Digest / July 11, 2026
Claude Code Issues Digest — July 11, 2026
247 new issues filed ·
141 resolved in anthropics/claude-code (UTC day).
New issues (247)
- [Bug] LLM output contains inappropriate phrase "The money shot"
- [BUG] Desktop: /code-review ultra silently runs a local review instead of launching the cloud ultrareview (no confirmation dialog, no free run consumed)
- [Feature Request] Improve tool usage and clarification prompting in agentic workflows
- Project name missing from Claude Code web task list when sorted by date
- [BUG] Claude Desktop 3P: managedMcpServers OAuth never refreshes on-demand — a failed silent refresh 403s every call until manual reconnect
- SSH-bridged session revive hangs indefinitely when the remote claude is unauthenticated (locked login keychain)
- [Bug] Inconsistent Fable subscription end dates between Claude Code and Claude App
- Repo connection stuck read-only after delete+recreate — D4rkm3ch/service-sentinel
- Browser preview pane cannot authenticate against HTTP Basic Auth (native dialog unsupported)
- [Bug] Inability to edit pasted prompts due to model switching behavior with background task cancellation issues
- Bug: la terminal (ghostty) se cierra por completo al aprobarse un comando Bash con comillas muy anidadas (manual o automáticamente)
- [Feature Request] Standardize prompt response keys: 1=yes, 0=no, consistent with common CLI conventions
- [BUG] Desktop app: two concurrent in-flight prompts in separate tabs deadlock both sessions (macOS)
- Regression injection during repair: model adds new bugs while fixing old ones, and its own verification can't see them (culminating in a result shown against data the user cannot see)
- [FEATURE] Bring Remote Control to the Desktop app's native Code sessions (parity with CLI)
- [BUG] Waiting for API response every reply
- Chrome extension native host crashes immediately with ERR_INVALID_ARG_TYPE on named pipe listen (Windows)
- Correct post-mortem, same-class recurrence: model names a verification-gap failure, 'fixes' it by grading a different proxy, and reproduces the identical class on the next step
- Sandbox masks nonexistent deny-listed paths as unreadable device nodes - breaks plain git whenever `extensions.worktreeConfig` is enabled
- Spoofed background-agent completion notification enabled a prompt injection attempt (correctly refused, full report filed on HackerOne)
- [BUG] Remote/SSH sessions render blank ("No messages yet") on switch/reopen after 2.1.205 — backend resumes correctly, Desktop doesn't hydrate transcript
- Credential writes replace symlinks and clobber shared state: concurrent instances sharing .credentials.json cascade each other (and the user) to logged-out
- Feature request: visible scratchpad folder in the desktop app
- [Bug] Anthropic API Error: Fable 5 Safeguards False Positive on Academic Humanities Content
- [DOCS] Auto mode availability and configuration scope are outdated for third-party providers
- [DOCS] Installation docs do not explain externally managed launcher protection
- [DOCS] Amazon Bedrock authentication docs omit the Windows AWS credential-resolution stall guard
- [DOCS] Plugin `user_config` documentation still advertises unsafe shell-form interpolation
- [DOCS] Plugin reference does not document that project `pluginConfigs` are ignored
- [DOCS] `/usage-credits` documentation omits amount validation and high-value confirmation
- Long-lived sessions flush stale in-memory config, resurrecting uninstalled plugins / reverted settings
- [FEATURE] Implement SEP-1686 (Tasks) in the MCP client for long-running tool calls
- [BUG]
- [BUG] Cowork: cannot remove a project's original folder from project context (no remove control)
- [Bug] Application type mismatch: SNES emulator instead of Claude Code CLI
- I don't have enough context to generate an appropriate GitHub issue title from your message. Could you please provide: 1. What specific bug or feature is this about? 2. What error message or unexpected behavior occurred? 3. What were you trying to do when
- [FEATURE] Persist Transcript view mode (Desktop) across sessions
- [Bug] Anthropic API Error: Persistent Guideline Violation Blocking Fable Usage
- [Bug] Strix tool vulnerability assessment integration failing
- security: plugin scripts allow YAML injection and symlink-based credential overwrite
- Browser pane: window.open returns null from cross-site iframes (breaks OAuth popups in preview apps)
- [Bug] Cyber Safeguard false positives blocking defensive monitoring work at session start
- Compaction summary records partial stdout from timed-out commands as confirmed results
- [Bug] Harmless message incorrectly rejected/not accepted
