Claude Code

issue2026-08-13T06:46:26Z

[BUG] Teammate subagents of a forked-skill agent cannot reply to their parent; SendMessage to "team-lead" false-succeeds into an orphaned inbox

## Summary Subagents spawned as named mailbox teammates by a forked-skill background agent have **no working reply path back to their parent** — and one of the failure modes returns `success: true` while dropping the message. The parent deadlocks waiting for replies that can never arrive, with success signals on both sides. ## Environment - Claude Code 2.1.2

importance 7@jacwrightopenbugplatform:macosarea:agentsOriginal evidence ↗

issue2026-08-13T02:37:04Z

[Bug] Claude Code ignores conciseness instructions and over-elaborates responses

**Bug Description** refuses to stick to instructions about being concise and keeps inventing complicated ways of explaining simple things **Environment Info** - Platform: win32 - Terminal: windows-terminal - Version: 2.1.228 - Feedback ID: 68a63514-36cf-4bca-903f-8377527d6726 **Errors** ```json [] ```

importance 7@FysicsTeachropenbugplatform:windowsarea:modelneeds-reproOriginal evidence ↗

issue2026-08-13T14:05:19Z

[Bug] Anthropic API Error: Content Flagged by Safety Filters on Legitimate Code Tasks

**Bug Description** I will continue to raise a bug report for each of these absolutely attrocious failures to oeprate within absolutely legitimate scopes: API Error: Fable 5's safeguards flagged this message (https://www.anthropic.com/legal/aup). Our intentionally broad safeguards allow us to deliver more capabilities faster, but can sometimes flag legitimat

importance 7@creativeheadzopenbugplatform:linuxarea:modelOriginal evidence ↗

issue2026-08-12T22:17:06Z

Cross-session messages are delivered but never trigger the recipient's turn (desktop app, Windows)

# Bug report — cross-session messages are delivered but never trigger the recipient's turn (Claude Code desktop, Windows) ## Summary `mcp__ccd_session_mgmt__send_message` reports `Message sent to session ...`, and the message **is** genuinely delivered — it appears in the recipient's queue via `list_events` and renders correctly in the recipient's window. Bu

importance 7@DensCadopenOriginal evidence ↗

issue2026-08-13T04:49:03Z

MCP OAuth loopback redirect uses 127.0.0.1 instead of localhost — breaks providers (e.g. Salesforce) that only allow localhost callback URLs

Related to #69326 and #66511 (both closed as not planned / stale), but with a concrete, current real-world case that isn't a hypothetical headless-server scenario. **Problem:** `claude mcp login` / `/mcp` builds its OAuth loopback redirect as `http://127.0.0.1:<port>/callback`. Several major OAuth providers reject IP-literal HTTP callback URLs outright but e

importance 7@alex-magillopenbughas reproarea:autharea:mcpOriginal evidence ↗

issue2026-08-13T02:25:23Z

[BUG] ScheduleWakeup has no fallback when tengu_kairos_loop_dynamic is off — self-pacing unreachable for all non-first-party deployments, and non-/loop callers fail opaquely

> **Rewritten twice on 2026-08-12.** The original report claimed a no-interval `/loop` dies after one tick (wrong — the harness degrades gracefully), and then attributed a regression to v2.1.202 (also wrong — the behavior is identical in every retrievable build). See the correction comments below; the surviving defect is described here. ### Preflight Checkli

importance 7@spencer-carveropenbughas reproarea:corearea:skillsarea:providersOriginal evidence ↗

issue2026-08-13T14:31:21Z

HTTP MCP: SDK client does not attach configured --header on tool calls (Claude Code 2.1.114, Windows)

## Summary On a Windows machine running claude-code 2.1.114, an HTTP-transport MCP server configured via `claude mcp add --transport http --header "Authorization: Bearer <token>"` stores the header correctly (visible in `claude mcp get`), reports `✓ Connected` in `claude mcp list`, and accepts the token on direct `curl.exe` calls. Tool calls initiated throug

importance 6@justanotherguymeclosedbughas reproplatform:windowsarea:mcpstaleOriginal evidence ↗

issue2026-08-13T14:29:31Z

Bug: injected system context hardcodes model name, ignores --model flag override

## Summary When launching Claude Code with `--model claude-opus-4-7`, the CLI correctly displays the model as **Opus 4.7** in the welcome banner. However, the system prompt injected into the model's context hardcodes `\"You are powered by the model named Sonnet 4.6. The exact model ID is claude-sonnet-4-6.\"` — causing the model to incorrectly self-identify

importance 6@urisnehclosedbugplatform:windowsarea:corestaleOriginal evidence ↗

issue2026-08-13T14:34:03Z

[FEATURE] persistence of memory over chat sessions between chat sessions

### 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 ## Problem Statement Claude Code CLI users lose 40-50% productivity between sessi

importance 6@mjc02840closedenhancementarea:corememorystaleOriginal evidence ↗

issue2026-08-13T14:29:33Z

[BUG] Inconsistent "extra usage" selections between Opus, Sonnet, 200K, and 1M contexts

