURE] Escalation request: end-user KYC pathway for Fable eligibility review
- Environmental Health Assessment: Asthma Trials vs Air Quality in San Francisco
- [FEATURE] `/usage` overloads the word "session": top "Session" block (this CLI run) vs "Current session" limit bar (5-hour window)
- [FEATURE] Escalation request: end-user KYC pathway for Fable eligibility review
- [Bug] WebFetch does not return response body on HTTP error codes
- Benchmark test: NightSkyTracker observation config verification request
- California Environmental Health & Data Infrastructure Assessment
- claude-sonnet-4-6
- [BUG] Strava connector appears under two different MCP server names across sessions, breaking permission rules
- [Feature Request] Restore Fable model access for Max tier users
- /model picker writes malformed model ID to settings.json
- NYC Location Assessment Report - Weather Stations and Literary Heritage
- [FEATURE] highlight fish shell syntax
- [BUG] Session keeps sending "waiting for task" messages autonomously, consuming tokens
- [FEATURE] URGENT UI/UX Copy move or add icon from to top to bottom of code block.
- [FEATURE] Native session recovery after Claude Code quota/rate-limit exhaustion
- Workflow fan-out inherits a premium-tier default with no per-agent cost ceiling, causing ~$1k in auto-purchased charges
- macOS desktop: file-link right-click menu disappears when hovering the "Open in…" submenu
- [BUG] Max plan: Opus 4.8 only shows 256k context, 1M option missing in model picker
- Workflow agent(): agentType with restricted tools blocks StructuredOutput when schema is set
- 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
- [Bug] Excessive consent interruptions breaking task flow and productivity
- Opus 4.8 (Claude Code): severe failure on a research/synthesis task — no analysis, defective quotes, fabricated handoff claims
- [Feature Request] Add option to exclude auto-injected user metadata from model context
- /clear is a no-op in Claude Agents backgrounded ("remote control") sessions — context not wiped
- Feature request: let the agent self-manage its context window when it detects degradation
- security-guidance SessionStart hook fails on Windows when CLAUDE_PLUGIN_ROOT is a backslash path
- [Bug] Usage limits not properly resetting across devices/sessions
- [Bug] Anthropic API Error: Inconsistent false positive policy violations on legitimate addiction-recovery educational content
- Test issue for benchmark - please ignore
- Research Report: Hokusai Great Wave Art Auction Data Verification
- Research note: Apple Inc. financial overview and market data analysis
- [BUG] Remote Control: slash commands (incl. built-in /clear) sent as plain text since the desktop "click-to-select skill" change
- Research Tracking Verification Test
- [FEATURE] [Feature Request] Ultraplan should only block new Remote Control sessions, not monitoring of existing ones
- [BUG] issue with cowork and code with arm64 (Snapdragon(R) X Elite) and win11
- [FEATURE] Add search widget in the Claude Panel
- Test issue for benchmark verification
- [FEATURE]
- Built MCP Server Toolkit - Direct integration framework for Claude Code
- [BUG] `claude install <channel>` removes working npm-global install even when the new native install isn't reachable on PATH (cleanup gated on PATH-detection that's already informational only)
- [BUG] Cannot log in to VS Code extension.