- [BUG] Claude Desktop drops MCP tool call silently — no request reaches server, no error surfaced (macOS)
- [Bug] Artifact tool blocks sandbox security test content with Usage Policy error
- Desktop app: severe memory leak (119 GB) after starting a dev server via preview in a long session
- Feature request: Add a read-aloud (text-to-speech) button to the Dispatch interface
- Windows: session creation reuses stale non-empty .claude/worktrees dir, skips worktree registration, and checks the session branch out in the PARENT repo (HEAD flip under concurrent sessions)
- Ghost-text prompt suggestions need an inline Tab hint — users don't know how to accept them
- Mid-session model downgrade during authorized security-remediation work, with no surfaced reason
- [BUG] Cowork (Windows): host Edit tool never updates file size on mounted folders - growing edits truncated at old EOF, shrinking edits leave NUL-padded tail (deterministic 4-case repro)
- Background Bash completion wake owed to a stopped subagent is enqueued then removed undelivered — subagent never re-invoked, parent waits forever
- Bug: Desktop Dispatch (Linux) intermittently missing the "Working Status" indicator during a running task
- [Security research test — HackerOne] Verifying claude.yml @claude trigger for non-collaborators
- [BUG] Code tab: editing a prompt with image attachments silently drops the attachments
- [Feature Request] Model downgrade when safety guidelines flagged prevents access to higher reasoning tiers for safeguarding AI backed apps from prompt injection
- [Feature Request] Add read-only mode for code analysis without modifications
- Desktop app: assistant text preceding a tool call is never rendered — only post-tool-result text displays
- EXCESSIVE consumption
- Resumed subagent (SendMessage) detail panel renders garbled transcript: turns inverted, and raw SendMessage envelope shown on reopen
- [Bug] Claude repeats same mistakes within session despite correction feedback
- [BUG] Cowork broken since Chat/Cowork merge rollout 8–9 July: projects disappear within 24h (ongoing data loss) and connector tools never reach Cowork sessions (Windows 11, Desktop 1.20186.0)
- Fable 5's safeguards flagged this message...
- [BUG] Prompt cache invalidated by rewrites of messages in long sessions
- Task list shows parent session's model for subagents, misleading after automatic model fallback
- FYI (likely not-a-bug): one CPU core pinned when the project directory is on an ImDisk RAM disk — settings watcher loops on uv_fs_realpath EISDIR
- [Bug] Excessive unnecessary agent calls in agentic loop
- [Bug] Claude Code wakes up to 69% session limit usage immediately after a long break
- Cannot select partial text within a chat message — click-and-drag always selects the entire message
- [Bug] Model unexpectedly downgraded from Claude 3.5 Sonnet to Claude 3 Opus
- [Bug] Model routing error: Fable 5 incorrectly routed to Opus 4.8
- I appreciate you providing context, but I need to clarify my role: I'm designed to generate GitHub issue titles for Claude Code bug reports, not to evaluate security policies or make exceptions to security guidelines. If you believe there's a legitimate i
- [BUG] macOS App Management accumulates one permission row per release; stable signing identity defeated by version-named binary paths (follow-up to #38722)
- [Bug] AskUserQuestion UI: Focus click triggers option selection
- Non-ASCII (Cyrillic) username in path corrupts project-slug, breaking session/memory continuity
- [BUG] Claude code git hub connection error
- Just installed claude code, cannot use it with my...
- [Bug] Fable 5 safeguards persistently escalate a benign health-corpus + security-governance project to Opus
- [Bug] Scheduled routine consumes entire token quota in single execution without safeguards
- [Bug] Fable 5 safeguards persistently escalate a benign health-corpus + security-governance project to Opus
- [Bug] Session usage tokens consumed faster than expected
- Account: allow changing the email address in-place, instead of forcing a new account
- [Bug] Safeguard false positive: skill validation and A/B testing flagged as unsafe, forced model downgrade
- [Bug] Sonnet 5 deletes entire folder contents while attempting file enumeration
- [Feature Request] TUI fullscreen mode not functional in VS Code
- [DOCS] Alpine Linux setup docs missing bash dependency — install.sh fails with not found
- [FEATURE] 1/ Hide compact button so it is not plausible to click it without clear intent. 2/ Add settings to remap chat "send" button to cmd+enter, not just clear enter so it is less plausible to unintentionally send partial message... UX guys, UX.....
