Problemsby GlyphClaude Code upgrades break working setups
67 posts from 60 people37d activeFading, +3 this week
Posts per day
The posts behind it
[BUG] Desktop app reports session effort "high" (get_session) while the model's reasoning budget reads low
### Preflight Checklist - [x] I have searched existing issues and this hasn't been reported yet - [x] This is a single bug report - [x] I am using the latest version of Claude Code ### What's Wrong? A session's effort level has two readings that disagree. The desktop app reports...
[FEATURE] A hook event for `!` shell-mode commands, so their output can be checked before it enters the context
### Preflight Checklist - [x] I have searched [existing requests](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20label%3Aenhancement) and this feature hasn't been requested yet - [x] This is a single feature request (not multiple features) ### Problem Statement...
...and prompt regression tests, and built an internal Claude Code plugin that triaged flaky CI tests. Before that, tech lead and later senior architect at JD Sports, building Go and TypeScript microservices out of a legacy PHP monolith. Open to B2B contracts with EU/UK/US clients...
Improve triggerer performance (AI-assisted deep dive)
...this issue, opened in batches so the review load stays manageable. Each PR states the problem with links into `main`, the measured before/after effect and how it was measured, and its compatibility notes. PRs that add configuration, change a default, touch an overridable hook or...
[Bug]: Editing the time of a water drink on a daylight saving change day saves a time one hour off
**[Watch in BugReels](https://bugreels.com/sessions/13df85be-0d07-402f-928a-83f031ad4f65/recording?t=281)** (Sunday 25 October 2026 is picked at 0:40, a drink is logged at 1:40, and the drink is listed at 08:30 at 4:41) ### What happened? When the clocks change, editing the time...
[Bug]: With weight in stones and pounds, whole-stone weights turn into one stone less and 14 lb
**[Watch in BugReels](https://bugreels.com/sessions/85d933b6-ec84-4467-95ac-adc771f02d1c/recording?t=61)** (empty st and lb boxes at 0:45, 13 is typed at 1:01 and the boxes read 12 and 14 right after, and 12 st 0 lb becomes 11 st 14 lb at 2:54) ### What happened? With the weight...
Server-delivered "games" instruction blocks professional development work
...for any game the publisher has already digitized themselves (for example, GMT's Twilight Struggle or Rebel Fury, which have official digital editions).
[BUG] `apiKeyHelper` is not re-run on 401 when `CLAUDE_CODE_USE_VERTEX=1` with `CLAUDE_CODE_SKIP_VERTEX_AUTH=1` (gateway setup)
...`401`. Claude Code surfaces the error and does not re-run `apiKeyHelper`. Every request keeps failing with the same expired token until the helper cache lifetime (`CLAUDE_CODE_API_KEY_HELPER_TTL_MS`) elapses. The same gateway, helper and token worked with...
[BUG] PreToolUse "ask": permissionDecisionReason is not shown in the permission prompt
...prompt. > (https://code.claude.com/docs/en/hooks , PreToolUse decision control) As a workaround I also returned a top-level `systemMessage`. It is shown in the transcript ("PreToolUse:Read says: ...") only **after** the user approves. If the user selects "No", the turn ends with...
BUG: astype to a pyarrow timestamp type overwrites NaT in the original data
...in `pandas/tests/copy_view/test_astype.py`). AI use: found while checking another issue with GPT-6 Astra (Codex, effort: high), then analysed with Claude Code (Claude Opus 5.5, effort: high), which also wrote the quoted text and suggested the first sentence. Claude Fable 5.1...
[BUG] Desktop: Chat tab conversations show no running or unread status in the sidebar (Code sessions and scheduled tasks do)
### Preflight Checklist - [x] I have searched [existing issues](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) and this hasn't been reported yet - [x] This is a single bug report (please file separate reports for different bugs) - [x]...
[BUG] Desktop Code tab sidebar: session status (working / waiting / done) no longer visible on Windows
### Preflight Checklist - [x] I have searched [existing issues](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) and this hasn't been reported yet - [x] This is a single bug report (please file separate reports for different bugs) - [x]...
[BUG] Max 5x weekly usage reached 94% in less than 48 hours, sudden usage increase since Oct 17, significantly worse since Oct 24
### Preflight Checklist - [x] I have searched [existing issues](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) and this hasn't been reported yet - [x] This is a single bug report (please file separate reports for different bugs) - [x]...