- Claude Fable 5 is not available
- Cowork rootfs.vhdx.zst checksum mismatch — server serves wrong artifact (expected 21237ca8, got 01b19a53) on Win11 Home, app 1.12603.1.0
- [BUG]
- [Bug] File checkpointing silently stashes and hard-resets working tree, destroying uncommitted edits
- [BUG] macOS version of Claude Desktop becomes corrupted, freezes, requiring full Reset App before it will work again
- [Bug] Notes text clipped when exceeding display limit in options window
- Desktop app omits Windows junction/symlink skills from the / command menu (agent loader resolves them fine)
- Test benchmark issue - can be closed
- Title: [BUG] Cowork bash sandbox stuck "Workspace still starting" indefinitely on macOS
- Benchmark-Test-Project Server Infrastructure Audit: Threshold Exceedances & Asset Verification
- [BUG]
- AskUserQuestion popup blocks the user from reading the preceding explanation
Resolved issues (326)
- [Bug] Text rendering garbled in terminal on Ubuntu with AMD GPU
- [DOCS] Hooks reference missing `SessionStart` `hookSpecificOutput.sessionTitle` support on startup and resume
- [DOCS] `MessageDisplay` hook event missing from Claude Code hooks documentation
- [DOCS] `/code-review` command docs still describe pre-`--fix` behavior and outdated `/simplify` aliasing
- [DOCS] Skills frontmatter docs omit `disallowed-tools` support
- [DOCS] Dynamic headers docs omit `otelHeadersHelper` failure diagnostics
- [DOCS] Auto mode docs omit the explicit `AskUserQuestion` exception
- [DOCS] `CLAUDE_CODE_SHELL_PREFIX` docs omit stdio MCP argument-preservation guidance
- [DOCS] `--resume` docs conflict on cross-worktree session resumption
- [Bug] Agent workflow resumes rerun completed tasks instead of resuming from checkpoint
- [Bug] Model downgraded to Opus during security checks blocking security vulnerability analysis
- [BUG] claude-fable-5 is unavailable
- [Bug] Model access lost without changes: claude-fable-5 unavailable on max plan
- [BUG] Hints advertise `btw-side-question` feature, which doesn't seem to work and isn't documented
- False positive: Fable 5 auto-switched model for Thai royal condolence popup (HTML/JS)
- Write/Edit tool hangs 2–4 min (ep_poll, no network/CPU) when a stdio MCP server is connected
- Calibration of when to act: agent overrides explicit narrow request, then overcorrects into refusing an authorized action
- [Bug] (Fable 5) Claude Code fails to audit user's own application code or protocols
- I'm ready to help generate GitHub issue titles for Claude Code bug reports. Please provide the bug report content and I'll create a concise, technical issue title following the guidelines you've specified.
- [Bug] TUI rendering corruption during session execution
- [Bug] Anthropic API Error: Safety Filter Incorrectly Flags Legitimate Security Audits
- [Feature Request] Add option to disable or customize work category classification for non-biology codebases
- [Bug Report] Unable to process request - insufficient information provided
- [FEATURE] per-project data directory (store sessions/memory inside project folder)
- [Bug] Claude Code security review request causes unexpected behavior or errors
- Phantom user message: input the user never typed appears as a user turn and gets executed
- [Bug] Terminal text display corrupts to "..." when typing long input in multiple choice prompts
- [Bug Report] Unable to process issue - no bug report content provided
- [Security] Assistant transcript message superseded with injected instruction text, attributed to claude-in-chrome MCP channel
- Feature request: compact/collapsible permission prompt — option to hide command preview & diff to reduce dialog height
- [Feature Request] Support for scientific computing workflows without safety restrictions
- [Bug] Fable 5 silently downgraded to Opus 4.8 for defensive security work without user indication or override option
- feat: add ANTHROPIC_DEFAULT_FABLE_MODEL env vars + hide unsupported models in /model picker when CLAUDE_CODE_USE_BEDROCK=1
- [Bug] Model switches unexpectedly during code analysis
- MCP tools/list_changed: tool index frozen at session create — newly projected tools invisible until reconnect
- I need more context to generate a GitHub issue title. Could you please provide: 1. A description of the bug or feature request 2. Any error messages or unexpected behavior 3. Steps to reproduce the issue 4. Expected vs actual results Once you share these
- [Bug] Input bar positioning unstable - text floats outside input box
- [Bug] Safety Filter Incorrectly Flags Benign Content