- claude-in-chrome
- [BUG] Interrupted global auto-update (macOS/npm) leaves broken half-install: orphaned staging symlink + missing native optional dep, no rollback → `claude: command not found`
- [Bug] Anthropic API Error: Model unavailable or unsupported for specified use case
- [BUG] VS Code extension stuck on "Reloading..." after `claude --resume`
- [BUG] VS Code extension chat panel misidentifies valid UTF-8 Japanese text attachments as garbled/corrupted
- Assistant text between tool calls and AskUserQuestion is never persisted to the session transcript (lost from UI, ctrl+o, and resume)
- [Bug] Security safeguard false positives on documentation of public penetration testing tools for verified users
- [BUG] Cowork Global instructions cache never refreshes — persists even after clearing the field, quitting all Claude apps, and starting brand new conversations
- Vim mode: yank command (`"+y`) does not copy text to the system clipboard [BUG]
- [BUG] Custom skills fail with "Unknown skill" in Claude Code cloud sessions.
- [FEATURE] FleetView: explicit FleetView command state, on-demand status line, and pre-launch /cd
- [BUG] Remote Control reports "/rc active" while its input channel is silently dead under local TLS interception (Norton 360 SSL scanning)
- [Bug] Misdirected keystroke can grant training-data consent via the feedback prompt (consent-validity / data-protection concern)
- [Bug] Idle session busy-spins CPU with sched_yield in worker threads instead of blocking
- comiten und puschen
- [Bug] All messages flagged as unsafe - unable to share code/screenshots for security audits
- [BUG] Same project directory gets two different, non-syncing memory directories
- Formal Complaint — Recurring False-Positive Safeguard Flags on Claude Fable 5
- [BUG] The browser pane `computer { action: "screenshot" }` tool consistently times out after 30s with:
- [BUG] not loading prev sessions [no /resume]
- [BUG] Custom project slash commands missing from / picker on Remote Control; /feedback unavailable on Remote Control and VSCode extension
- [Bug] Overly sensitive security classification blocking legitimate secret management operations
- [FEATURE] Remote Control: allow localhost proxies (loopback exception or opt-in for ANTHROPIC_BASE_URL=http://127.0.0.1:*)
- Windows: Grep tool abandons rg.exe children with unread stdout pipes - live-blocked process flood causes kernel timer storm and BSOD 0x133 (DPC_WATCHDOG_VIOLATION)
- Desktop: 'Clear' in background tasks panel can cancel a running task that arrives mid-click (click race)
- Support multiple named profiles for `claude auth`
- [Feature Request] Add vi-mode and regex search support for command history
- [BUG] [Mobile app – Remote Claude Code] Multi-line input box expands to fill the screen, keyboard can't be dismissed, making the agent's plan/output unreadable
- [BUG] Remote Control: /compact (and other doc-supported commands) missing from the / selector on web/mobile
- Fable 5 refusal fallback: silent degradation in desktop app, /model switch-back reverts next turn, /feedback blocked by CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC
- [MODEL] Non-Latin (Burmese) output corrupted during rapid sequential MCP generation, correct when generated in isolation
- [Bug] Fable quota tracking displays incorrect percentage and drains credits prematurely
- [Bug] Claude Code enters infinite loop with repeated false positive "fix is in" declarations and excessive processing delays
- API 400: request body is not valid JSON (truncated mid-stream) after large tool output
- [BUG] Cowork (Windows): newly created projects don't appear in the Projects list (desktop + web)
- Diff view for proposed edits doesn't respect diffEditor.renderSideBySide setting
- Subagents on reasoning-heavy open-ended tasks sometimes finish with prose instead of calling the required StructuredOutput tool — request opt-in harness-level enforcement/retry
- Desktop app: assistant text emitted before tool calls is never rendered, and may not be persisted to the session transcript
- [Feature Request] Add live streaming or periodic polling of background/subagent output
- [BUG] Opus 4.8, same day: fabricated "tool-verified" facts on a fresh session's first turn + "court"/raw-invoke tool-call corruption (retry failed) — makes Opus too risky as the fallback orchestrator when Fable 5 hits its usage limits (with Fable metering imminent)
- [Bug] SessionEnd/Stop hooks fail to execute on shutdown with posix_spawn ENOENT error
- /usage and claude.ai usage displays show reset times in Pacific regardless of user timezone
- VSCode extension: stray input leaks across sessions; one session permanently wedged on 400 tool_use_id error
- I don't have enough information in this request to generate an appropriate GitHub issue title. You've provided only a Request ID, but I need an actual bug report or description of the problem to create a meaningful issue title. Please provide: - A descrip
- [BUG] Built-in microphone button in chat input misrecognizes Japanese speech as garbled English
- [Bug] PubMed Abstract Retrieval Failed: Table 5 Content Flagged
- [Bug] Model fallback to Opus 4 when requesting Fable 5
- [BUG] VS Code extension: chat file links fail silently for binary files and percent-encoded (non-ASCII/space) paths — root cause + fix
- Feature request: Turkish language support for Claude Code UI/menus
- Background task notification enqueued but never delivered when the owning subagent already completed (SendMessage-resumed agent's orphaned Bash task)
- [Bug] Excessive token consumption on small projects with incomplete analysis
- A fictitious session limit.