### 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] I am using the latest version of Claude Code ### What's Wrong? The /model picker

importance 6@GeoffNordlingclosedbugstaleOriginal evidence ↗

issue2026-08-13T14:34:05Z

[Bug] Claude Code repeatedly requests permissions and forgets environment variables across sessions

**Bug Description** Whenever tests are run I need to use environment variables. Claude keeps forgetting this. I am also prompted to give permission for grepping results for tests. It makes it impossible to be AFK and leave it running. I've asked claude to asjust permissions but it keeps asking. Also it has taken well over an hour and used lots of extra token

importance 6@Malc75closedbugplatform:macosplatform:intellijarea:permissionsstaleOriginal evidence ↗

issue2026-08-13T14:29:29Z

[Bug] Token usage spike with non-Opus models after Opus 4.7 release

**Bug Description** Since the release of opus 4.7, my token usage has spiked, even if I'm not using opus at all, but models like sonnet. I went from barely reaching 6-10% during a regular session to reaching 6-10% on only a messages. On the Max20x, that is unexpected and, for the cost, not appropriate. Thank you. **Environment Info** - Platform: win32 - Term

importance 6@MeterH1closedbugplatform:windowsarea:coststaleOriginal evidence ↗

issue2026-08-13T14:29:23Z

Status line not refreshed after `/compact` slash command

## Summary After running `/compact`, the status line continues to display the **pre-compact** context size, even though the conversation has actually been compacted. `/context` correctly shows the new (much smaller) context size, but the status line stays stale until something else triggers a re-render. ## Reproduce 1. In a session where context has grown la

importance 6@metricorbit-codeclosedbugplatform:macosarea:statuslinestaleOriginal evidence ↗

issue2026-08-13T14:30:47Z

Worktree background full checkout fails with 'not-a-git-repo' when repo has `submodule.recurse = true`

## Summary When creating a Claude Code worktree against a repo that has `submodule.recurse = true` set in its local git config, the staged background full checkout always fails with `Background full checkout failed: not-a-git-repo`. The worktree directory ends up partially populated (only the selective paths — `CLAUDE.md`, `.claude`, `.mcp.json` — are checke

importance 6@thomasqbradyclosedduplicateOriginal evidence ↗

issue2026-08-13T14:29:35Z

[BUG] Claude Code VS Code Extension crashes on Windows 11 with error 3221225477

### 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] I am using the latest version of Claude Code ### What's Wrong? **Version:** 2.1.

importance 6@L0kiz0closedbugplatform:windowsarea:authplatform:vscodestaleOriginal evidence ↗

issue2026-08-13T14:34:23Z

[BUG] VSCode reopen appends duplicate entries to session JSONL, corrupting /insights analysis

## Description When VSCode is closed and reopened with an active Claude Code session, entries from that session are re-appended to the JSONL log file. Each replayed entry is byte-for-byte identical to the original (same `uuid`, same `timestamp`, same content) with only minor context fields differing (e.g. `cwd`). ## Minimal example — same UUID at two differe

importance 6@An0nyaclosedbughas reproplatform:vscodestaleOriginal evidence ↗

issue2026-08-13T14:34:54Z

[BUG] 2.1.129 install subcommand intermittently fails to create ~/.local/bin/claude

### 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] I am using the latest version of Claude Code ### What's Wrong? On 2026-05-06, tw

importance 6@AlexanderMuenchclosedbugstaleOriginal evidence ↗

issue2026-08-13T14:33:09Z

[BUG] Agent tool spawn crashes with K.length on first invocation in v2.1.136

## 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] I am using the latest version of Claude Code (2.1.136) > Note on the first checkb

importance 6@francisco-perez-sorrosalclosedbughas reproplatform:macosarea:agentsapi:anthropicOriginal evidence ↗

issue2026-08-13T14:33:24Z

[BUG] Cowork shows "Virtual Machine Platform not available" — yukonSilver marked as unsupported on Windows 11 Pro

