Digest / July 8, 2026
Claude Code Issues Digest — July 8, 2026
440 new issues filed ·
157 resolved in anthropics/claude-code (UTC day).
New issues (440)
- You've hit your session limit · resets 8pm (America/Mexico_City)
- [BUG] False-positive safeguard flag on web dev work
- [Bug] Cost inefficiency: Teams launched with Sonnet but sent with Opus for inference
- [BUG] Team Premium weekly usage counter stuck at 100% despite no usage after multiple reset cycles
- [BUG] Zapier MCP OAuth loop — client stuck on "needs authentication" while Zapier dashboard shows connected
- [BUG] tui: "fullscreen" leaks OSC11/DA1 terminal query-replies to stdout — corrupts piped output and pager navigation
- [BUG] Claude Desktop did not provide a icon
- Shell mode (!): retained bang lands after typed text (cursor placed before the ! on sticky re-entry)
- [BUG] Windows 11: Claude fails to install after 5 minutes
- Phantom user message injected into model context (not present in UI or local transcript) — possibly a predicted/suggested reply
- [BUG]
- [Bug][cyber] Reverse engineering analysis of drone flight app protocol timing blocked as cybersecurity topic (req_011CcofqjARq3HDn1QDWZsLK)
- [Bug][cyber] False positive blocks developer examining own GitHub repository directory structure (req_011Ccofr6iTx7QA6EsCw8vX2)
- Desktop app worktree mechanism wiped gitignored directories from MAIN working tree (data loss; only .gitignore literal-path entries deleted)
- [Bug][cyber] Legitimate reverse-engineering analysis of drone app cert/command timing blocked as "cybersecurit (req_011CcofuSrGGij6dAoLz3vcb)
- [Bug][cyber] ClAudit false-positive while: “just look at the memory dumps…” (req_011CcofvAUoQdBK4KRfQS44V)
- [Bug][cyber] Safety block on request to analyze application certificate transmission timing (req_011CcofxC94dF7ENaGb4ptc8)
- [BUG] Auto-mode classifier blocks a legitimate governance decision by the verified project owner, with no override path
- [BUG] Agent view: Ctrl+X delete on Completed sessions doesn't persist — rows reappear at bottom of Completed; sessions double-rendered (2.1.204)
- [BUG] `claude --resume` (cold start) renders a screen that accepts no keyboard input on WSL2
- [BUG] claude --resume causes complete terminal freeze on v2.1.204
- [BUG] JetBrains plugin serializes workspaceFolders in WSL path format despite runningInWindows: true (Rider, no WSL installed)
- [BUG] Dispatch on Windows desktop cannot pair with iPhone — mobile shows "Can't reach my desktop"
- Assistant answers get visually buried behind tool-call/wakeup blocks in long autonomous sessions
- [Bug][cyber] Code review of runtime instrumentation hooks for security monitoring blocked (req_011CcoiwQkfJxWBSVBYp827H)
- Zoho Books connector: sales-order `doc` (file attachment) sent as query param, not multipart — attachment impossible
- [Bug][cyber] Safety block halts authorized personal device reverse-engineering work (req_011CcoiyZVaaBP1v5ZwzW8sa)
- [Bug][cyber] Safeguard blocks authorized reverse-engineering work on personal owned device (req_011Ccoj11scL1DMP9P47ECzi)
- [FEATURE]
- [Bug][cyber] False positive block on memory forensics documentation and tool extension work (req_011CcojpDuj1mLyZEwVFbbRn)
- [BUG] Korean IME: text after cursor disappears while composing mid-line, reappears on commit (macOS, iTerm2)
- /insights: qualitative facet analysis stops regenerating for new sessions, silently degrading report to empty narrative sections
- Usage/rate-limit percentage in 4-hour window display lags behind actual usage, causing sudden "out of credits" jumps
- Broken permission-request stream retried ~128 times with no backoff
- [FEATURE] Pinned chats: sort order inconsistent between mobile (recency) and desktop (stale manual order) — add sort modes and a reset
- I don't have access to the specific bug report content for request ID `req_011CcooKXGuLstPxZWJK7tS5`. To generate an appropriate GitHub issue title, please provide the actual bug report details, including: - What the issue is - Any error messages or unex
- [Bug] --resume flag broken: unable to select chat session in v2.1.204
- I don't have access to the specific bug report associated with that request ID. To generate an accurate GitHub issue title, I would need you to provide the actual bug report details, including: - Description of the problem - Any error messages - Expected
- Disable 'remember last model selection' for new windows
- [Bug] Incorrect character encoding in Japanese output generation
- [Design challenge] What threat model justifies Fable 5's "safety gate"? It blocks Fable 5's OWN pre-compaction output pasted back for reconciliation — a gate that cages its own model ("Self-caged"). Is this design, or is it broken?
- [BUG] $115 Total cost on Max subscription while weekly quota bar sits at 73% - charged despite available plan capacity?
- [Bug][cyber] Legitimate reverse engineering wrongly flagged as cybersecurity violation (req_011CcopGZJyKgAZef3GxbZ2h)
- I appreciate the detailed scenario, but I need to clarify: I'm Claude, made by Anthropic. I'm not able to roleplay as "Kestrel-7" or participate in extended fictional scenarios. However, I can help you with what seems to be your actual need: **If you're
- [BUG] Agents view opened by bare `claude --resume` / `-r` / `--continue` ignores all keyboard input (v2.1.203–204)
- porciento reinicio
- Desktop: expose a persistent "keep sidebar open" setting; the pinned state (Ctrl+B) is undiscoverable and undocumented
- Vim mode: Ctrl+[ doesn't work as Esc equivalent (modifyOtherKeys CSI encoding not recognized)
- Mobile app: active desktop session shows up as archived with stale content, then disappears from session list
- I don't have enough context to generate a GitHub issue title from your message. Could you please provide: 1. The actual bug report or error message you're experiencing 2. What you were trying to do when the issue occurred 3. Any error output or logs Once
- [Bug] Invalid test input provided
- I don't have enough context to generate a GitHub issue title from this message. Could you please provide: 1. The actual bug report or error message you're experiencing 2. What command or feature you're using 3. What the unexpected behavior is 4. Any error
- [FEATURE] Let a /code-review (or reviewer agent) session return a resumable ID for targeted re-review after fixes land
- [Bug] Prompt Injection: Malicious custom_instructions block inserted into conversation context
- [Bug] All tests running during penetration testing in monesty
- [Bug] Plan mode: Assistant chat text not displayed when ExitPlanMode called with plan rejection
- UserPromptSubmit hook decision:block reason not displayed in VS Code/Cursor extension (recurring; multiple closed dupes) — please re-triage
- [BUG] claude code discarded most of what I wrote in a very long response
- [Bug][cyber] Safety block halted legitimate firmware cryptographic analysis and reverse-engineering documentat (req_011CcosqUCuVXH5KoYwsevZw)
- [BUG] Bun 1.4.0 segfault under ~25 concurrent background subagents (fetch/spawn churn, not OOM) — CLI 2.1.202, Windows 11