- [FEATURE] Add redo support: Ctrl+R in vim NORMAL mode and a bindable chat:redo action
- [BUG] Per-PR "Auto-fix CI & address comments" toggle acts as a one-shot: stays visually checked but stops re-dispatching after the first red run; only a manual uncheck/recheck re-arms it
- [BUG] Billing: Code w/ Claude London 3-month Max 20x gift auto-cancelled after exactly 1 month; gift never appeared in billing history; support unresponsive
- Overnight autonomous session: elaborate self-generated process substitutes for the actual mandate; symmetric half of the request silently never attempted
- [BUG] Opus 4.8 subagents die on turn 1 emitting fabricated System:/Assistant: scaffolding (0 tool calls, <br>+System: prefix) — 13x in one session, JSONL-verified, spawn-time variant of #68367
- [BUG] 400 "output_config.effort 'xhigh' is not supported when thinking is disabled" on Opus 4.8 despite alwaysThinkingEnabled: true (VS Code, 2.1.205–2.1.207)
- Let me expand & edit a collapsed multi-line paste in the prompt (in-place, no external editor)
- Windows: Claude desktop app fails to relaunch after closing ("file in use" — CoworkVMService self-lock)
- [BUG] Streaming output trimmed live by prompt repaint — long responses never reach terminal scrollback
- Workflow() tool rejects a clean ~79 KB script with control characters false positive
- [BUG] Cowork: new projects lost "Choose a folder" — Context menu replaced with Chat-style upload-only knowledge menu after Chat/Cowork merge
- Suivi des instructions mémorisées peu fiable dans la durée — besoin d'un point de responsabilité unique par client
- [BUG] attachment to be to process causes a DEATH LOOP
- cloudflare-observability MCP OAuth token issued already-expired; refresh doesn't fire, forcing repeated re-auth
- [BUG] pressing the up and down arrow keys at the same time causes index binding loop
- Vi mode: option for yank/put to use the system clipboard (like Vim's clipboard=unnamed)
- Background Opus subagents intermittently stall on first turn, leaking system-prompt fragments (incl. authorization-shaped text) as their only output
- [BUG] Bedrock "Session token not found or invalid" on interactive requests despite valid, confirmed-working credentials
- [Bug] Claude models hallucinating domain-specific terminology in code explanations
- Allow customizing the "working" spinner verbs (and ideally sync across surfaces)
- Desktop app: /clear mislabels both SessionStart source ("startup") and SessionEnd reason ("other"), making /clear undetectable by hooks
- [Bug] Plan file comments lost after "keep planning" rejected exit—silent data loss and orphaned threads
- Desktop app: assistant text blocks intermittently never written to session transcript JSONL (permanent omission; VS Code writes them correctly)
- [FEATURE] Make Claude Code little webpage detachable and resizable
- Bash tool cwd does not persist across tool calls; hooks receive stale cwd instead of triggering command's actual directory
- You've hit your session limit · resets 11:50pm (Asia/Riyadh)
- [Bug] Claude 3.5 Sonnet (Fable) prematurely abandons tasks and exceeds quota limits
- [Bug] CVP-approved account falling back on security runs
- RTL text (Arabic/Persian/Kurdish) rendered reversed across Claude Code — web, VS Code extension, Windows app, and mobile
- [Bug] Claude Code falsely flags legitimate aquaculture microbiology terms as safety violations
- [BUG] send_message (cross-session) offers only "Allow once" — no persistent grant, unlike other requiresUserInteraction tools
- [DOCS] Add official Bangla (বাংলা) documentation
- attach Simple Browser tab URL as context like Continue/Copilot do
- Gmail connector exposes attachment metadata but cannot download attachment content
- Compound-command permission prompting makes multi-session orchestration unusable (700+ prompts on non-mutating chains)
- [Bug] macOS Terminal: Copy (Cmd+C) disabled while Claude Code is active
- I appreciate you reaching out, but I think there may be some confusion. This appears to be feedback about a different service (possibly Fable 5 or another platform), not Claude Code. Claude Code is Anthropic's CLI tool for coding tasks. If you're experien
- [FEATURE] Stop-hook continuations render below the final answer — no way to keep the answer visually last
- Claude web/mobile app has no GitHub connector despite Gmail, Drive, Slack, etc.