### 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] I am using the latest version of Claude Code ### What's Wrong? What's Wrong: Cow

importance 6@ImperialCommandclosedbugplatform:windowsarea:coworkarea:desktopstaleOriginal evidence ↗

issue2026-08-13T14:29:47Z

Worktree-isolated subagents leak Edit/Write into parent checkout (Windows)

## Summary When `Agent({ isolation: "worktree", ... })` dispatches a subagent on Windows, the subagent's `Edit` / `Write` operations sometimes write to the **parent checkout's** files instead of the isolated worktree's files. The same pattern affects subagent-issued git operations — commits land on the parent's branch ref instead of the worktree's branch. Th

importance 6@JBudgeMEclosedbughas reproplatform:windowsarea:toolsarea:agentsOriginal evidence ↗

issue2026-08-13T14:29:02Z

[Bug] Claude Desktop Dispatch: No way to start new session, persistent session ID across restarts

**Bug Description** I have a problem when using CLaude Desktop on Windows 11, specifically related to the Dispatch feature. I have been using it for weeks, and Claude Desktop is issuing warnings that my dispatch session is VERY large and I need to start a new one. But I have not been able to do so, and Claude cannot tell me how to do it either. Finally I ask

importance 6@GraceFortMillITclosedbugplatform:windowsarea:coworkarea:desktopstaleOriginal evidence ↗

issue2026-08-13T14:31:58Z

[BUG] Claude can not explain "ultraplan" because the text is stripped.

### 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] I am using the latest version of Claude Code ### What's Wrong? [BUG] I cant ask

importance 6@webcloudstudioclosedbugplatform:macosarea:tuistaleOriginal evidence ↗

issue2026-08-13T14:31:56Z

[BUG] apikeyhelper script not getting terminated on exit

### 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] I am using the latest version of Claude Code ### What's Wrong? ## Issue On Windo

importance 6@nuernberclosedbughas reproplatform:windowsarea:authstaleOriginal evidence ↗

issue2026-08-13T14:29:59Z

[Bug] Token usage occurring during idle periods without active requests

**Bug Description** the token is burning even when I am not using it **Environment Info** - Platform: darwin - Terminal: iTerm.app - Version: 2.1.139 - Feedback ID: d3f9b23e-7e21-46ca-a42c-80a0983bcfbb **Errors** ```json [{"error":"Error: NON-FATAL: Lock acquisition failed for ~/.local/share/claude/versions/2.1.139 (expected in multi-process scenarios)\n at

importance 6@Hao-Juiclosedbugplatform:macosarea:costneeds-infostaleOriginal evidence ↗

issue2026-08-13T14:28:44Z

[Bug] Claude 3 Opus exhibits inconsistent reasoning with verification requests and self-corrections

**Bug Description** check how many times opus is finding him self mistaking many times, each time I verify something opus push him self back and correct him self, anytime opus is asked to verify his claim he go another direction , not sure if i can trust it now **Environment Info** - Platform: darwin - Terminal: WarpTerminal - Version: 2.1.139 - Feedback ID:

importance 6@omaraltaliclosedbugplatform:macosarea:modelstaleOriginal evidence ↗

issue2026-08-13T14:33:52Z

[Claude Desktop] Background token consumption without user interaction — 5% weekly usage lost overnight

### 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] I am using the latest version of Claude Code ### What's Wrong? Weekly usage incr

importance 6@daynitonclosedbugplatform:windowsarea:coststaleOriginal evidence ↗

issue2026-08-13T14:35:50Z

[BUG] Windows 10: "Interrupted" error on every restart - Cannot read properties of null (reading 'effortLevel')

### 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] I am using the latest version of Claude Code ### What's Wrong? ## Bug Descriptio

importance 6@akamkvclosedbughas reproplatform:windowsarea:corestaleOriginal evidence ↗

issue2026-08-13T14:32:22Z

Allow workspace authors to suppress named platform skills during custom onboarding flows

### 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 When a Cowork workspace ships its own onboarding flow — a custom skill that walks

importance 6@cameron-bales-telus-healthclosedenhancementarea:coworkarea:skillsstaleOriginal evidence ↗

issue2026-08-13T14:29:37Z

[FEATURE] optional raw/verbatim mode for WebFetch

### 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 ## Summary WebFetch routes fetched pages through a smaller extractor model before

importance 6@flohococlosedenhancementarea:toolsstaleOriginal evidence ↗

issue2026-08-13T14:37:22Z

[BUG] Localisation is poor

### 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] I am using the latest version of Claude Code ### What's Wrong? The German transl

importance 6@ygoeclosedbugplatform:windowsarea:uiarea:desktopOriginal evidence ↗

issue2026-08-13T14:29:57Z

Claude CLI subprocess silent-hang: zero stdout for 3+ minutes (intermittent)