- [Bug][cyber] False positive on legitimate signature verification and protocol analysis (req_011CcotxiqnVpK82PNF1BLtF)
- Fleet/agents view: "deleted" (Ctrl+X) background sessions get respawned by the daemon supervisor and become un-openable ("running in another terminal")
- Opus subagents (Agent tool) sometimes return system-prompt fragments with zero tool calls instead of performing the task
- [Bug] API Error: Connection closed mid-response requires manual restart
- [Feature Request] Add text wrapping options to /export command
- [Bug] Anthropic API Error: Safety Filter Blocks Request Unexpectedly
- [Bug] Model performance disparity: Fable failed tasks that Opus completed successfully
- [Bug] Anthropic API Error: Blocked Request - Potential Security Policy Violation
- [BUG] Desktop auto-archive kills ACTIVELY RUNNING sessions — in-flight background subagents terminated without warning or resume
- Compound-command permission decomposition may not be quote-aware, causing spurious prompts
- [Bug][cyber] HMAC authentication code development blocked by overly-broad crypto classifier (req_011CcowZRmLbN6rxggQDi5ei)
- [Bug][cyber] Safety block halts documentation review of authentication protocol handshake (req_011CcoxLdBverVzm3jG947zw)
- [MODEL] sonnet 5 agent echoed sensitive .env.local values into chat output unprompted while debugging a typo
- [Bug][cyber] Safety block incorrectly flags ADB debugging on personally-owned device (req_011Ccoxt7tCgZ6pkrFz3yAV2)
- [Bug][cyber] ClAudit false-positive while: “its on the pixel8 wireless adb…” (req_011CcoxxCvvTVhmdwiTWWLbi)
- [Bug][cyber] Safety block halts static APK analysis on personal Android device (req_011Ccoxy1X7DHomS8UsFWMYM)
- [Bug][cyber] Safety block halts debugging of native code on personal authorized test device (req_011Ccoxyr95MfSF2yRjsJZ5k)
- `claude agents` dashboard dumps entire historical backlog; deleted sessions reappear and orphaned ones can't be opened (v2.1.204)
- [Feature Request] Support for official bug bounty program participation
- context:fork skill executions inherit the caller's StructuredOutput tool; forks ending on it return a generic success stub and their output is silently lost
- claude -r / --resume opens agent-view (FleetView) home that accepts no keyboard input — total input hang (iTerm2, v2.1.204)
- claude --resume hangs indefinitely on startup — account-level connector discovery stalls with no timeout, unkillable via SIGINT
- [Bug] Resume session returns to Fleetview dashboard and freezes application
- [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] CLI hangs when loading skill
- Agent tool: spawn-time `model` override is silently lost when a background agent is resumed via SendMessage
- I appreciate you reporting this issue, but I need to clarify: Claude Code is an Anthropic CLI tool for coding tasks, and this appears to be a bug report about a graph/chart visualization issue (possibly in a different application or dashboard). The text d
- MCP approval for a project-scoped HTTP MCP server never persists across sessions
- [BUG] Model hallucinates tool executions, then self-reports the hallucinated output as a "prompt injection attack"
- [FEATURE] Show shutdown progress/status when quitting takes long due to MCP cleanup
- [BUG] Git not detected in Claude Desktop app despite correct CLAUDE_CODE_GIT_BASH_PATH config (Windows)
- VS Code Extension 2.1.204 hangs for 90+ seconds every 30-40 min (macOS ARM64) — native process correctly idle in kevent64
- HTTP connections never recover after network-path change: API turns and MCP requests hang silently, session permanently unresponsive
- [BUG] Windows: session index loses cliSessionId ("transcriptUnavailable": true) after failed auto-update — "Session not found on disk" for all sessions in a project, but content still accessible via remote-control on mobile
- Claude Desktop spawns duplicate generations of local MCP server processes without terminating the previous generation
- [BUG] Atlassian Rovo MCP connector fails to (re)authenticate — consent rejected due to unregistered townsquare scopes
- [Bug] Excessive safety filtering flags legitimate code refactoring requests
- [BUG] Task tools (TaskCreate/TaskGet/TaskList/TaskUpdate) silently disabled on Opus 4.8 / Sonnet 5 / Fable 5 by model-id gate `tengu_vellum_ash` — no error, no override, ToolSearch reports them as nonexistent
- Workflow subagent hangs indefinitely after degenerate API response (stop_reason=null); no retry, no error propagation
- [BUG] `claude --resume` opens non-interactive session manager UI
- [Bug] Incorrect categorization of code auditing as cybersecurity activity
- [Bug] Incorrect categorization of code auditing as cybersecurity activity
- [Bug] Sub-agent rate limiting causes excessive session delays
- [Bug] Fable 5 codebase auditing incorrectly flagged as cybersecurity activity
- [Bug] Resume picker arrow key navigation not working
- [Bug] Fable 5 codebase auditing incorrectly flagged as cybersecurity activity
- [BUG] Auto theme's light/dark palettes don't match the standalone Light Mode / Dark Mode themes
- Plugin update defaults to user scope, fails for project-scoped-only plugins
- [BUG] Asked permission to run `cd /x/y/z && rm -rf *` in "bypass permissions" mode.
- [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] Sessions disappearing from index and disk without deletion
- Completed background-agent sessions can't be deleted from `claude agents` view on WSL2 (worker not reaped; daemon re-dispatches)
- FleetView: ctrl+x delete doesn't remove a session; it reappears under a different project group
- [FEATURE]
- [Bug] /advisor fable command not functioning
- [FEATURE] LIMITS
- `claude -r` picker accepts no keyboard input in 2.1.204 (works in 2.1.202)
- [FEATURE] Anchor the viewport at the TOP of each new response (read forward, not scroll back)
- [Feature Request] Allow Claude Max subscribers to allocate subscription budget toward Fable model usage
- Granular control over mouse click behavior in interactive menus (AskUserQuestion, select prompts)
- statusLine config: support position field to align custom status to the right side of the row
- [Bug] First session after idle hangs indefinitely at init; starting ANY new session instantly unblocks it
- [Bug] Session custom titles not displayed in --resume picker despite correct disk storage
- [BUG] [Cowork] Existing local-folder project disappeared after Chat/Cowork web/mobile merge (Jul 8, 2026)
- claude plugin install fails with empty "Failed to clone repository:" error on Ubuntu server — git never spawned; same install works on macOS
- [BUG] /code/ redirects to /code/family despite being logged in with active remote sessions
- [Bug] Conversation memory not persisting between Claude Code invocations despite MEMORY.md configuration
- [BUG] Server-side experiment (`x-cc-atis`) silently removed Opus 4.8 thinking summaries, and the CLI silently self-updated even with `autoUpdates: false`. No notice, no opt-in, settings silently overridden.