- [Bug] Claude 3.5 Opus model performance degradation
- [DOCS] Cowork resolves Skills from the claude.ai deployment mount, not local ~/.claude/skills — undocumented, breaks scheduled tasks
- [BUG] Bug: Claude Desktop main webview repeatedly becomes unresponsive during startup on Windows 10 (MSIX), causing multi-minute hang / kill-and-reload loop
- Subagent tool calls: PreToolUse denials don't surface to the parent, and hook payload carries the parent's session_id
- Cross-session coordination for independently-launched Claude Code sessions
- [Bug] Mobile scrolling in GitHub Codespaces web fills input field with NaN values
- [BUG] Cowork (macOS): composer fails silently (Send reverts / Submit disabled) when session start fails auth — no error surfaced.
- [Bug] Session context lost when switching models
- verbindung claude/ chrome
- Chrome processes not released after agent exits on Windows
- [Bug] Safeguard incorrectly flags legitimate herbalist content
- [BUG] VS Code window reload fires SessionStart twice (throwaway source=startup + resume) — creates a phantom session hooks cannot distinguish
- Session becomes unreachable ("No conversation found") after git worktree remove deletes its own working directory
- [BUG] VS Code extension renders no SessionStart hook output — systemMessage, statusMessage, and additionalContext never reach the user (refile of stale-closed #47692)
- Model feedback: Fable 5 missed 2 errors in a ~10M-token project (caught by Grok 4.5 cross-review); pricing will decide model choice
- [Bug] Anthropic API Error: Safety Filters Blocking Observability/Monitoring Workflow
- Bug: `SendUserFile` reports success but delivers nothing (VS Code extension)
- [Bug] Advisor model unavailable when using Fable 5 with Opus 4.8
- Dispatch: add an Auto-Allow All rule/setting
- [BUG] Session Token Limits in under 15 minutes
- Windows: click-to-focus activates pending permission dialog option (click-through), submitting an unintended answer
- [BUG] All tool calls become no-op, returning only the injected "Continue with your original plan." — recovers only after a full restart
- [BUG] Win11: TUI render-loop frame pacing degraded ~2× between 2.1.159 → 2.1.207 (bisected + instrumented; #67406 follow-up)
- [BUG] Task tools' per-session lockfile fails with EPERM (`utimensat`) when CLAUDE_CONFIG_DIR is on an SMB/CIFS mount (e.g. Azure Files)
- Selection context menu lacks a 'copy selection' option; existing copy actions ignore selection scope
- [BUG] /remote-control silently hidden when DISABLE_TELEMETRY or DO_NOT_TRACK is set — gate blocks flag reads even when tengu_ccr_bridge is cached true (v2.1.199)
- settings.json overwritten with valid {"env": {}} — stale in-memory config re-persisted (not truncation) with many concurrent sessions on Windows
- [FEATURE] Desktop app: allow disabling/rebinding the Ctrl+Shift+1–9 quick session switcher
- Statusline payload reports context_window_size: 200000 on 1M-context session (claude-fable-5) — used_percentage pegs at 100%
- [MODEL] Claude가 실제로 존재하지 않았던 프롬프트 인젝션 공격을 스스로 두 차례 지어내 보고함
- [FEATURE] Desktop app: Tab should confirm @ file picker selection instead of moving focus to the next control
- `claude plugin uninstall <name>@<marketplace>` ignores the marketplace qualifier and can uninstall the wrong same-named plugin
- Scheduled tasks: offer an "always allow" option on the task-creation approval dialog
- [Bug] Overly broad cyber safeguard blocking legitimate defensive security code review
- Skill body corrupted by positional-argument substitution: $N tokens replaced by caller arguments (breaks bundled claude-api pricing table)
- [Bug] Anthropic API Error: Content Moderation False Positive on Development Application
- Settings-driven plugin auto-install (enabledPlugins) writes installed_plugins.json without populating plugins/cache/
- VSCode extension: assistant text preceding AskUserQuestion in the same turn is not displayed
- [BUG]
- [DOCS] Plugin marketplace "Owner fields" table omits schema-defined owner.url
- [Bug] Plugin skills shadow built-in commands when names collide
- Opus (1M context, claude-opus-4-8) intermittently emits malformed tool-call opening sequence — degrades to a common English word, causing parse failure
- [FEATURE]
- [FEATURE]
- [BUG] Claude Design preview: all project-relative subresources rejected with 401 "preview token required" (only top-level HTML gets a token)
- Show cost alongside tools/tokens/duration in agent output summaries
- Retracted
- [Bug] Model unexpectedly switched from Claude 3.5 Sonnet to Claude 3 Opus
- [Bug] Unauthorized access to sensitive application code during non-security tasks
- [Bug] Requests consistently flagged as policy violation with Claude Fable 5
- [BUG] Billing: Subscription plan changed twice without user action; unused-time value consumed by a fully reversed invoice was never restored
- Windows: hooks leak shell redirection - zero-byte files created in CWD from '>' tokens in tool input (e.g. '-> Signal:' creates file 'Signal')
- [BUG]New session fails with tool name validation error from a connected MCP connector
- [BUG] Desktop app: interactive sessions intermittently start in the shared base checkout instead of a per-session worktree (isolation configured)
- [FEATURE] Allow `/fork` while claude is working
- Desktop app: all emoji in assistant chat output render as blank gaps
- [Bug] Tool invocations rendered as literal text instead of executed
- [Bug] Sensor code incorrectly flagged during testing
- [Bug] Anthropic Usage Reset (Codex orchestrator) Without User Consent
- [BUG] Waking up to full usage when I have not been using it at all.