- [Feature Request] Add search/filter option for claude agents sessions
- Re-authentication required after every clean restart due to auto-update wiping credentials
- [BUG] /clear typed in one CLI session is delivered to another concurrently-running session in the same cwd, causing context loss
- Usage limit expired bug
- [Bug] Claude Code incorrectly flags legitimate crossword generator requests as cybersecurity violations
- macOS 2.1.170 binary hangs before startup (_dyld_start), install.sh stuck at 'Setting up Claude Code'
- You've hit your session limit · resets 1:40pm (Asia/Jerusalem)
- [BUG] Regression: Bash tool again escapes "!" → "\!" (regression of #32864, fixed in 2.1.87)
- [Bug] Model context not persisting across task switches after security hardening
- API-error retry bookkeeping orphans the successful turn from the parentUuid chain; --resume then silently drops it from context
- TUI: copying assistant output strips spaces & drops wrapped lines on paste (Copilot/Gemini panes unaffected)
- [Bug] Assistant text blocks dropped when streamed with tool_use in remote-control mode (Fable 5)
- Tool calls occasionally leak as literal <invoke> text (missing namespace prefix) instead of executing
- [Bug] Agentic tool building blocked by overly restrictive safety filters
- [Bug] Overly aggressive security flagging for legitimate code review tasks
- [Bug] Fable safety filter falsely flags legitimate development code as unsafe
- [Bug] Cyber classifier false positives triggered by infrastructure terms in workspace context
- [BUG] Session data lost after app update — "디스크에서 세션을 찾을 수 없음"
- [Bug] Claude hallucinating non-existent user messages and refusing to retract after correction
- [Feature Request] Support for Claude models other than Opus in Claude Code
- [Bug] Unable to process request - insufficient information provided
- [Feature Request] Improve agentic reasoning capabilities for domain-specific document analysis in sports science applications
- Feature Request: Localize permission prompt UI (e.g., Korean language support)
- Output degeneration: stray token 'court' emitted before tool calls, escalating to runaway repetition loop in long sessions (Opus 4.8)
- [Bug] Korean character encoding breaks in output display
- [Bug] Opus model produces incoherent code suggestions despite improved reasoning capabilities
- [BUG] Cowork: Desktop update wipes all projects, instructions and folder connections - Windows 11
- [Feature Request] Add option to suppress Anthropic API safety flags for user prompts
- [BUG] Claude Desktop: MCP tool calls intermittently never dispatched — silent 4-min hangs, all local servers wedge until full restart (both-side logs)
- [Bug] Automatic model fallback to Opus without user consent during execution
- [Bug] Anthropic API Error: Tool result missing due to internal error
- [Bug] False Positive Rate in Content Classifier for Data Pipeline
- Advisor/model mismatch (400) silently hidden by fallbackModel — misleading "could not be retried" on Fable 5
- [BUG] Cyber-security usage detected during normal work (non-security related).
- [Feature Request] Dynamic workflows should auto-select appropriate model instead of inheriting parent model
- [Bug] Skill-specific memory not persisting across session restarts
- [Bug] Mythos incorrectly flagging non-cybersecurity content as security-related
- [Bug] Fable 5 model incorrectly flagged as unsafe, forcing unnecessary Opus 4.8 fallback
- I need more information to generate a GitHub issue title. Please provide details about the bug or problem you're experiencing, such as: - What were you trying to do? - What error message did you receive (if any)? - What happened vs. what you expected? - A
- [Feature Request] Display task summary and specifications before long-running operations
- [Bug] Anthropic API Error: Model triggered security review causing request rejection
- [Bug] Model unexpectedly switches to opus-4.8 when using fable-5
- [Bug] Model unexpectedly switched to Opus during security audit task
- [Feature Request] Prevent model from attempting to decompile Claude Code binary
- [Bug] Mythos flag incorrectly triggered during connection method selection
- Fable 5 refuses to assist with 'Application Security Architect resume' editing
- [Bug] Inappropriate content filter blocking legitimate research review requests
- YURI production Flagged for some reason while using fable
- [Bug] Anthropic API Error: Server rate limiting requests
- [FEATURE] persist session color (/color) across sessions via settings.json
- [FEATURE]
- [Bug] Anthropic API Error: Usage Policy violation on reverse engineering detection