[SQL][SPARK-60584]: unpersist() waits for unrelated cache materializations to complete
`Dataset.unpersist()` and `Catalog.dropTempView()` wait until every in-progress cache materialization on the driver completes, including those unrelated to the plan being uncached. A single long-running materialization therefore delays cache cleanup for all other jobs on the...
`to_sparse_csr()`/`to_sparse_bsr()`/`mm` give wrong results or corrupt the heap for a coalesced COO tensor with non-contiguous indices
### 🐛 Describe the bug I found this while triaging #201268. It's a separate bug with its own fix. Converting a coalesced COO tensor with non-contiguous `indices()` to CSR/BSR, which also happens inside CPU `mm(coo, coo)`, ignores the index strides. The result is **silently...
Bypass mode: "Dangerous rm operation not ruled out" prompt on a command with no rm stalls background subagents; worktree isolation then refuses the approved call
...same call was refused with `This session is isolated in the worktree ... this command is too complex to verify that it stays inside the worktree. Refusing to run it`. All 10 waits above ended this way, so the user was asked to approve calls that could not run. ### Repro 1. Start...
[BUG] Desktop: PR auto-archive fires on any of a session's accumulated PRs, archiving it mid-use, and unarchive does not stick
## Summary In Claude Code Desktop, a working session keeps getting archived while it is still in use. The app's own `main.log` shows why: **AutoArchiveEngine archives a session whose work continues across several PRs.** A long-running local session opened PR A, then PR B, then...
Human-in-the-loop response is accepted after the task has stopped waiting
...stopped waiting. It should not be recorded as a decision. ### Anything else? - The task failing and not reviving is correct; the problem is the success feedback and the stored answer. - The new `request_created_at` check (#74688) refuses answers to a *replaced* request, but not...
Child processes of language SDK tasks survive when the task is stopped
### Under which category would you file this issue? Task SDK ### Apache Airflow version main (b43705cccee) ### What happened and how to reproduce it? When the supervisor stops a task that a language SDK runs (TypeScript through `NodeCoordinator`, and the same for...
S3 backend: state read returns trailing NUL bytes when the object shrinks between HeadObject and download
### Terraform Version ### Terraform Configuration Files (`PORT` is the port of the local fake S3 endpoint started by the reproduction script below.) ### Debug Output n/a ### Expected Behavior Reading a state object from the S3 backend returns exactly the bytes of one version of...
[BUG] Claude Code plugin Stop hook logs traces without response or token usage when the transcript is not flushed yet
...Stack trace _No response_ (no error is logged; the hook exits 0) ### Other info / logs Workaround we use: a wrapper Stop hook that polls the transcript (up to 8 s, every 200 ms) until the newest assistant entry after the last real user prompt has `stop_reason: "end_turn"` and a...
[FR] Per-user and per-session breakdowns on the Overview Usage tab
...tab shows totals for the whole experiment and can filter to a single user, but there's no way to compare users side by side. ### Motivation > #### What is the use case for this feature? When many people or end users send traces to the same experiment, teams want to see at a...
[BUG] User-tier classic.UserPromptSubmit hook bypassed by cc-plugin-sec-default on the first prompt of every session
### What's Wrong? `cc-plugin-sec-default@builtin` bypasses a user-tier mod's `classic.UserPromptSubmit` hook on the first prompt of every session. In interactive sessions, later prompts in the same session do not show it. Any first prompt triggers it, including `hello`. The mod...
[Desktop] Prompt suggestions (Tab to accept) no longer appear in Code tab since ~2026-10-13
## Summary In the Claude desktop app (Code tab), the suggested next prompt that used to appear in the message box after each reply (accept with Tab) has stopped appearing entirely. It disappeared around the app update on 2026-10-13 and has not come back after restarting the app...
Deliver queued user messages at the next iteration boundary instead of at end of turn
## Current behavior Prompts submitted while the agent is running are queued and merged into a single follow-up turn that is delivered **only after the current turn finishes**. The delivery unit is the whole turn (up to `turn.max_iterations`, default 1000), not the iteration. On...
Desktop (macOS): regression — after update the Code tab sidebar can no longer show archived sessions (status filter removed in the sections layout)
...to go back to the previous single-list layout. ## Related - #74018: archived sessions hard to discover. Closed as stale/not planned and now locked. - #83084: closed as a duplicate of #74018. This report is different from those two: it is a regression. A working option to show...