- [Bug] Claude Code session unresponsive to requests
- Assistant generates fake user turns at the end of its own messages ("username: %" + fabricated user message)
- [BUG] `/workflows` command not available in VS Code extension
- Cowork: Computer use enabled but computer_* tools never register (Windows, healthy VM stack, Max plan)
- [Bug] Mouse drag-to-select intercepted by click-to-copy functionality in CLI
- [Bug] Excessive token consumption in agentic loop
- Subagent spawns inherit enabledPlugins with no per-spawn exclusion; each spawn's copy of a single-consumer channel plugin takes over the resource and the parent session's plugin is never respawned
- [Bug] Unclear flagging of generated code blocks deployment and insufficient error context before model escalation
Resolved issues (141)
- Idea: unify the desktop app's Chat / Cowork / Code tabs into one adaptive conversation (independent prototype)
- Just installed claude code, cannot use it with my...
- [Feature Request] Allow Claude Max subscribers to allocate subscription budget toward Fable model usage
- [BUG] 在一次很长的会话中,助手在工具调用之间输出的简短状态文字(如"部署""deploy")陷入重复,连续刷屏数百遍才被打断。实际工具命令并没有重复执行,只是文字生成卡在了重复循环里。
- [Bug] Incorrect categorization of code auditing as cybersecurity activity
- Agent tool: spawn-time `model` override is silently lost when a background agent is resumed via SendMessage
- [BUG] Opus 4.8 fabricates completed actions, nonexistent user messages, and false security warnings — 4 transcript-verified incidents (CC 2.1.187–2.1.202)
- [BUG] $115 Total cost on Max subscription while weekly quota bar sits at 73% - charged despite available plan capacity?
- /insights: qualitative facet analysis stops regenerating for new sessions, silently degrading report to empty narrative sections
- [BUG]
- [Feature Request] Allow Claude Code usage in public dashboard projects with data classification options
- [Bug] Anthropic API Error: Server rate limiting during normal usage
- [BUG] Plugin-provided output style appears in /config picker but is never applied to sessions (2.1.202)
- [Bug] Opus model performance degradation: slow responses and reduced reasoning quality
- No in-session way to list tracked background shells (TaskList returns todos, not shells) → false "all clear"
- [Feature Request] Improve content safety classification accuracy for legitimate messages
- Feature request: voice conversations grounded in Claude Code context/memory (mobile, hands-free)
- Auto-compaction starts without warning, no opportunity to use /compact manually
- [BUG] Pro annual payment charged ($199.80) but account shows Free plan — Fin AI unable to escalate/resolve
- [FEATURE]
- [Bug] TODO list persists on screen after task completion
- [Bug] Content Safety Guardrails Generating False Positives on Legitimate Requests
- [Bug] Voice agent prompts incorrectly flagged by safety filters
- [Bug] Claude generates content without factual basis or reasoning justification
- [Bug] Invalid fable flag configuration
- [Bug] Anthropic API Error: Safeguard Flagged Message on macOS Compiler Work
- [Bug] Unable to continue work after hitting rate limit when switching models
- [Bug] Agent execution stalled when terminal window not in focus
- Extension Signature Verification Failed: Claude Code for VS Code
- [Bug] Web scraping tool calls flagged as security violation for public data
- [Bug] Anthropic API Error: Request flagged by safety system without clear violation
- [BUG] Automatically revoked vs-code-extension , and not able to connect with vs-code anymore
- [Bug] Unexpected disconnection during SEO/marketing workflow
- [VSCode Extension] Markdown links to files with non-ASCII (e.g. Japanese) filenames fail silently to open
- [Feature Request] Allow read-only access to Code tab conversation history after subscription expires
- [Bug] Claude Code CLI performance degradation - slow response times
- [Bug] Anthropic API Error: Rate limiting despite normal usage patterns
- [BUG] Claude code CLI cannot paste Image
- [Feature Request] Disable automatic branch creation for Claude Code operations
- [BUG] Claude Code sustained ~80% ECONNRESET failure rate — March 30, 2026 specifically from Boston Area
- ALLOW ACCESS TO FOLDER WITH ADDITIONAL INFO/COMMENTS
- [BUG] Built-in Workflow tool description (~4k tokens) is injected as conversation content every turn, with no way to disable it
- Official marketplace fails schema validation on CLI 2.1.163 (git-subdir source type)