- [Feature Request] Add security scanning exemptions or allow suppression of false positives
- [BUG]
- [Bug] Vulnerability scanning not working with current model
- Mac terminal: text selection for copying intermittently stops working
- [BUG] trouble connecting with my remote claude.ai account
- [Bug] Unexpected model switch to Claude Opus 4.8 during task execution
- [Bug] Fable 5 model safety filters causing false positives on benign messages
- Terminal UI rendering breaks on Bengali (Bangla) / complex-script text — corrupted gutter, misaligned columns, overlapping glyphs
- [BUG]
- [Bug] Anthropic API Error: Tool result block missing corresponding tool use block
- VS Code extension: two-column layout (or height caps) so input + long messages don't cover Claude's replies
- [BUG] WSL remote on Windows: false `fork/exec ... no such file or directory` on ccd-cli (binary is present); old chat windows can't reconnect after remote-server restart
- [Bug] Claude Code ignores explicit implementation instructions and user directives
- Config dialog scrolls to bottom when switching pages instead of maintaining focus position
- [FEATURE] Stop/SubagentStop hook payloads should expose token usage, cost, and duration natively
- Feature request: organize conversation/session history into folders or tags
- [Bug] Agent view displays incorrect timestamp and status after session deletion
- Built-in command-safety analyzer false-positives on quoted Windows paths with spaces
- [BUG] Repeated <ip_reminder> injections in CCR web session instruct model to conceal content from user
- [Bug] GitHub MCP Connection Failed
- ad[Bug] Multi-line code block copy-paste formatting inconsistencies
- [Feature Request] Session limit accounting for skill creation should be separate from general usage limits
- Missing conversations in desktop version
- Feature request: concurrent personal + work account sessions
- [Bug] Claude Code stops abruptly during code generation without completing changes
- [Feature Request] Granular per-category notification preferences (desktop/mobile) + don't suppress blocking prompts on mobile when terminal is focused
- [BUG]
- [MODEL] Multiple behavioral failures in single session: instruction violations, blame-shifting, in-session memory loss, and unethical problem-hiding (claude-sonnet-4-6)
- Model quality degradation: overconfident false claims, failure to acknowledge corrections, rationalization of incorrect conclusions
- Git hook output (pre-commit tests) not surfaced to user — model sees it, user doesn't
- [Windows] Relaunch to Update fails with "Another program is using this file" (MSIX file lock issue)
- Plan mode: add a "save and exit" option that doesn't trigger implementation approval
- [BUG] Claude Desktop installation fails with HRESULT 0x80073CFF on Windows 11 Pro 10.0.26200 (fully updated)
- [Feature Request] Keyboard-driven in-chat search + visual-select/yank in the Claude Desktop Chat app (vim-style)
- [FEATURE] Hot-upgrade Claude Desktop
- Client-side exception loading claudemarketplaces.com in embedded webview
- [BUG]
- [BUG] API Error: Extra usage is required for 1M context · run /extra-usage to enable, or /model to switch to standard context
- [Bug] Anthropic API Error: Extra usage required despite MAX Plan subscription
- Add persistent usage/credit display to Claude Code
- [BUG] Random context usage at start of new sessions and on "/clear"
- Plan mode: expose 'auto-classifier collapses plan body' as opt-in setting
- [Bug] Fullscreen TUI: Cmd+Left/Right still broken in Warp (re #50503)
- Asana MCP OAuth fails: redirect_uri not registered for application
- [FEATURE] University for Agents
- [Feature Request] Display wide GFM tables as raw markdown or scrollable instead of expanding to field-value list
- [BUG] Ultra-Review Failing
- [BUG] Opus 4.7 - An AI model with Down syndrome
- [BUG] Claude-spawned child shells on Windows can't complete new outbound TLS handshakes to non-Anthropic endpoints (breaks Perforce)
- [FEATURE] Starred & Recents Roll-up
- [BUG] Auto mode (shift+tab cycle) has blank label and only partially bypasses permissions -- Bash approved, Edit still prompts
- Claude Preview viewer regression: viewport switcher gone, full URL hidden behind click
- Hide or dim "Sign In" button when user is already authenticated
- [FEATURE] MTLS Support with Apple KeyChain and Credential Support
- [BUG] /memory opens external editor but hangs Claude Code; Ctrl+C does not exit
- [BUG] Issue: /ultrareview returned empty findings 3 times in a row, all 3 free reviews consumed without usable output.