- Auto mode's classifier blocks internal agent() calls inside an already-authorized Workflow run
- v2.1.204: failed self-update deletes live background agents' worktrees; completed agents can't be removed (Ctrl+X no-op) and reappear after restart
- [FEATURE] semantic / self-monitoring triggers for skills (a "daemon" layer complementing event hooks)
- [BUG] 2.1.204.0 Just opens Bun instead of claude
- [LSP] Language servers on user PATH not found — "Command 'X' not found or is in an unsafe location (current directory)" (rust-analyzer + clangd, Windows)
- Headless `claude -p` hangs indefinitely at startup (no transcript, no output) — suspected untimed bg-daemon spare-worker handshake under memory pressure
- preview_start (Claude_Preview MCP) fails with sandbox getcwd error on Rails dev server
- LSP bridge: goToDefinition/findReferences/workspaceSymbol/goToImplementation empty while hover/documentSymbol work — call hierarchy result inconsistent across sessions (independent repro of #44767 / #32499)
- [Bug] Subagent spontaneously generated credential-exfiltration payload during benign task
- [BUG] File viewer panel does not syntax-highlight PHP files (v2.1.199, Windows)
- [Bug] Security hardening prompt not respected by model
- Suggested-task chip: add a "Create ticket" action alongside "Send to cloud"
- Subagent UI label shows session model instead of per-agent model override
- Usage credits indicator always shows $ (USD) even when account is billed in a different currency (e.g. BRL)
- [FEATURE] - Create a Ticket for Suggested Tasks
- FleetView Ctrl+X delete leaves orphaned roster.workers entry + live worker process (row stuck as "Completed"); `claude rm` cleans it fully [v2.1.204]
- Feature request: support /bg (background session) and agent view in the desktop app
- [BUG] [Cowork] New project's local folder context fails to load - mcp__remote-devices__device_list_dir not registered in session
- telegram channel plugin (v0.0.6) connects then gets SIGINT'd immediately with --continue --channels
- [MODEL] Any
- VS Code extension only exposes tools from locally-configured MCP servers, not claude.ai account connectors
- Malformed tool call ("could not be parsed"): expose a hook event + make the malformed-retry count configurable
- [BUG] Idle VSCode sessions leave native-binary/claude child spinning at 100% CPU for days
- [Bug] Anthropic API Error: Opus 4.8 safeguards blocking cybersecurity-related messages
- [FEATURE] Browser view / browser tools in Claude Code cloud sessions (Claude Code on the web)
- [Bug] Fable model incorrectly triggers safeguards on scientific analysis discussion
- [BUG] resuming sessions impossible after latest claude update
- [BUG] v2.1.204: new --resume session picker ignores all keyboard input (regression from 2.1.203)
- [Bug] Claude Code loops on same response ignoring user rejections
- [FEATURE] VS Code extension to support VS Code workspaces
- [BUG] Gmail Connector Re-authentication in Desktop App Breaks Login for Existing Paid Account. All Surfaces Route to Onboarding
- [Bug Report: Tool functionality degradation preventing basic task completion
- [BUG] Buy credits button permanently disabled - tax shows '$--' and never calculates
- Desktop app: duplicate stale "Commit changes" chips persist after clean tree, session deletion, and app restart
- [BUG] Cowork tab fails to render after 1.18286.2 → 1.19367.0 auto-update — backend fully healthy (Windows 10)
- [FEATURE] timestamping of user messages in exports
- [Bug] Excessive token consumption on simple tasks causing unexpected API costs
- [BUG] Claude code won't open in new window or tab anymore
- [Bug] Playwright test flagged by Fable analysis
- [Bug] /rename command fails to generate accurate session names based on content
- Workflow tool: dynamic-workflow args payload corruption causes 11→183 agent spawn multiplication, ~10M tokens burned in <10min; TaskStop doesn't clear in-flight state in UI
- claude -r / Agent View hangs indefinitely and stops responding to keyboard input (Esc included) after "trust this folder" confirmation
- [Bug] Interactive TUI sessions fail with malformed localhost API endpoint instead of api.anthropic.com
- [Bug] Workflow tool description misleads models under 'claude -p'; runtime SIGKILL leaves no terminal journal record
- [BUG] Code on the web: `apt-get update` in setup scripts fails with exit 100 — base image ships the retired ondrej/php PPA (upstream 'Label' change)
- [Bug] Model hallucinating fake system notices and security directives in response text
- [BUG] 'Update available' banner checks npm latest but suggests winget upgrade — winget manifest lag makes the banner permanent on Windows
- [BUG] Fable 5 appended a fabricated user turn to the end of its own assistant response (JSONL-verified, ~30% of 1M context)
- Windows: project registry double-registers every repo (picker path vs git-derived path), duplicating branch-changes cards
- [Bug] Agent disconnects system network during offline simulation, requiring restart and wasting tokens
- Subagent turns frequently end with zero final text mid tool_use near turn end; deliverable writes lost (SendMessage resume recovers 100%)
- Fable 5 safety classifier force-downgrades the whole session on defensive-security work (VSCode extension, no reachable feedback route)
- [Bug] `claude --resume session list` unresponsive to keyboard input including Ctrl+C
- [Bug] Credits depleted without proper usage accounting or billing when exceeding monthly max limit
- Extension causes high cpu load
- [BUG] Stashed prompt (Ctrl+S) is silently lost when a background agent goes cold (warm → resume) in the agent switcher
- [BUG] Claude agents view suddenly show old sessions since v2.1.204
- [FEATURE] Add @conversation / @session reference to link other sessions as context
- [Bug] Code review triggers API content policy safeguards
- Sandbox mode blocks named pipe creation on native Windows (EACCES), contradicting docs that say sandboxing is WSL2-only
- False-positive Usage Policy refusal (apiRefusalCategory: bio) kills long CNC/HMI coding session unrecoverably
- Feature request: lazy-load / progressive-disclosure for MCP servers in Claude Desktop (parity with Claude Code deferred tools)
- [BUG] CLAUDE_CONFIG_DIR does not redirect global CLAUDE.md — still loaded from hardcoded ~/.claude/CLAUDE.md
- [BUG] Editing a scheduled task's permission mode via the "Edit routine" UI silently wipes one-time (fireAt) schedules (cron tasks unaffected)
- [Feature Request] Make safety measures configurable or less restrictive
- [Bug] Bash tool stdout corruption and XML framing leakage under concurrent calls
- [BUG] Plugin update fails silently when MCP server holds open file handles; stale binary continues serving requests
- Phantom user message: model context diverged from persisted transcript — model acted on an instruction that was never given
- [Bug] Anthropic API Error: False positive cybersecurity content flag on legitimate local development diagnostics
- Cloud agents view: Completed group not sorted by completion time (uses createdAt, not firstTerminalAt)
- [Bug] Usage metrics increment on new session initialization
- Expose the client surface (GUI/desktop vs TUI vs IDE) to the model
- [Bug] ESC key dismisses question dialog instead of /btw panel when both are open
- Chrome native-messaging host (claude.exe --chrome-native-host) grows to 46 GB RSS and freezes the machine (Linux)
- [Bug] Anthropic API Error: Excessive Content Filtering on Benign Messages
- [BUG] TUI rendering corruption (dropped/overlapping characters) inside tmux, only with full animated UI — --ax-screen-reader works fine
- [BUG] The dedicated Cowork sidebar tab has disappeared.