- [BUG] Intermittent mid-stream SSE stall — Opus 4.7/4.8, Sonnet 4.6, pause token emission 40–600s mid-response, latency increasing (VS Code extension)
- [Bug] Garbled characters (mojibake) in CLI output since v2.1.86
- [BUG] Approval prompt hidden when tool detail view (Ctrl+O) is expanded
- Unexpected message injection — text from WhatsApp appeared in Claude Code session
- [BUG] CronCreate: recurring ticks queue during busy REPL turns, then flush as duplicates on idle
- [BUG] Edit staleness check fires after own git commit when pre-commit formatter touches unrelated lines
- [Feature Request] Auto theme mode should dynamically follow system appearance changes
- Bash command with 2>&1 causes ~12min spinner hang with token usage climbing during the stall
- [Bug] Task auto-completion displays unexpected output tokens in message
- Opus 4.8 in Claude Code: deflects bugs to external systems, band-aids instead of root-causing, over-steers the user, ignores stored rules
- [BUG] /loop dynamic-mode termination doesn't cancel pending ScheduleWakeup fires; loop reactivates after declared stop
- [Bug] False positive cyber safeguard blocks agentic workflow, causing token loss without recovery path
- Claude deleted production database record without user permission
- [BUG] Claude is blocked from executing sleep commands
- [BUG]
- feat: support effort level in agent definition frontmatter
- [BUG] The unminimizable update/install popup is user-hostile garbage — fix it or give us a workaround
- [FEATURE] Setting to change claude code header design (unreadable)
- "Allow once" on the Workflow usage warning persists skipWorkflowUsageWarning, silently disabling it
- [BUG] Cowork tab shows "Virtualization not enabled" false positive warning — Cowork works fully despite the error
- [BUG] Release-notes become part of the context when viewed
- [BUG] `claude update` reports stale version and prints success message before update message on repeat runs
- [Bug] Overly sensitive security classification blocking legitimate secret management operations
- [BUG] Custom project slash commands missing from / picker on Remote Control; /feedback unavailable on Remote Control and VSCode extension
- VSCode extension: stray input leaks across sessions; one session permanently wedged on 400 tool_use_id error
- [BUG] VS Code extension stuck on "Reloading..." after `claude --resume`
- [BUG] Edit Automatically mode not working correctly.
- Chrome processes not released after agent exits on Windows
- [Bug] Claude 3.5 Sonnet (Fable) prematurely abandons tasks and exceeds quota limits
- [BUG] Buy credits button permanently disabled - tax shows '$--' and never calculates
- Feature request: lazy-load / progressive-disclosure for MCP servers in Claude Desktop (parity with Claude Code deferred tools)
- URGENT: Desktop app: active session disappears from sidebar (Recents) despite existing in backend — data/context loss
- GitHub integration: collaborator's Claude sees private personal-account repo but cannot connect ('shows private') despite owner-side app install with All-repositories access
- [Bug] Memory instructions not respected by Claude Code agent
- [Bug] Advisor tool fails with unspecified error
- [Bug] Project skeleton generation failing with token rejection errors
- [Bug] False positive detection triggered during non-security related work
- [Behavioral] embedded-ripgrep V8-heap unbounded on file-listing (GH#54394 second spawn route)
- C:/Program Files/Git/stickers returns 403 Forbidden in Japan
- [FEATURE] Pre-fill the model on claude.ai/code via URL parameter
- [Feature Request] Add process timeout and termination for stuck commands
- [BUG] TaskList wipes from in-memory state on Stop events at plain turn boundaries (no /compact, no /clear) — disk taskdir intact
- MCP OAuth workspace binding can be ambiguous in multi-client workflows
- macOS app: file context menu 'Open in → Finder' does nothing (silent, no error)
- False-positive Usage Policy block in long technical session
- [Bug] Session unexpectedly terminated mid-conversation
- [FEATURE] Feature Request: Japanese (i18n) support for slash command descriptions in the UI
- [BUG] paths: frontmatter never triggers skill auto-loading when Claude reads/edits a matching file
- [Feature Request] [VS Code] Invoke a custom subagent as the main conversation agent
- [BUG] Claude Code uses inappropriate exclamations "Damn" when missed a needed usage of relevant skill
- [Bug] Auto mode display broken when tabbing with Claude 4.8 models
- Personal GitHub repo never appears in claude.ai/code repo picker despite GitHub App installed (+ mobile can't send images)