- [BUG] "Edit/Write permission prompts in VS Code extension missing 'Always allow' option"
- [BUG] MCP OAuth fails: Protected Resource domain mismatch when server uses different public URL vs internal resource URL (Asana MCP V2)
- [BUG] Claude Code 2.1.119 silently switches you to claude-opus-4-7[1m] even if you pick claude-opus-4-7
- [BUG] Model picker shows Opus 4.7 (200K) but session silently runs in 1M context — burned 18% of 5-hour Max limit on a small prompt
- [MODEL] sonnet 4.6
- Support pasting images directly into the VS Code extension chat input
- [BUG] Claude Slack app returns "No repositories found" for non-coding queries without GitHub connected
- [BUG] Running `claude sessions list` via Bash tool triggers recursive process fork bomb, consuming entire usage quota
- /resume returns empty picker when launch-cwd encodes differently than original session cwd (path-encoding nondeterminism in fh())
- Documentation contradiction: multi-hook firing order is documented as "parallel" in Agent SDK docs but "sequential with short-circuit" in hooks-guide
- [Feature Request] Restore claude -p support or provide alternative pricing for agentic workflows
- Allow plugin monitors to opt out of the exit confirmation prompt
- [BUG] Claude Code CLI hangs with whitespace-only prompt and empty stdin
- Restart Computer with insufficient warning
- Mobile app: new session creation hangs at "allocating sandbox" against local remote-control daemons
- Claude in Chrome: set.or.th (Stock Exchange of Thailand official website) blocked by safety API
- [Bug] GHES contribution metrics not ingested by claude.ai analytics after webhook delivery
- [BUG] JE VIENS DE PERDRE UNE MARGE DE 25K EUR À CAUSE DE CLAUDE.
- [BUG] my message text is stuck in the claude code desktop app when i scroll.
- [Feature Request] Track and label advisor escalations for model training dataset improvement
- [BUG] Model repeatedly begins implementation without explicit user approval
- [FEATURE] Cowork artifacts should not be auto-pinned; add option to create artifacts unpinned
- [BUG] 2.1.141 release regression: output panel auto-reveal, debug log spam, settings not persisting
- [BUG] Claude Code adds Co-Authored-By footer that links commits to unrelated GitHub account
- [BUG] Unable to view Claude's thinking process as of newest "claude code for VS code" version 2.1.141
- [Bug] Curly quotes and apostrophes converted to straight quotes in file operations
- Add Windows Terminal to OSC 8 hyperlink auto-detect allowlist
- Auto-updates silently break ~/Documents access on macOS (TCC denial without dialog)
- [BUG] Deny List not Honored for Edit/Write
- [BUG] Claude leaves behind locked panels in VSCode
- [BUG] Claude Code repeatedly creates backdoors and noops critical security code
- [BUG] Cycling modes in Agents view does not work
- [BUG] VS Code Extension Error: "Received message_stop without a current message"
- Auto-mode classifier inconsistently rejects `Co-Authored-By: Claude` trailer in subagent commits
- Setting to auto-pin newly created chats in the agents view
- [FEATURE] Native SCP for claude code desktop app
- [FEATURE] Add toggle to hide active PR card in Code session chat footer
- [BUG] Claude code web Setup scripts: a failed script is recorded as success, its env exports don't reach agent shells, and its output is never logged
- [DOCS] CronCreate durable: true accepted but silently no-op — session-only crons on macOS
- FleetView: support Shift+Tab to cycle permission modes (plan / accept-edits) in the new-task input
- Local routines: support custom cron expressions in the UI
- Ultrareview crashed with zero output — requesting refund of free attempts
- PreToolUse hooks and --allowedTools not enforced in headless -p mode
- [FEATURE] Multiple independent chat sessions / tabs in a single window
- [Bug] Ultrareview crashes before producing findings, incorrectly consuming free attempt quota
- [BUG] JetBrains plugin: Assertion failed when removing selection listener on null editor
- [BUG] Active session auto-archives within seconds of window focus loss (Windows)
- Allow 'bypass permissions' toggle (Shift+Tab) in agent sessions or /bg background jobs
- Opus 4.7: thinking content invisible AND requested work not executed — repeated silent stalls
- [FEATURE] UI Localization / Language Support
- [BUG] Claude output appears to be another user's query
- Skill tool redundantly invoked after slash command loads skill, fails with disable-model-invocation error
- Sort completed background jobs by last interaction time, not session duration
- [MODEL]
- claude.ai MCP-Apps host silently drops `_meta.ui.csp.frameDomains`, blocking all third-party inline iframe embeds (YouTube, Salesforce, etc.)