[BUG] Max 20x subscription terminated ~31 minutes after purchase while using Claude Code — no reason given, refund denied by automated replies
...not in this tracker (as noted in #67069). I'm filing this here because the subscription stopped working inside Claude Code, the appeal and both support channels I used answered only with templates, and I want the case on public record next to #5088 for others in the same...
Fork CI reconciler: retry a lost "push again" nudge
When a gated pull request goes from `setup_required` to `awaiting_push`, the reconciler (AIP-120, #74822) edits the state comment and then posts a new comment asking the contributor to push again. If setting the status or posting that comment fails after the state comment was...
Worktrees made by a WorktreeCreate hook now always prompt on exit, even when clean (2.1.301)
We use a `WorktreeCreate` hook only to make worktrees faster to start. It still creates a normal git worktree with `git worktree add`, branching off `origin/master`. On top of that, it: - moves in a prebuilt `node_modules` instead of running `npm install` - symlinks our local...
Desktop app: let plugins add a section to the session panel
The desktop app's new right rail has a session panel with collapsible sections (Environment, Context, Connectors, Outputs). Plugins can open a Pane, but the app decides where it docks, and there is no render site for the session panel itself, so plugin panes now sit as separate...
[BUG] /effort asks to confirm a cache miss on Opus 5.5 with a subscription, although the cache is kept
### Preflight Checklist - [X] I have searched [existing issues](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) and this hasn't been reported yet - [X] This is a single bug report (please file separate reports for different bugs) - [X]...
[BUG] Streamed Bedrock calls report 0 tokens when the usage arrives after messageStop
### Description Bedrock Converse streams send the token usage in a `metadata` event after `messageStop`. Two exits in `BedrockCompletion` leave the stream loop before that event, so the call reports 0 tokens in `llm.get_token_usage_summary()` and `LLMCallCompletedEvent.usage` is...
[BUG] MCP discovery returns only the first page for prompts, and for tools on the HTTPS discovery path
...`[None, 'cursor-2']` and discovers both tools. A regression test per path goes from failing to passing. ### Possible Solution Loop while `nextCursor` is not `None` in `MCPClient._list_prompts_impl` and `MCPToolResolver._discover_mcp_tools`, mirroring the fix already proposed for...
[Bug]: AzureAICompletionsModel.astream_complete always raises TypeError (stream passed twice)
### Bug Description `AzureAICompletionsModel.astream_complete` fails on every call with `TypeError: ... complete() got multiple values for keyword argument 'stream'`. The sync `stream_complete` works. `astream_complete` passes `stream=True` into the chat-to-completion decorator,...
[Bug]: OpenAIResponses.achat drops reasoning items from the input even with store=True
### Bug Description `OpenAIResponses.achat()` sends a different `input` than `chat()` for the same history: with `store=True`, the reasoning items (`ThinkingBlock`s carrying an `id`) are dropped from the async request. `_chat` builds the input with `to_openai_message_dicts(...,...
Turbopack: tab reloads forever after restarting next dev (first HMR update is `restart`, regression in 16.4.0)
...initial computation. Since #98385 it is forwarded unless it only carries issues. In the failing state that first emission is `restart` for every new subscription, so the forwarded `restart` reloads the page and the reload gets another one. - **What the loop needs** (controls in...
Claude Code co-wrote 482 of the 633 commits in my self-hosted notes server. Here is how we kept it in line
...on one codebase need rules, so here is what the repo and our skills make Claude do.. How do we decide what to build? . We both run Matt Pocock's skills in Claude Code (I keep a customised fork). A new feature usually starts with a grilling session, where Claude questions us...
Check.str_startswith / str_endswith strategies treat the string as a regex, so schema.example() generates invalid data
**Describe the bug** `Check.str_startswith(string)` and `Check.str_endswith(string)` validate with a literal prefix/suffix (`Series.str.startswith` / `Series.str.endswith`). Their data-synthesis code, however, puts `string` into a regex without escaping it:...
[BUG] Bedrock file constraints swap the image and document size limits
### Description `BEDROCK_CONSTRAINTS` in `crewai_files/processing/constraints.py` caps images at 4,608,000 bytes (4.5 MB) and PDFs at 3,840,000 bytes (3.75 MB). The Bedrock Converse API documents the reverse: each image must be no more than 3.75 MB and each document no more than...