## Summary We invoke `claude.exe` as a subprocess from a PowerShell automation framework (Windows 11, Claude Code v2.1.x range). Intermittently, the subprocess produces **zero stdout bytes for 3+ minutes** before either eventually producing output or being killed by our wall-clock timeout guard. Same prompt + same model + same environment will succeed on one

importance 6@GoSunriseCEOclosedbugplatform:windowsarea:clistaleOriginal evidence ↗

issue2026-08-13T14:30:01Z

[BUG] Windows Desktop: Previous prompt block persists as floating overlay on subsequent responses

### 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] I am using the latest version of Claude Code ### What's Wrong? On the Windows Cl

importance 6@denys-progclosedinvalidstaleOriginal evidence ↗

issue2026-08-13T14:34:51Z

commit-commands plugin uses session working directory instead of the git repo where changes were made

### 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] I am using the latest version of Claude Code ### What's Wrong? When Claude Code'

importance 6@daniel769closedbughas reproplatform:macosarea:skillsstaleOriginal evidence ↗

issue2026-08-13T14:33:01Z

[DOCS] Documentation table rows for Project Instructions and User Instructions are in incorrect order

### Documentation Type Missing documentation (feature not documented) ### Documentation Location https://code.claude.com/docs/en/memory ### Section/Topic Choose where to put CLAUDE.md files ### Current Documentation The official documentation table at https://code.claude.com/docs/en/memory shows: Row 2 → Project Instructions (./CLAUDE.md) Row 3 → User Instru

importance 6@kannanwisenclosedbugdocumentationarea:docsstaleOriginal evidence ↗

issue2026-08-13T14:33:45Z

[BUG] Worktree isolation fails with "Background full checkout failed: git-crypt-locked" in git-crypt repos

### 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] I am using the latest version of Claude Code ### What's Wrong? When starting a n

importance 6@johanhalebyclosedbughas reproplatform:macosarea:desktopstaleOriginal evidence ↗

issue2026-08-13T14:35:53Z

Claude Code does not enforce CLAUDE.md project instructions — regressions shipped despite explicit rules

I set up CLAUDE.md with explicit, non-negotiable rules to prevent regressions in my project. Claude Code read these rules at the start of every session and acknowledged them, but did not follow them during implementation. This caused two production regressions that required multiple fix cycles and wasted significant session usage. **Rules that were in place

importance 6@prathmesh312closedbugarea:modelarea:corestaleOriginal evidence ↗

issue2026-08-13T14:34:58Z

[BUG] PreToolUse hook payload `session_id` / `transcript_path` alternate between current and stale UUIDs within one conversation

## Bug Within a single conversation, the `PreToolUse` hook payload's `session_id` and `transcript_path` fields **alternate** between the currently active session UUID and a stale UUID (frozen from earlier in the same conversation). Different `PreToolUse` invocations seconds apart receive different values. ## Evidence Captured via the [`@nizos/probity`](https

importance 6@wazumclosedbugplatform:macosarea:hooksstaleOriginal evidence ↗

issue2026-08-13T14:36:25Z

[Bug] Agent generating hallucinated responses instead of using available tools

**Bug Description** main agent is making things uo **Environment Info** - Platform: darwin - Terminal: Apple_Terminal - Version: 2.1.140 - Feedback ID: 1de4fb55-27a9-4b0a-96f9-d062da35c4d9 **Errors** ```json [{"error":"Error: Request was aborted.\n at makeRequest (/$bunfs/root/src/entrypoints/cli.js:51:6192)\n at processTicksAndRejections (native:7:39)","tim

importance 6@saif-mohtasibclosedbugplatform:macosarea:modelneeds-reprostaleOriginal evidence ↗

issue2026-08-13T14:35:20Z

claude-sonnet-4-6 /context reports 200K — same registry-stale bug already fixed for Opus 4.7 in v2.1.117