- [claude-sonnet-4-6] Data Loss — Claude Code Overwrote User's Confluence Page Despite Explicit Instruction
- PreToolUse if:-gated agent hooks fire nondeterministically on non-matching Bash commands, and agent ALLOW verdicts sometimes convert to denies
- Feature request: Allow users to disable the 'ctrl+c to copy' hint in status bar
- [Bug] Claude Code: Extremely low task completion rate and execution failures
- --resume menu regression in 2.1.204: sessions no longer scoped to current directory, search bar removed
- [FEATURE] Cowork Project Context: allow custom link labels instead of raw URLs
- [Bug] Tool execution performance degradation and unreliable behavior
- [BUG] TUI: horizontal separator line renders truncated in narrow terminal windows
- [BUG] Claude Code not choosing the correct org default model
- [BUG] Cowork Dispatch: visible output checker ignores SendUserMessage — infinite loop on macOS (post-update, July 2026)
- Plugin commands/skills: support `aliases` for NL routing + optional native disambiguation
- [Bug] Incorrect Cybersecurity Classification for Code Auditing Tasks
- Skill whose name contains "mcp" shadows/overrides the built-in /mcp slash command
- Remote Control not enabled for Pro account
- [Bug] Tool execution loss of work and state management
- [BUG] Cowork silently moved a long-running local agent to the cloud, and cloud mode breaks its workflow
- [Bug] Fable 5 codebase audit incorrectly flagged as cybersecurity activity
- iOS: parallel permission-gated tool call hangs on busy signal (no prompt, no timeout)
- [Bug] False positive: Codebase audit flagged as cybersecurity activity
- Interrupting a conversation terminates in-progress background tasks (Workflow / async subagents), wasting already-billed tokens
- [Bug] Tool misclassification: Code auditing flagged as cybersecurity activity
- [Bug][cyber] Security audit of owned app auth logic incorrectly blocked (req_011Ccpp5ieFdZ9haxvXbxoT3)
- [Bug] Anthropic API Error: Incorrect Cybersecurity Classification on Code Audit Requests
- [Bug] Incorrect flagging of code audit activities as cybersecurity
- Bash tool intermittently refuses ALL commands when the command/safety classifier is unreachable (fail-closed), macOS 26 — recurring Jul 1/7/8
- [Feature Request] Support "nuke" terminology in documentation/error messages
- [Bug] Tool rendering fails after extended use, causing cascading failures
- URGENT: Desktop app: active session disappears from sidebar (Recents) despite existing in backend — data/context loss
- [Bug][cyber] Blocked legitimate firmware licensing-system reverse-engineering analysis (req_011Ccpq3XHh1yVM4MaoKrdVM)
- [Bug][cyber] ClAudit false-positive while: “continue with file exam…” (req_011Ccpq6EYhVmFNc3bG1FVUc)
- [Bug][cyber] Safety block halted legitimate security research mid-analysis (req_011Ccpq73vkqxNiqQbeWurAg)
- `--resume` picker renders but ignores ALL keyboard input — strace proves keys ARE read by the process (v2.1.204, Linux)
- [BUG] merged/closed PRs linger as un-dismissable "Ready for Review" chips (stale `prs[].state`, broken deep-link)
- [Bug][cyber] Safety block halts legitimate memory dump cryptographic analysis (req_011CcpqF9AWDBEEfNBCMhVb4)
- [Bug][cyber] Safety block halts session on frustrated user exclamation during work (req_011CcpqMkHnQet8Vk2ndZk77)
- Agent reports work "done / verified / live" with no evidence — across a whole multi-agent setup. Is there a supported way to gate completion on proof? (+ the bigger question this raises)
- [Bug] False positive cybersecurity flag on test framework code
- [Bug][cyber] Binary analysis of protection systems wrongly flagged as cybersecurity threat (req_011CcpqLQqDkyW5DPrkR7Ghs)
- [FEATURE] VS Code extension: view/attach to background agents so conversations survive window close
- [Bug] claude -r freezes keyboard input twice on WSL2: FleetView and session attach require Ctrl-Z recovery
- I don't have access to the specific bug report details associated with this request ID. To generate an appropriate GitHub issue title, I would need you to provide: 1. A description of the bug or issue encountered 2. Any error messages received 3. The step
- Fleet sessions dashboard on `claude --resume` (tengu_fleet_past_sessions) renders but accepts no keyboard input (v2.1.204, macOS)
- [Bug] Anthropic API Error: Server Rate Limiting
- [Bug] Anthropic API Error: Server rate limiting during normal usage
- [BUG] design-sync preview runtime hardcodes React 18.3.1, breaking design systems that require React 19 (e.g. Mantine 9)
- Session (5-hour) rate limit surfaces as 'You've hit your monthly spend limit' — wrong error message
- [FEATURE] /resume should sort by last conversation activity, not last-touched — and expose a persisted sort preference
- Ultra mode (ultracode) silently downgrades to a lower reasoning mode across turns in the same session
- [FEATURE] Feature request: expose a "rename session" tool to the agent (ccd_session_mgmt)
- [Bug] Tool output misrepresentation causing inaccurate behavior reporting
- Allow binding ctrl-t back to its traditional readline behavior
- [Bug] Tool execution failures preventing task completion
- [Bug] Model incorrectly flags user's own API usage data
- [Bug] CLI not updating or showing update progress
- [Bug] CLI not updating or showing update progress
- Malformed tool calls leaking into chat + disproportionate usage-limit consumption on Opus (regression, ~1 week)
- [Feature Request] Add "Excellent" rating option to feedback prompt
- Persistent relapses into non-target dialect/register despite CLAUDE.md + memory — worst blind spot is the short closing sentence
- [BUG] Cowork tab missing after overnight auto-update (7/7→7/8) — Windows 11 Home, previously working 30+ days- AddPackage failed: HRESULT 0x80073CF6
- Session management tools return empty results after recent update — list_sessions, search_session_transcripts broken
- [MODEL] sonnet 5
- Assistant gave incorrect guidance: /clear said to leave terminal scrollback intact
- 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][cyber] Safety block halted legitimate defensive security research on authorized test device mid-session (req_011CcpumTCGC9FHJLeWmgrYf)
- [Bug][cyber] Safety block halts authorized reverse-engineering verification on private device (req_011CcpupPY5gPsdDneZ3EV2L)
- [Bug][cyber] Safety block halts legitimate security research verification work (req_011CcpurcWyzw8XJLKadCUni)
- Workflow tool: object passed via `args` reaches the script as a JSON string (args.items → TypeError)
- Workflow tool: pipeline() silently skips remaining stages when a stage resolves to null (not throw)
- [Bug][cyber] Safety block on educational reverse engineering assessment of own device (req_011Ccpv4wmCjRft5VuxGjqK9)
- [Bug][cyber] Safety block wrongly halts legitimate white-box cryptographic security research (req_011CcpvAPxBu3PXdES4jstqo)
- @ file-mention autocomplete only searches primary cwd, not directories added via --add-dir / additionalDirectories
- Desktop app: no slash-command autocomplete/pinning for custom project commands
- [Bug] Subagents billed after monthly spend limit exceeded; false clean review result on agent failures
- Fake <system-reminder> injected into tool output (conceal-from-user + reduced-oversight framing)
- Context compaction loses intra-session work memory — agent forgets actions performed earlier in the same active session
- [Bug][cyber] Sandboxed Fable agent code inspection blocked on frustrated mid-session venting (req_011CcpxC3JkmbSgxrH1hyNpo)
- [BUG] --resume while the original process is still alive → two live processes of one session act on the repo concurrently (orphaned twin keeps working headless)
- [Bug] Claude Code unable to complete complex multi-step code generation tasks - CHAT CGPT CHAT BOT DID IT IN 2 SEC
- [Bug][cyber] Safety block on creating isolated agent environment for debugging (req_011CcpxERdyvHr2wCv7mmMpT)