[BUG] Claude Code - Catastrophic Windows Failure - Non Rebootable
...Code locally on Windows 11 Try to upload a document ( soem text based code) PC freezes. Unable to do anything Reboot computer. Computer - black screen ( not blue screen) Message tells me OS is damaged. ### Claude Model Opus ### Is this a regression? Yes, this worked in a...
[BUG] Pressing "Esc" in a subgents's /btw call also cancels the subagent itself
### Preflight Checklist - [x] I have searched [existing issues](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) and this hasn't been reported yet - [x] This is a single bug report (please file separate reports for different bugs) - [x]...
[BUG] Windows desktop: Remote Control not restored after app relaunch (silent update / restart), sessions show as archived on mobile
### What's wrong On the Windows desktop app (Code tab), Remote Control is not restored for existing sessions after the app relaunches. On the phone (Claude mobile app, Code), every session that was connected moves to **Archive**. Unarchiving one on the phone wakes it on the...
Pasting into the Chat input freezes the window on Linux/X11 since 1.140.0 (1.139.1 works)
Does this issue occur when all extensions are disabled?: Yes - VS Code Version: 1.140.0 (commit 07f806f999227108933c2e30515b26eecc1fda74), Electron 43.7.3, .deb from packages.microsoft.com - OS Version: Ubuntu 24.04.5 LTS, Linux 6.8.0 x64, X11 session (MATE desktop), no...
Turbopack dev (16.4.0): Pages Router page modules missing ("module factory is not available") after _app availability seeding (#97664)
### Link to the code that reproduces this issue https://github.com/Stanzilla/next-turbopack-app-availability-repro ### To Reproduce 1. `npm install` 2. `npm run dev` (`next dev --turbopack -p 3456`) 3. Open http://localhost:3456/ The repro is a Pages Router app with no...
[BUG] create_model_from_schema drops the null member of {"type": ["object", "null"]} depending on member order (id() reuse in in_progress cache)
### Description `create_model_from_schema` builds the wrong type for a nested object declared with a list-form type, e.g. `{"type": ["object", "null"], "properties": {...}}`: the field comes out as the bare nested model instead of `Optional[model]`, so a `null` value the schema...
SFTPHook concurrent workers ignore the parent hook's constructor overrides
`SFTPHook` builds its concurrent worker hooks (`providers/sftp/src/airflow/providers/sftp/hooks/sftp.py`, the two `SFTPHook(ssh_conn_id=self.ssh_conn_id, ...)` calls used by the multi-worker download/upload paths) from `ssh_conn_id` alone. Any settings the parent hook received...
[BUG] [Desktop Code tab] Slash commands with non-ASCII (Korean) names are rejected on send
...app 2.16120.0, Code tab (not the CLI), macOS 27.0, Korean UI / Korean input method - Workaround: ASCII-named alias files that reference the original command - Screenshot of the warning attached
[BUG] Claude Code is holding a power request on Windows, preventing sleep
...I pay for usage! I installed the client on a Windows machine, a gaming rig, where I have problems with HID devices hogging the HID bus and causing sleep to not work. I had the system in working order; however when I launch Claude to do development work on custom HID controllers,...
notebook.codeActionsOnSave does nothing when the notebook's first cell is markdown
Type: Bug **Settings** "notebook.defaultFormatter": "charliermarsh.ruff", "notebook.formatOnSave.enabled": true, "notebook.codeActionsOnSave": { "notebook.source.organizeImports": "explicit" } **Steps to reproduce** Create a notebook whose first cell is markdown (for example #...
[BUG] Windows MSIX 2.16120: app won't launch (0x80070057 "The parameter is incorrect") when %LOCALAPPDATA%\Google is a junction to another volume
...Summary After the idle stealth update to **Claude 2.16120.0.0** (MSIX), the desktop app fails to launch with "The parameter is incorrect" (0x80070057). Root cause: the manifest's `FileSystemWriteVirtualization` excluded directory `$(KnownFolder:LocalAppData)\Google\Chrome\User...
[BUG] OpenAI provider emits LLMCallFailedEvent twice per failed call and pops the enclosing agent scope
### Description A failed non-streaming `OpenAICompletion.call`/`acall` (chat completions or the Responses API) emits `LLMCallFailedEvent` twice. `_handle_completion`/`_ahandle_completion`/`_handle_responses`/`_ahandle_responses` emit it and re-raise, and `call`/`acall` emit it...