- Background job (daemon) Bash cwd overridden to ~/.claude, ignoring job's recorded cwd/originCwd
- Conversation history no longer persists after Developer: Reload Window in VS Code
- [FEATURE] In-session keybinding/command to terminate the current session (equivalent of Agents-view Ctrl-X)
- [BUG] Hook `if` field: path patterns with explicit `/` separators never match on Windows
- [FEATURE] Collapse/expand the AskUserQuestion popover
- [BUG] `/loop` scheduled cron fires at wrong time — offset from session start instead of declared cron minute
- Remote Control: built-in slash commands fail because a "Message sent at" system-reminder is prepended ahead of the leading slash
- [BUG] Slack connector not seen by /schedule
- [BUG] MCP OAuth tokens not persisted to disk — re-authentication required every session
- [FEATURE] Expose rate limit / 5-hour usage data for status line and external tooling
- /clear is a no-op in Claude Agents backgrounded ("remote control") sessions — context not wiped
- [BUG] Unexpected 5-hour usage consumption after Claude for Windows update
- [BUG] awaiting instructor review - Claude Certified Architect – Foundations Certification Exam
- [FEATURE] Bulk session lifecycle management: batch archive/delete/group via MCP tools and UI multi-select
- [FEATURE]
- Subscription shows canceled after successful payment – Invoice WJT1HDYH-0004
- [BUG] Google OAuth on Desktop and web routes paid subscriber to new-account onboarding; support ticket unanswered since May 19
- [BUG] Chat Session RENAME Not Working
- [BUG] Claude Desktop EBADF: bad file descriptor on git-worktrees.json rename – Windows 11 Pro 24H2 MSIX
- Claude Desktop: add shell configuration support (env.SHELL or per-server shell override)
- [BUG] ArchicadTapir MCP server consistently fails to connect on startup with "Could not attach to MCP server ArchicadTapir"
- [BUG] dispatch stuck on Asleep cant reach desktop from IOS
- [Bug] Report: Insufficient information provided
- [BUG] API Error: Server is temporarily limiting requests (not your usage limit) · Rate limited
- VS Code extension stalled ~5 minutes without executing any tools after receiving user query
- /reload-plugins breaks inbound channel notifications
- [BUG] CCD-worktree sessions never surface PR status (chip + sidebar icon stay grey); cloud sessions work
- [Feature Request] Implement weighted command history based on execution frequency
- Windows: --chrome native host and channel feature broken for months
- [BUG] Claude Code Exploit: Claude Code Becomes an Unwitting Executor
- Docs: clarify whether /goal is supported in the Claude Code desktop app
- Claude Code commits and pushes without permission, ignoring explicit CLAUDE.md prohibitions
- Chrome MCP per-domain prompt re-triggers every session despite tool allowlist — no settings.json syntax to allow domains
- [BUG] Cowork "Virtualization is not available" — HCS error 0x80370102 despite all Hyper-V components fully enabled (Windows 11)
- [BUG] Intermittent "Tool permission stream closed before response received" on Bash tool calls
- [BUG] `claude install <channel>` can leave fresh npm-global installs with no functional binary (cleanup runs after attempted native-installer switch fails)
- UX: Add project/branch indicator, real-time task visibility, and multi-agent awareness
- [FEATURE] MCP connector sync should work with CLAUDE_CODE_OAUTH_TOKEN from setup-token
- [BUG] Chris Mundwiller's bug report for financial analysis. Plug-in
- Windows: Polish (and other AltGr-dependent) characters not typed in CLI input on native Windows / Conpty
- [BUG] AltGr+E and AltGr+plus intercepted in chat input on Windows IT keyboard, block € and ] symbols
- [BUG] macOS Keychain warning on every launch: add-generic-password missing -U flag causes -25299 "item already exists"