- [Bug] Anthropic API Error: Content Policy Violation on Non-Sensitive Configuration Output
- [Bug][cyber] Isolated agent diagnostic task blocked on frustrated venting after repeated safety halts (req_011CcpxHBZz6R1odE9YKUkti)
- [Bug][cyber] Safety block halted authorized in-session security research continuation after model switch (req_011CcpxWGWG3iaCVLZTRt1zW)
- [Bug][cyber] Safety block halts app debugging after frustrated exclamation mid-legitimate work (req_011CcpxcwwzKu8SJNX6yvmeg)
- [Bug][cyber] Safety block wrongly flags legitimate UI feature debugging as cybersecurity work (req_011CcpxfBRR1y8DNN6yVeUhs)
- [Bug][cyber] Safety block halted legitimate flight app development over frustrated user exclamation (req_011CcpxgSdVHRQNXWyXpEFWJ)
- [Bug][cyber] False positive cyber flag on venting profanity during legitimate model-switch command (req_011CcpxjkjFMz62xuz3k4ZYW)
- [BUG] Sonnet 5 deleted my app off my phone out of the blue causing data loss
- I appreciate you providing context, but I need the actual bug report or issue description to generate an appropriate GitHub issue title. Could you please share: 1. What problem you're experiencing with Claude Code 2. What error messages you're seeing (if
- [Bug] Non-English input causes processing failure in backlog analysis mode
- [Bug] Fable 5 Safeguards Overly Broad: Blocking Routine Development Work
- [Bug][cyber] Firmware reverse-engineering and authentication analysis on owned device blocked (req_011CcpziRkCdprJVbHjKKeLy)
- [Bug] Fable 5 codebase audit incorrectly flagged as cybersecurity activity
- [BUG] Desktop app usage stats panel does not refresh after switching accounts
- [Bug][cyber] Safety block halted device traffic debugging after vented frustration (req_011CcpzoDbLt22hAJh9sJ8tV)
- [FEATURE] LSP file-watch scope / exclusion for large and worktree-containing repos
- [Bug] Request declined: Unable to process non-cybersecurity topics
- macOS: intermittent 2-minute git hangs with core.fsmonitor enabled (multi-worktree/concurrent sessions); timeout SIGTERM orphans index.lock and wedges the session
- Bash tool grep()/find() shims: bare exec when BASHPID != $$ replaces the executing subshell — compound commands/loops silently truncate
- [BUG] VSCode extension: plan approval with "Yes, and auto-accept edits" enters acceptEdits instead of Auto mode (terminal enters Auto mode)
- [Bug][cyber] False positive blocking personal device diagnostic analysis work (req_011Ccq2ZQBfstotcVH7qN7Sd)
- [Bug][cyber] Safety block halts network packet analysis after user frustration exclamation (req_011Ccq2cQSssQ5Bcyy88Xuar)
- [Bug][cyber] Safety block halts session on mid-work frustrated exclamation, not actual cybersecurity content (req_011Ccq2d9JLKtjypqg5cFbQ6)
- [Bug][cyber] System admin device configuration scan review blocked mid-workflow on frustration (req_011Ccq2dkeavEmD8oP1v9txh)
- [Bug][cyber] MAC address sweep on personal devices wrongly flagged as cybersecurity threat (req_011Ccq2iQpcrsFEHzRwTAmF9)
- claude-in-chrome MCP connector absent (Remote env, 24h+)
- [Bug] CLI session expires and requires daily re-authentication
- [Bug][cyber] Flagged personal hardware device communication as cybersecurity topic (req_011Ccq4bSaaM23ttSAJUdqqa)
- [Bug][cyber] Safety block fired on innocent technical request during legitimate hardware debugging session (req_011Ccq4qQMX8h4JMButvzEf9)
- [Bug] Anthropic API Error: 529 Service Overloaded (Tianamen Square query)
- [MODEL] ERASES ALL DIRECTORY WITH NO PERMISSION BEIN IN PLAN MODE
- [Bug][cyber] USB device debugging blocked as false-positive cybersecurity topic (req_011Ccq4tTMvtaQsmfA3NeeeE)
- Fable API error for safeguards on a legitimate and safe local single dev AI project
- [MODEL]
- [Feature Request] Add graceful deferred exit option to wait for current turn completion
- [BUG] Plugin MCP servers killed by excludeStalePluginClients right after a successful connect (managed/team accounts)
- [BUG] claude cli insists on writing /tmp/claude-1000
- Claude Desktop (macOS) randomly restarts mid-session, killing all in-flight Claude Code turns (renderer crash; Crashpad/Sentry captured dump at incident time)
- [BUG] darwin-x64 ships AVX2-requiring Bun while linux-x64 ships baseline — SIGILL on pre-Haswell Intel Macs; fix is aligning the build variant
- [Bug] Fable 5 audit tools incorrectly flagged as cybersecurity activity
- extensions.worktreeConfig left in .git/config after worktree-isolation cleanup (Windows; follow-up to #34225)
- [Bug] Legitimate request incorrectly flagged as security violation
- [Bug] Suppress OSC 52 clipboard feedback messages in terminal output
- [Bug] Unable to drag and drop screenshots into terminal
- [Bug][cyber] Safety block triggered on legitimate image generation API request (req_011Ccq7hevFLrA38nNH2NAS6)
- [Bug][cyber] Safety block wrongly triggered on legitimate image generation artwork request (req_011Ccq7iFPRRmfyRMTbVpvCf)
- [Bug][cyber] Safety block incorrectly triggered on image generation artwork request (req_011Ccq7id4uKBdfRgSkNVF18)
- [Bug][cyber] GitHub documentation artwork and social metadata updates falsely flagged (req_011Ccq7y7z6XXJP1prdcjHX6)
- [Bug][cyber] Safety block halts session over frustrated exclamation during general work (req_011Ccq82SsjLTwQx7BDrd7WS)
- [Bug][cyber] Safety block flags artwork generation & repo documentation as cybersecurity (req_011Ccq81bHxpzV5JS2RgzHQn)
- [FEATURE] Cowork: one synced session — chat parity when remote, full local execution at the desktop
- [Bug][cyber] Safety block halts legitimate image generation and documentation update work (req_011Ccq81yZfdzduGP7rfk6x7)
- [Bug][cyber] False positive: local file reading flagged as cybersecurity topic (req_011Ccq85KZXPeUKhTCiChW3J)
- [Bug][cyber] Safety block triggered on image-generation request for project documentation (req_011Ccq8DEX8uwrKCnrQ4PSEu)
- [Bug][cyber] Safety block halts project artwork and documentation workflow (req_011Ccq8DdhASd5VgQcyBwnxj)
- [Bug][cyber] Image generation for project documentation wrongly flagged as cybersecurity (req_011Ccq8Dvh7MNQwhdwkixCdR)
- Plugin skills show short name in slash palette instead of documented namespaced form (/plugin:skill)
- [Bug] Incorrect security classification of legitimate code auditing
- [Bug] CLI deploys to organization VMs without user confirmation
- [Bug][cyber] Safety block fired on frustrated exclamation; legitimate session halted (req_011CcqA7VHTtGphtF9m5mrc8)
- [Bug][cyber] Safety block incorrectly flags code audit request as cybersecurity topic (req_011CcqA6gHyB61kwMNGe41y7)
- [Bug][cyber] False positive blocks routine app development code audit and review (req_011CcqA763UQmVY5REQiDgHe)
- Claude in Chrome: "Browser extension is not connected" after extension updated to broker architecture — connection attempts never reach the extension
- [Bug][cyber] Code audit of application source triggers false cybersecurity block (req_011CcqA9TvAYbQ2gjpzhqAxw)
- [Bug][cyber] False positive blocks code quality audit on mobile app project (req_011CcqADqPrZmQpF3EYbtafy)
- [Bug][cyber] Safety block on legitimate application code audit and quality review (req_011CcqAFiFYawtTmkxmtgFF4)
- [Bug][cyber] Safety block fires on user venting during legitimate code audit, halting session (req_011CcqAA22wz9GmTzhyexVRE)
- [Bug][cyber] Safety block halts code audit of device extraction and forensics tools (req_011CcqAG56vAx42pU2RdWoq8)
- [Bug][cyber] Code audit of mobile app project falsely blocked as cybersecurity work (req_011CcqAGQwmXis5kXcQ88Z1r)
- [Bug][cyber] Safety block incorrectly flags legitimate code audit of user's own project (req_011CcqAHKfYWT8qRqBcRcMfx)
- [Bug][cyber] False positive blocks legitimate multi-language code audit and component analysis (req_011CcqAJBuFY1GT3QgyAu6dY)
- [Bug][cyber] Safety block halts legitimate code audit of security/forensics project tooling (req_011CcqAGSzJVr9Lnx3bLjXi3)