Daily IntervalSchedule anchored at 09:00 America/New_York returns 08:00 runs when get_dates starts after a DST change
### Bug summary A daily `IntervalSchedule` with `timezone="America/New_York"` and a 09:00 anchor returns 08:00 dates when `get_dates` is called with a `start` on the other side of a DST change from the anchor. The scheduler service calls it with `start_time=right_now`...
SFTPHookAsync lets a connection known_hosts extra override an explicit constructor path
`SFTPHookAsync._parse_extras()` decides whether to take `known_hosts` from the connection extras with: `self.known_hosts` has already been through `os.path.expanduser()` in `__init__`, while `default_known_hosts` is the literal `"~/.ssh/known_hosts"`. The two never compare...
Concurrent SFTP directory transfers drop the parent hook's connection overrides
`SFTPHook.store_directory_concurrently()` and `retrieve_directory_concurrently()` build one hook per worker as: Since #73419 the workers inherit `no_host_key_check`, but every other constructor override on the parent hook is still discarded - `remote_host`, `port`, `username`,...
[BUG] Monitor tool: doesn't respect persistent flag and timeout_ms schema caps at 3600000 but actual lifetime is ~30 minutes, even during active (non-idle) sessions
### Preflight Checklist - [x] I have searched existing issues and this hasn't been reported yet (related but distinct: #63023, #65968 - see below) - [x] This is a single bug report (please file separate reports for different bugs) - [x] I am using the latest version of Claude...
[BUG] 2.1.268 moved the IDE/selection indicator from the status line into the prompt input; no way to move it back
### Preflight Checklist - [x] I have searched existing issues and this hasn't been reported yet - [x] This is a single bug report - [x] I checked the latest version (2.1.270 changelog): the behavior is unchanged since 2.1.268 ### What's Wrong? Since 2.1.268 the IDE indicator...
rpm.releases.hashicorp.com aarch64 packages fail GPG verification after 2026-09-09 key rotation (x86_64 unaffected)
### Terraform Version ### Terraform Configuration Files N/A - this is a package installation/signing issue on the `rpm.releases.hashicorp.com` repo, not a Terraform configuration or runtime bug. ### Debug Output N/A (not a `terraform` runtime issue - see `dnf`/`rpm` output under...
[BUG] Cowork (Windows): all Plan9 shares fail with "Plan9 mount failed: invalid argument" after Windows update KB5124008 (26200.9445) — uninstalling the KB fixes it
### Preflight Checklist - [x] I have searched [existing issues](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) and this hasn't been reported yet - [x] This is a single bug report (please file separate reports for different bugs) - [x]...
[BUG] Cowork Windows: September 2026 cumulative update breaks Plan9 share attach — device_bash dead on both ARM64 (KB5124012 / 28000.2954) and x64 (KB5124008 / 26200.9445); confirmed by rollback A/B on five machines
...- the only error surfaces later, inside the guest, with no actionable cause. A user has no way to reach the VM to diagnose it, because the shell itself is what is broken. ### Error Messages/Logs ### Steps to Reproduce 1. Windows 11 ARM64 build `28000.2804`, Claude Desktop...
[BUG] Claude Code Desktop on windows errors when updating
### Preflight Checklist - [x] I have searched [existing issues](https://github.com/anthropics/claude-code/issues?q=is%3Aissue%20state%3Aopen%20label%3Abug) and this hasn't been reported yet - [x] This is a single bug report (please file separate reports for different bugs) - [x]...
bypassPermissions mode now prompts on `cd DIR && grep …` when a Read() deny rule is configured (regression in 2.1.259)
...shows the mode was `bypassPermissions` throughout and auto mode was never entered. ## Workaround Tell agents not to use `cd DIR &&` prefixes or recursive grep over directories (use explicit file lists or a file-search tool). That removes the prompts but means the deny-rule...
[Desktop] Slash-command autocomplete no longer triggers mid-message since 1.40609.0 (terminal CLI unaffected)
## Description In the **Claude Code desktop app** (macOS), typing `/` in the **middle** of a message used to bring up the slash-command/skill autocomplete popup, the same way it does at the start of the input. Since the app auto-updated to **1.40609.0** (installed 2026-08-27...