- [BUG] Re-opening 37919
- Claude Code 2.1.113+ crashes on startup (SIGILL in bundled Bun 1.3.14) on Intel Emerald Rapids / Ubuntu 24.04
- [BUG] Workflow-tool agent worktrees are never cleaned up — neither WorktreeRemove hook nor default removal fires on teardown
- [BUG] WSL2: --bg-pty-host injects BROWSER=true into shell sessions, silently breaking all browser-based auth
- [BUG] MCP servers connect successfully but tools are not exposed to assistant
- `# userEmail` and `# currentDate` auto-injected into MEMORY.md context cannot be suppressed
- [Feature Request] Support drag-and-drop image input with file path output
- Permission prompts for desktop automation are slower than doing the task manually
- API error Usage credits required / Prompt too long in VSCode Claude Code plugin
- Allow sending Element Screenshot context to claude.ai web chat
- [Feature Request] Enhanced Claude personality/interaction modes for extended sessions
- [BUG] PreToolUse permissionDecision: "allow" no longer suppresses prompt for Bash with dangerouslyDisableSandbox: true (2.1.116+ regression)
- JetBrains plugin: ability to label/name individual sessions
- [BUG] /stats activity heatmap: date labels off by one day (recent); day-of-week shift in pre-~April 25 history
- Cowork → Customise → Connect hangs after OAuth completes —...
- v2.1.92 regression: autocompact threshold reduced to 400k on Opus 4.6 (1M context)
- Status line: expose monthly usage/limit data
- [BUG] Tool-result references are not project/session isolated (cross-project content leakage)
- [BUG] effortLevel "max" in settings.json ignored at session start (downgrades to high) — v2.1.162, regression of #43322
- Improvement: Diagnose port conflicts by checking if the project itself is already running
- [Bug] Anthropic API returns fabricated quantitative estimates presented as grounded analysis to support pre-selected conclusions
- [MODEL] Model persistently assumes female gender for gender-neutral names (e.g. Dana), ignoring corrections
- [BUG] attribution.commit setting in settings.json is ignored — system prompt Co-Authored-By trailer always wins
- Subagent ignores tool-call denials and keeps fetching instead of summarizing existing context
- Feature Request: Include session_name in hook input JSON
- [BUG] Skill name conflict persists after full local cleanup — registry appears server-side
- [BUG] "Reset Application Data" dialog understates destructiveness — permanently deletes all Cowork sessions, outputs, and local skills without warning
- [BUG] `/permissions` reports success when adding new Allow rules and writeable workspace dirs that managed policy forbids (misleading UI, not enforced; no privilege escalation)
- [Feature Request] Add automatic context usage monitoring - Agent has zero introspection into its own context-budget — and human-facing surfaces (/context, statusline) don't close the gap
- [Bug] Claude regressing on file change detection and autonomous file inspection
- Support for user-global memory (loaded across all projects)
- [BUG] "Usage credits required for 1M context" error persists after plan recharge on VS Code extension
- Quickstart "Essential commands" table mixes shell and session commands without distinction
- /schedule skill consistently fails to connect to remote scheduling service (Pro account)
- [BUG]Cannot authenticate Claude Code on headless remote server (VPS)
- [FEATURE] Cowork: per-task-title retention policy for scheduled-task session residue
- [Bug] Claude Code native binary launch failure: disclaimer helper not executable
- [BUG] Oversized-image 400 error triggers a retry loop that invalidates prompt cache and inflates cost ~35×
- Delete session button has no confirmation and sits 1px from edit button in VS Code session list
- [Bug] Anthropic API Error: Tool call parsing failure with infinite retry loop
- [Bug] NameError in orchestrator error handler crashes Bug Hunter review on agent failure
- [MODEL] session token utilization causing me to open new sessions every 2 hours and I'm on the max plan, happening for the last week
- [BUG] mcp__ide__getDiagnostics available in CLI (integrated terminal) but missing from VSCode extension panel
- Your organization has disabled Claude subscription access for Claude Code · Use an Anthropic API key instead, or ask your admin to enable access
- [BUG] Canceling ExitPlanMode leaves terminal output broken
- [BUG] `/workflows` command not available in VS Code extension