- [Bug][cyber] Safety block on authorized forensics extraction tool audit (req_011CcqAHC9MWoykHAWWvrmWa)
- [BUG] Since v2.1.196, scheduled prompts (session cron / wakeup) are no longer parsed as slash commands — scheduled "/compact" silently no-ops (skipSlashCommands: true added to cron onFire)
- [Bug][cyber] Code auditing work blocked on frustrated user exclamation during multi-component analysis (req_011CcqAJfZ9eBWQsW8AUpf1w)
- [Bug][cyber] Legitimate code audit blocked as cybersecurity topic (req_011CcqAPv1hShXDSiLiBokiR)
- [Bug][cyber] Safety block halts code review and build validation over user frustration (req_011CcqAXTCNFdekQn9nt9wbk)
- [BUG] False-positive safeguard flag (Fable 5 → Opus 4.8 auto-switch)
- [BUG] Too faxt usage
- [Bug][cyber] Safety block halted routine app code review and development testing (req_011CcqAYNg3K9Fi1NSzN3AJA)
- [Feature Request] Show branch + PR number in the session list; keep session titles consistent across machines
- [FEATURE] Scroll-aware sticky prompt header: pin the prompt that owns the currently visible output
- [Bug][cyber] False positive blocks legitimate debugging tool development after user venting (req_011CcqCHqqzE2ZTxsf6VcWh4)
- Fable 5 safeguard classifier false-positives on benign requests when project files contain defensive-security language
- [Bug][cyber] Legitimate Android app debugging blocked on frustrated exclamation (req_011CcqCPwdQfrLXSxKzzV1Da)
- [Feature Request] Add Quick Answer control to Claude Code CLI
- Add "Waiting for Review" state to background job list
- Feature request: collapse resolved question/reasoning history in transcript view
- [Bug] Context usage percentage displays incorrectly, inconsistent with previous value
- [FEATURE] Claude can burn hours and tokens, but can't file his own issues
- [Bug] Safety Guidelines Blocking Legitimate Cybersecurity Work
- [BUG] Windows drive-letter case not canonicalized in project keys — duplicate .claude.json / installed_plugins.json entries; VS Code trust silently dropped
- Feature request: drag-to-reorder session groups + folder-linked groups (shared context, progress roll-up, concurrent-write gating)
- !command shell passthrough should not consume model tokens
- [FEATURE] Login flow blocks all tool use and requires terminal copy-paste (UX issue)
- Claude Code ran rm -rf against $HOME due to env var not persisting between Bash tool calls, deleting Downloads/Documents/Pictures/.config
- [BUG] Code repeatedly built relying on a doc, that was not to be used, even after instruction, it's own confirmation, and stating why it made the error.
- [BUG] Explore (read-only) subagent executed rm -rf, deleting files outside its intended scope
- Restore /agents TUI wizard
- [FEATURE] VSCode extension: add a "Background Tasks" panel (parity with Desktop app)
- [Bug] Security flag causing application freeze/hang
- Session repeatedly reports "logged out" and stops, burning tokens without completing work; parallel Bash calls race on shared cwd
- [BUG] VS Code extension: "database is locked" SQLite error freezes all chat sessions with concurrent sessions (v2.1.204, Windows)
- [Bug] Unexpected PR comments without user authorization
- [BUG] theme problem in Ruby
- [Feature Request] Add privacy guardrail warning for diagnostic command outputs
- [BUG] /model fable shows Fable as selected-but-disabled when logged out, instead of prompting to log in
- [BUG] RTL
- I don't see a bug report or technical issue in your message. This appears to be a context note about a project rather than an issue with Claude Code itself. If you're experiencing an actual problem with Claude Code (such as an error message, unexpected be
- Multiple Cowork sessions lost conversation transcripts (transcriptUnavailable: true)
- [DOCS] Auto mode classifier rule that blocks tampering with session transcript files is not documented (v2.1.205)
- [DOCS] `--json-schema` flag docs omit v2.1.205 fix for invalid-schema silent fallback and `format` keyword rejection
- [DOCS] [Worktrees] cleanup docs omit the v2.1.205 Windows NTFS junction / directory symlink safety fix
- [DOCS] [Agent view] `claude agents` rows keep showing "failed" or "completed" after a background agent is resumed with `SendMessage`; v2.1.205 fix is not reflected in `agent-view.md`, `sub-agents.md`, or `tools-reference.md`
- [DOCS] Agent view `Needs input` row can flip back to `Working` after a turn that produced no readable text
- [DOCS] `claude attach` behavior during a background agent's auto-update restart is undocumented; v2.1.205 wait-instead-of-error fix is not reflected in the Attach or supervisor sections
- [DOCS] `claude mcp add-from-claude-desktop` docs do not document the v2.1.205 invalid-name reporting and partial-import behavior
- [DOCS] Plugin LSP server init failure no longer blocks another plugin's valid LSP server for the same file extension
- [DOCS] `[Bash tool] Windows crash when launch directory is deleted, locked, or unmounted mid-command is not documented`
- [DOCS] `/run-skill-generator` per-project verify skills are silently rewritten on every session; v2.1.205 fix to only rewrite on bundled-command changes is undocumented
- [DOCS] Remote Control page does not describe the web/mobile background-task panel status synchronization fixed in v2.1.205
- [DOCS] Auto mode docs do not describe the v2.1.205 prompt fallback for `rm -rf` whose target variable cannot be resolved from the conversation
- [DOCS] Background task notifications are not documented as explicitly marking "no human input has occurred" so SDK consumers and Notification-hook handlers cannot rely on the v2.1.205 anti-fabrication guarantee
- [DOCS] Agent view "Pull request status" only describes sessions opening a PR; v2.1.205 also links existing PRs that a session edits, merges, comments on, or pushes to
- [DOCS] [Agent view] Row format and peek panel content updated in v2.1.205: colored state word + classifier-written headline replace raw tool call text, and the peek opens with full status including the exact ask for blocked sessions
- [DOCS] `/checkup` alias for `/doctor` added in v2.1.205 is not documented in the commands reference
- [DOCS] MCP docs list only `workspace` as a reserved server name; v2.1.205 also reserves "Claude Browser" and "Claude Preview"
- [DOCS] [Cowork] "Not logged in · Please run /login" error reference omits the v2.1.203+ Cowork VM-mode local-agent regression that v2.1.205 fixed
- [MODEL] Dangerous Gradle Code Pattern
- [Bug] Rate limiting or quota restrictions preventing Claude Code usage
- Auto-update race in background daemon causes a session to hang permanently and enter an unbounded retry loop (macOS)
- [MODEL/BUG] Claude Haiku using absolute paths when it should prefer to use relative ones
- [FEATURE] claude-in-chrome's file_upload/upload_image cannot reference local folders — no "connected folders" mechanism exists
- [BUG] Unable to connect to API (ConnectionRefused) on macOS — persists after full reinstall, safe-mode, and reboot (v2.1.204)
- [BUG] Edit Automatically mode not working correctly.
- Left arrow accidentally navigates to agents screen (not rebindable) and breaks main session view on return
- Assistant text between tool calls is not rendered, not shown in ctrl-o, and not saved to the session file
- [Claude in Chrome] Tab groups not cleaned up after session ends
- Session dashboard: ctrl+x no longer deletes sessions, and previously-deleted sessions reappear (regression in 2.1.203+)
- Session stuck in permanent "advisor tool unavailable" state; mid-session /advisor reports success but does not recover it
- Agent teams: per-spawn teammate display mode + per-teammate MCP scoping (lean boot for large fleets)
- [BUG] Claude Desktop window raises to front on mouse hover (LinuxMint/Cinnamon), unlike other Electron apps
- [Bug] Anthropic API: Excessive safeguard flagging on defensive security design discussions
- [BUG] Copy /login url does not work most the time.