- [BUG]
- Feature request: Add skill for editing Apple Keynote (.key) files
- [FEATURE] Live artifacts: allow sidebar display name to be renamed independently of id slug
- GitHub Copilot reading .claude/ configuration directory without explicit consent
- Memory leak: ~930 MB/min growth in v2.1.101
- [FEATURE] IsReadInSession(path) query tool / --check-read-state flag
- [Bug] Claude generates hallucinated content without grounding in provided context
- Agent made destructive YouTube changes applied wrong data and falsely reported completion
- /init hangs / times out on an empty directory instead of failing fast
- [BUG] 529에러
- ultrareview: failed Find phase consumed free quota slot (empty results, ~10s heartbeats)
- [BUG] Cowork "Virtualization is not available" — Virtual Machine Platform not responding
- Feature Request: Option to hide/collapse tool call panel headers in Agent SDK environment
- Feature request: lazy skill loading for plugins with many skills
- Feature Request: Hierarchical skill resolution via directory traversal
- Claude Code reflexively dismisses test failures as "pre-existing" instead of investigating — pattern persists across sessions despite corrections
- False positive "Blocked by endpoint security" SIGKILL screen on clean personal Mac (no MDM/EDR)
- [BUG] v2.1.113+ on linux-x64 behind HTTP CONNECT proxy: Bun runtime fetch race condition + 3 cascading regressions
- [FEATURE] Align weekly usage limit reset with billing cycle renewal
- [BUG] Claude Code reverts to English after context compaction
- [BUG] Cowork (Windows 11): drag-and-drop files are not uploaded to the session uploads folder
- [BUG] Make Claude Desktop's auto-worktree behavior optional
- Claude Code declarou problemas resolvidos repetidamente sem verificação real — desperdício significativo de sessão
- REPL crashes with unhandled TypeError when Write tool receives a \?\ long-path prefix on Windows
- [BUG] Serious usage tracking bug: quota increased from 64% to 100% with no active tasks
- [Feature Request] Desktop mode does not support webhook hooks for notifications
- Windows: plugin hooks spawn visible cmd.exe/conhost.exe windows despite windowsHide:true
- [BUG] Detaching VSCode window leaves a "Ghost Session" due to unclosed stdio pipes
- [Feature Request] Session title rename and keyboard shortcut parity in agent view
- [BUG] conector "Claude in Chrome"
- [BUG] Windows: model is told Shell=PowerShell while hooks execute under bash — cross-shell foot-gun for hook authors
- False-positive safety refusal blocks legitimate plant-virology and SARS-CoV-2 bioinformatics prompts; same prompts succeed in main session but fail in subagent context
- Skill tool call returns 'Tool result missing due to internal error' and agent gets stuck
- Severe agent failure in Claude Code: data destruction, content fabrication, and repeated self-rule violations across an ~8-hour session
- [BUG] Plugins not showing up on Mac desktop app
- [Bug] Intermittent OAuthError on Claude Code session start on macOS
- Ultrareview: free credit consumed despite rate-limit failure (no report produced)
- Plan-mode and auto-memory file paths inconsistent across sessions — agent writes to wrong dir when two adjacent .claude-* config dirs exist
- Taskbar icon for Claude desktop app is heavily pixelated on Windows (MSIX missing targetsize assets)
- UX feedback: TaskCreate reminder cadence + permission-block opacity
- [FEATURE] More /color options please (ideally allowing hex number as fallback)
- Ctrl+C does not copy selected text in Claude desktop app — always interpreted as 'stop generating'
- [FEATURE] Add non-interactive /config for temporary mid-session setting changes