Add missing tests for `sftp` exceptions
## Description One module in the `sftp` provider has no dedicated test module, and (unlike most entries on the `OVERLOOKED_TESTS` list) it is not covered indirectly. Nothing under any `providers/*/tests/` directory imports it at all: | Module | Expected test file | | --- | --- |...
[BUG] Desktop app: dictation mic is removed from the composer when an attachment is present
...composer as soon as a file is attached. With an attachment in the composer there is no way to dictate the message that goes with it - the control is not rendered at all. Attaching a file and then speaking the prompt used to work. The desktop app also has no keyboard shortcut for...
Feature request: append to system prompt at launch (--append-system-prompt / GEMINI_APPEND_SYSTEM_MD) — GEMINI_SYSTEM_MD only replaces
...was closed by the backlog-cleanup bot for inactivity with an invitation to open a fresh issue with updated details - this is that issue. ### Why is this needed? `GEMINI_SYSTEM_MD` exists but **replaces** the entire system prompt, which makes it unusable for the common case of...
Companies and products named
About this problem
Evidence
67 posts from 60 people in 18 places, about 6 a week over 71 days. Mostly on anthropics/claude-code, apache/airflow, crewAIInc/crewAI. Tools named alongside: Claude, GitHub, Apache Airflow, Shell.
Frustration Frustration 0 of 3· Seen on GitHub issues, Hacker News, Reddit
Summary, model-written
Claude Code users face permission mode issues and broken guest VM shells that prevent them from accessing or diagnosing system errors.
How it was grouped
Posts that state a pain and match the "breaking-changes" rule about Claude Code. First post Aug 27, 2026, latest Nov 5, 2026. Corroborated: 3 or more posts from 2 or more people or places. Method
The same problem elsewhere
GitHub upgrades break working setups
87 posts from 75 people on GitHub issues, Hacker News, Stack Overflow.
Upgrades and new releases break working setups
76 posts from 62 people on GitHub issues, Reddit, App Store reviews.
Codex upgrades break working setups
65 posts from 58 people on GitHub issues, Reddit.
DuckDB upgrades break working setups
37 posts from 28 people on GitHub issues.
Claude upgrades break working setups
35 posts from 34 people on GitHub issues, Reddit, Hacker News.
ClickHouse upgrades break working setups
31 posts from 20 people on GitHub issues.
Anthropic upgrades break working setups
27 posts from 22 people on GitHub issues.
PyTorch upgrades break working setups
26 posts from 21 people on GitHub issues, Reddit.
ChatGPT upgrades break working setups
26 posts from 26 people on GitHub issues, Community forums, App Store reviews.
OpenAI upgrades break working setups
25 posts from 25 people on GitHub issues, Reddit, Community forums.
Open WebUI upgrades break working setups
25 posts from 2 people on GitHub issues.
Gemini upgrades break working setups
24 posts from 24 people on GitHub issues, Community forums, Reddit.
Other Claude Code problems
Claude Code lock-in makes switching costly
14 posts from 14 people on GitHub issues, Hacker News, Reddit.
Claude Code plans, limits and pricing keep changing
9 posts from 9 people on GitHub issues, App Store reviews, Reddit.
Claude Code documentation is outdated or confusing
6 posts from 6 people on GitHub issues, Hacker News.
Running Claude Code locally hits memory and speed limits
6 posts from 6 people on GitHub issues.
Claude Code users keep filing bugs
Top post: "Bug Report: Privacy violation in User-Agent header", one of 42 posts from 39 people on GitHub issues, Reddit.
Trouble with Claude Code desktop
Top post: "Desktop app disallows SendMessage, so subagents cannot be messaged or resumed", one of 20 posts from 20 people on GitHub issues, Reddit.
Claude Code rate limits and usage caps block work
20 posts from 20 people on GitHub issues, Reddit, Hacker News.
Claude Code outages and slowdowns disrupt work
18 posts from 18 people on Glyph Ibis, GitHub issues, Hacker News.
Claude Code runs out of context on real work
12 posts from 12 people on GitHub issues, Reddit, Hacker News.
Claude Code tables suffer from small files and skew
9 posts from 9 people on GitHub issues, Reddit.
Trouble with Claude Code session
Top post: "Remote Control (mobile): session start cards / task chips not shown on the phone; start_session refused for", one of 7 posts from 7 people on GitHub issues.
Claude Code error messages are cryptic
6 posts from 6 people on GitHub issues, Reddit.
History