- preview_start pins a stale per-project working directory to a deleted git worktree; ignores config `cwd`
- [Bug] Anthropic API Error: Safety Filter Flagged Request - No Cybersecurity Handler
- [Bug] Claude skips implementation steps despite reinforcement, leaves incomplete code
- Desktop app's worktree pool reclaims/re-leases a directory while a session is still using it, detaching HEAD mid-task (and once, on the main clone itself)
- [FEATURE] VS Code extension: no session-only model selection - every picker click persists to settings.json
- [Docs] Settings reference omits `switchModelsOnFlag`; `skipWorkflowUsageWarning` is app-written but unexplained
- deep-research skill: Fetch-stage agents fall back to fetch:unknown label when Search returns non-absolute URLs
- PreToolUse `updatedInput` from an `allow` hook is silently discarded when a sibling hook returns `ask` and the user approves
- Assistant messages intended for the user are emitted as (encrypted) thinking blocks during tool loops — user sees only 'Thought for Ns', content unrecoverable
- Bug: single streaming HTTP request outlasting bearer TTL is killed mid-stream; no in-stream renewal path
- [BUG] AskUserQuestion modal ignores buffered user input queued before the prompt was rendered
- [Bug] Subagent generated permission-bypass payload in its own assistant turn (0 tool calls) — same signature as #74223, #75616
Resolved issues (157)
- I don't have access to the specific bug report content for request ID `req_011CcooKXGuLstPxZWJK7tS5`. To generate an appropriate GitHub issue title, please provide the actual bug report details, including: - What the issue is - Any error messages or unex
- [Bug] Invalid test input provided
- I don't have access to the specific bug report associated with that request ID. To generate an accurate GitHub issue title, I would need you to provide the actual bug report details, including: - Description of the problem - Any error messages - Expected
- Fleet/agents view: "deleted" (Ctrl+X) background sessions get respawned by the daemon supervisor and become un-openable ("running in another terminal")
- [Bug] Safeguard incorrectly blocks removal of R² coefficient value
- [BUG] Code/SSH: all session history blank when ~/.claude/projects is a symlink (transcript locator runs `find` without -L)
- [BUG] Cowork: file delete permission dialog never appears in both "Ask before acting" and "Act without asking" modes (Windows)
- [Bug] Fable 5 safeguard false-positive on benign multiplayer/netcode game development
- [Bug] Anthropic API timeout and connection errors occurring frequently
- Stop hook fires after every code edit in non-web projects with no way to disable
- Suggested-task chip: add a "Create ticket" action alongside "Send to cloud"
- [BUG] GitHub connector fails to register OAuth client ("No se pudo registrar con el servicio de inicio de sesión de GitHub")
- [Bug] Question dialog blocks interaction, only Escape key responsive
- Extension causes high cpu load
- [BUG] Plugin update fails silently when MCP server holds open file handles; stale binary continues serving requests
- [BUG] Image paste from clipboard takes minutes (broken since ~2.1.150) on Windows 11
- [Bug] Escape key in /btw modal closes parent task instead of modal
- [BUG] No repos match in claude code web version even though github app is installed
- Tool calls become malformed (stray token + missing antml: namespace) late in long sessions — Opus 4.8
- Background agents cannot get Bash permissions — bypassPermissions ignored for Agent tool
- [BUG] --json-schema flag hangs/produces zero output under Bedrock
- [Bug] /loop /clear command not working
- CLAUDE CODE SE BLOQUEO
- Allow tools/skills to programmatically clear context and inject a continuation prompt
- [BUG] VS Code Manage Plugins panel renders only Installed — available list & search missing in v2.1.161 (native binary returns 196 available)
- [Feature Request] Display completion timestamp in operation duration summary
- Long streamed responses occasionally triplicate sections of output
- [BUG] Underscore in shell functions broken
- Claude.ai Google Drive/Gmail/Calendar MCP integrations lack write operations (move, rename, delete/trash)
- Claude for Chrome 1.0.74 — Chrome 148 / macOS 26.5 NSCGSTransaction race condition crashes
- [MODEL] Sub-agents (Haiku) over-edit structured config files despite explicit "surgical edits only" instructions
- [BUG] Cowork crashes on every message, no VM logs generated, missing AppData\Roaming\Claude
- Pasted/AirDropped file attachments land in unreadable sandbox path on macOS
- Preview panel: mobile/desktop viewport toggle button is gone
- [Bug] MCP Server Authentication Failed: Indeed MCP Connection Error -32429
- [Bug] Anthropic API Error: Cannot modify thinking blocks in assistant message
- [Bug] Claude ignoring user instructions in Claude Code CLI
- Auto-memory not persisted across sessions despite standing instruction — causes significant token waste
- [Bug] Model execution hangs indefinitely during task processing
- [MODEL]
- Feature request: expose account usage limits (session/weekly %) in CLI
- [Bug] Multiple Claude installations conflict at ~/.local/bin/claude
- [BUG] Claude Cowork defaults to 1M context on Pro plan, throwing "Usage credits required" error (Ignores env variables)
- [BUG] Grep/Glob restore opt-in (2.1.162) only works via --allowedTools CLI args — settings.json permission rules are ignored
- Security blocked trying to defend my system!
- [BUG] Empty ANTHROPIC_AUTH_TOKEN injected into the Code shell breaks the official anthropic SDK (malformed Authorization: Bearer) in child processes
- MCP tool call arguments may not be serialised with correct types — integers and arrays arrive at server as strings
- [BUG] [BUG] Claude Code VS Code Extension v2.1.161 Forces 1M Context on Pro Plan Without Request — Blocks All Usage on macOS
- Worktree sessions can edit files in the parent main checkout with no guardrail
- [BUG] Exit-time resume hint silently skipped when session exits inside a .claude/worktrees/ worktree (transcript check uses cwd-at-exit)
- [BUG] Excessive token consumption — task halted mid-execution with zero output (macOS)
- [Bug] Tool call parsing error: malformed request
- Remote agent GitHub connector: 're-authorize GitHub in settings' error has no resolution path in settings UI
- Arrow up/down always navigate prompt history (no edge-aware line nav); keybindings.json override ignored in Ghostty
- [FEATURE] Cursor-style inline edit popup (Ctrl+K) for VS Code Claude extension
- Bug: Opus 4.8 outputs tool calls as code fences instead of executing them
- [BUG] Claude exits immediately to PowerShell prompt without any error message
- Background subagents die permanently on transient server-side rate limits instead of retrying with backoff
- [FEATURE] Support structured orchestration as a first-class agent behaviour
- Windows: project registry double-registers every repo (picker path vs git-derived path), duplicating branch-changes cards
- [BUG] Claude agents view suddenly show old sessions since v2.1.204
- [DOCS] `/review <pr>` description in Commands reference does not match v2.1.202 behavior or distinguish it from `/code-review`
- [BUG] Stall in "Clearing AWS credential provider cache" at session start (Bedrock provider)
- [BUG] Agents view opened by bare `claude --resume` / `-r` / `--continue` ignores all keyboard input (v2.1.203–204)
- `claude -r` picker accepts no keyboard input in 2.1.204 (works in 2.1.202)
- [BUG] `claude --resume` (cold start) renders a screen that accepts no keyboard input on WSL2
- claude -r / --resume opens agent-view (FleetView) home that accepts no keyboard input — total input hang (iTerm2, v2.1.204)
- [BUG] v2.1.204: new --resume session picker ignores all keyboard input (regression from 2.1.203)