- [BUG]
- [BUG] Artifacts and Scheduled Tasks fail when ~/Documents is a symlink (virtiofs cannot mkdir through symlinks)
- [BUG] Selection injection from `ide` MCP server has no documented suppression mechanism, leaks env file contents to conversation transport
- [BUG] gRPC OTLP exporter ignores otelHeadersHelper-generated headers (request metadata not injected)
- [Bug] Claude for Chrome extension incorrectly blocking non-restricted sites with false "organization policy" error
- [BUG] Claude Code Desktop v1.3561.0 - "process exited with code 1" on every tool request (macOS Apple Silicon)
- Bun segfault crash after extended session
- [Feature Request] Add Account Settings gear icon and configuration management UI
- Bug report: Ink rendering corruption during heavy text streaming — cumulative frame buffer contamination
- [Bug] grep shell wrapper silently overrides -E (ERE) with -G (BRE) via ugrep — breaks scripts using extended regex
- Gmail connector cannot switch google accounts "you have some access already" is blocking re-authentication
- [Feature] FleetView: support pasting images into the 'start a task in the background' input
- [BUG] cd-prefixed compound commands bypass permission system (ask/deny rules ineffective)
- MCP stdio servers never auto-reconnect after disconnect
- `compactOnExit: true` — one setting, three problems solved
- Telegram plugin (and all stdio MCP plugins) disconnect after minutes of inactivity - makes messaging channels unusable
- [BUG]
- [BUG] Max 20x: Neither 2x session (May 6) nor 1.5x weekly (May 13) limit increases applied to account
- Mark conversation as resolved without sending a message
- Unhandled case: [object Object] shown when SDK stream ends with had_error after API stall — breaks conversation continuity
- [BUG] Keep getting context management extra inputs are not permitted
- [MODEL] Asked why I am not allowed to remove a disk because it is still using it and deleting random files.
- [BUG] Command history from other sessions appears when using arrow keys in a new session
- [BUG] /clear does not preserve previous session for /resume, contradicting in-app description
- VSCode extension renderer disconnects from live stream during long multi-step tasks; work completes successfully in background
- [BUG] Windows 11 / v2.1.142: JSON Parse error during init - Unexpected identifier "C" after bash-path detection (telemetry dropped)
- [BUG] Plan Mode Clear Context deletes all previous conversations
- [BUG] /agents view on Windows leaks orphan Windows Terminal XAML host on exit (taskbar tile reading "DesktopWindowXamlSource")
- [MODEL] Opus 4.6 breaks plan mode, lies about it under direct questioning — Critical #18986 unresolved after 3 months
- [BUG] Plan Mode (Opus Plan): Claude executes file edits instead of only planning when iterating on the plan
- [Bug] Claude ignores plan mode and continues executing despite acknowledgment
- [MODEL] Claude Code can get access to files with denied permissions through IDE extensions
- Claude: Fraud victim — 14 days no human response after initial support request — chargeback auto-suspension risk
- [Bug] Bulk `claude plugin marketplace update` leaves orphan child processes alive for 10+ minutes
- v2.1.142 disconnects on remote-SSH after extension update — claude binary accumulation + MCP server flapping
- VS Code extension leaks claude binary processes on remote-SSH session reconnect
- [Bug] `/mcp` command only displays global servers, omits project-specific servers from .mcp.json
- Test issue for benchmark verification
- [BUG] Claude Code repeatedly fails to follow CLAUDE.md session branch + push discipline
- Desktop app omits Windows junction/symlink skills from the / command menu (agent loader resolves them fine)