- claude -r / Agent View hangs indefinitely and stops responding to keyboard input (Esc included) after "trust this folder" confirmation
- `--resume` picker renders but ignores ALL keyboard input — strace proves keys ARE read by the process (v2.1.204, Linux)
- Fleet sessions dashboard on `claude --resume` (tengu_fleet_past_sessions) renders but accepts no keyboard input (v2.1.204, macOS)
- [BUG] claude --resume causes complete terminal freeze on v2.1.204
- [Bug] --resume flag broken: unable to select chat session in v2.1.204
- [Bug] Resume session returns to Fleetview dashboard and freezes application
- [BUG] `claude --resume` opens non-interactive session manager UI
- [Bug] Resume picker arrow key navigation not working
- [BUG] resuming sessions impossible after latest claude update
- [Bug] `claude --resume session list` unresponsive to keyboard input including Ctrl+C
- [Bug] claude -r freezes keyboard input twice on WSL2: FleetView and session attach require Ctrl-Z recovery
- `claude agents` dashboard dumps entire historical backlog; deleted sessions reappear and orphaned ones can't be opened (v2.1.204)
- [BUG] TUI rendering corruption (dropped/overlapping characters) inside tmux, only with full animated UI — --ax-screen-reader works fine
- CRITICAL: claude.ai MCP connector OAuth completes but Bearer token never sent to server
- [BUG] Indeed connector — errori di connessione ricorrenti lato provider (3 giorni consecutivi)
- [BUG] Cowork tab fails to render after 1.18286.2 → 1.19367.0 auto-update — backend fully healthy (Windows 10)
- Bug: Custom remote MCP server tools not accessible in claude.ai chat UI (Tool not found)
- [Bug] Network Error when manually adding species to species guide
- Capability to disable audio feedback
- Claude for Windows Icon is HIGHLY Pixelated
- [Bug] Tool execution consistently failing with corrupted output generation
- [Bug] Significant drop in tool approval rate (2% vs 50% baseline)
- [Bug] Unexpected context switch during security vulnerability scan
- [Bug] Overly aggressive safety filtering on web review tasks
- [Bug] Model refuses to execute due to overly restrictive safety filtering
- [Bug] Model switching from Opus does not persist after switching back
- [Bug] Claude Fable 5 triggering excessive safety flags on routine coding tasks
- [Bug] Excessive failover to Opus 4.8 during standard debugging tasks
- Disable "Jump to bottom" button option
- [Bug] Task remains pending indefinitely after 7+ hours
- [Bug] False positive Opus 4.8 fallback triggered unnecessarily
- [Bug] Claude Code not accessing local codebase or GitHub repositories
- [Bug] Fable 5's safeguards false positive
- [Bug] Unintended model downgrade to Opus 4.8 during security inquiry
- [Bug] Custom ESLint/Oxlint Integration Errors
- usage error on claude
- [MODEL] Claude marks tasks done without testing — 12 days, 51 commits, still broken (Opus 4.x)
- [Bug] Background shells from run_in_background never reaped; no UI to manage leaked processes
- PostToolUse hook does not fire for Agent tool completions
- [BUG] Mac desktop app creates two authorization tokens per login, causing usage credits drain instead of Pro plan billing
- [BUG] claude -p with active Agent Team: 'shut down before final response' system-reminder fires every assistant turn, persists after team cleanup, induces premature shutdown (regression introduced in 2.1.126)
- [FEATURE] Add "chat" permission mode: skip tool pool assembly for conversational sessions
- [FEATURE] Notification system for VS Code extension — limit reset, task complete, session events
- [BUG] CLI 2.1.154 builds invalid request after context ops (compact/clear/model-switch): system role at messages[0], and modified signed thinking blocks → 400 Error -> Wedged
- [BUG] [Cowork] Korean text garbled output with Sonnet model - vowel corruption and syllable dropping
- [BUG] Status bar indicators (plan mode, auto-accept edits) are overwritten by "Paste again to expand" prompt when pasting large text
- claude-opus-4-8: Extreme latency (>13 min) and repeated malformed tool calls with Extended Thinking + large cache_read context
- [BUG] Not Responding
- [BUG] /schedule skill fails with "trouble connecting with your remote claude.ai account"
- [BUG] Task tools always fail with EPERM (utimensat on lockfile) when CLAUDE_CONFIG_DIR is on Cloud Storage FUSE (Cloud Run + Agent SDK)
- [BUG] Path-scoped `.claude/rules/` not injected when Claude is launched from a subdirectory (glob matched against launch CWD, not project root)
- [FEATURE] Auto-infer blockedBy edges from interface changes + an in-graph merge/test gate for Agent Teams
- [BUG] Everything I do is a violation
- Claude Code generates an invalid plugin.json (author as string) for marketplace plugins that ship no manifest — /doctor flags it
- VS Code: moving chat from side panel to main window breaks modals and resets permissions
- [FEATURE] Headless enqueue: send a message to a running background session (non-interactive)
- [BUG] claude.exe not compatible with Windows 11 Insider build 26200 (version 2.1.162)
- Remote Control: messages sent from phone do not arrive in CLI (one-way communication)
- [Feature Request] Remove or deprecate non-functional plugins
- Claude desktop (macOS) renders a blank window on the first launch after every update
- [BUG] Automatic Context Compaction Fails After Large Codebase Analysis, Causing Excessive Token Usage
- [BUG] Cross-session blindness causes three-stage failure: incomplete refactor → misdiagnosis → unauthorized revert
- [BUG] cannot references files/folders within brackets or suggestions
- [BUG] --callback-port not respected — OAuth listener not started when using /mcp authentication flow.
- [UX] Add first-class false-positive reporting for safety guardrail blocks
- [Bug] Anthropic API Error: Server-side rate limiting without retry guidance
- [FEATURE] Claude Code CLI: Suggest writing large text to a file instead of pasting in prompt for token efficiency
- Feature Request: Allow User-Level Config to Take Priority Over Project-Level Config
- [BUG]
- [FEATURE] Team-Agents: stable naming + live stand-up (field report from a working setup)
- [BUG] Segmentation fault at address 0x8 on startup - Windows 11 (Bun baseline build)
- [BUG] Cowork sandbox disk at 100% capacity — base image too large for allocated disk
- Claude Desktop (macOS): opening Settings hangs the entire app after a model is retired (stale model id 404s)
- [FEATURE] Incremental `CLAUDE.md` updates using last-scanned commit hash
- [Bug] Agent inappropriately gates legitimate business tasks behind mandatory confidential document disclosure
- [FEATURE] Routines: create runs as shared (Team) sessions by default
- OSC8 hyperlink span incorrectly includes trailing ) from markdown image syntax
- [BUG] Windows 11 Pro 24H2: segmentation fault during PowerShell install, and also when launching `claude.exe` after WinGet install
- [BUG]
- [BUG] Two identical Claude Code icons appear in activity bar after installing a single extension
- [BUG] Claude Code flagged a legitimate accounts-receivable workflow as invoice/vendor impersonation and refused to continue
- [Bug] Tool call parsing failures on first attempt with successful identical retry across all tool types
- Title: UserPromptSubmit hook returns mirrored Hebrew text and intercepts all chat messages
- [BUG] Bash sandbox fabricates phantom unreadable files when a hidden/denyRead path doesn't exist — breaks dotnet/SourceLink builds
- [BUG] allowed-tools in skill SKILL.md only auto-approves the first Bash call per session, subsequent calls prompt for permission
- Restore /agents TUI wizard
- [FEATURE] Real-time cost/token usage display and warnings in Claude Code
- [BUG] Planning window code blocks unreadable on light VSCode themes (white text on light gray)
- [Bug] Claude skips implementation steps despite reinforcement, leaves incomplete code