**Claude Code version:** 2.1.140 (Windows native build) ## What I expected `/context` should show `1,000,000` for `claude-sonnet-4-6`. Per the **2.1.49** changelog: > Sonnet 4.5 with 1M context is being removed from the Max plan in favor of our frontier Sonnet 4.6 model, which now has 1M context. Please switch in /model. The same registry-stale bug was alrea

importance 6@bse-aiclosedbughas reproplatform:windowsarea:modelstaleOriginal evidence ↗

issue2026-08-13T14:35:39Z

[BUG] Claude Desktop - Code - MCP InputSchema with nullable arrays has type silently dropped

### 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] I am using the latest version of Claude Code ### What's Wrong? In Code mode (i.e

importance 6@abirminghamclosedbugplatform:windowsarea:mcparea:desktopstaleOriginal evidence ↗

issue2026-08-13T14:36:22Z

[BUG] Agentic loop terminates after clean tool_result on empty-text/empty-thinking tool_use turn — Windows v2.1.138, long post-tool inference

### 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] I am using the latest version of Claude Code ### What's Wrong? On Windows Claude

importance 6@jatoranclosedbughas reproplatform:windowsarea:corestaleOriginal evidence ↗

issue2026-08-13T10:28:37Z

[FEATURE] Allow hooks to spawn subagents / background agents (programmatic dispatch from hook events)

### 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 Hooks are Claude Code's mechanism for **deterministic** automation in the agentic

importance 6@blocwaveclosedenhancementarea:hooksarea:agentsstaleOriginal evidence ↗

issue2026-08-13T14:39:47Z

[Feature Request] AskUserQuestion option descriptions truncate on mobile, hiding key differentiators

**Bug Description** Title: [Mobile] AskUserQuestion option descriptions truncate to ~2 lines, hiding the differentiator Type: Feature request / UX What happens: On the Claude Code mobile app, the option descriptions in a multiple-choice question (AskUserQuestion) are clamped to ~2–3 lines and hard-truncated with an ellipsis. When the distinguishing detail is

importance 6@julienbonastreclosedduplicateOriginal evidence ↗

issue2026-08-13T14:37:48Z

[BUG] Claude No longer works on New worktree

### 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] I am using the latest version of Claude Code ### What's Wrong? After recent upda

importance 6@tdsolomonsclosedbugplatform:macosarea:desktopOriginal evidence ↗

issue2026-08-13T14:30:32Z

AskUserQuestion dialog hides preceding assistant message text until the question is answered

## Describe the bug When the assistant outputs a text message and then calls the `AskUserQuestion` tool **in the same turn**, the interactive question dialog is rendered immediately, but the assistant's preceding text (which contains the context needed to answer the question) is not shown until after the user answers or dismisses the dialog. In our case the

importance 6@Mrazotticlosedbugduplicateplatform:windowsarea:tuiOriginal evidence ↗

issue2026-08-13T14:41:49Z

Third-party marketplace with autoUpdate:true not refreshed at session start; new release not picked up across restarts (v2.1.175)

## Environment - Claude Code **v2.1.175**, macOS (Darwin 25.5.0), native installer - Third-party marketplace: **private self-hosted GitLab**, added via SSH URL (`git@<internal-host>:<group>/<repo>.git`), one plugin installed at **user scope** - `autoUpdate: true` confirmed in `~/.claude/plugins/known_marketplaces.json`; the `/plugin` UI shows: *"Auto-update

importance 6@dimkin-euclosedOriginal evidence ↗

issue2026-08-13T10:28:35Z

Feature: Named sessions via --session <name> flag for easier session management

### 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 **Summary** Add support for named sessions so users can create, resume, and manag

importance 6@sanalthomas47closedenhancementarea:clistaleOriginal evidence ↗

issue2026-08-12T22:18:30Z

[Bug] Simplified Chinese kanji glyphs appearing in Japanese responses

**Bug Description** Title: Simplified Chinese characters leaking into Japanese output (字形 mix-up) Description: When responding in Japanese, Claude occasionally renders individual kanji in Simplified Chinese glyphs instead of the correct Japanese forms. In my session the word "審問官" was output as "审问官" (with 审/问 in Simplified Chinese), and it was inconsistent

importance 6@teshnakamuraclosedbugplatform:linuxarea:modelstaleOriginal evidence ↗

issue2026-08-13T10:28:06Z

claude remote-control --permission-mode bypassPermissions is silently ignored on mobile (native Windows)

### Scenario `claude remote-control` (alias `claude remote`) accepts `--permission-mode bypassPermissions`, and the flag is listed in `claude remote-control --help`. Starting a server-mode remote session with it and connecting from the Claude iOS app does not suppress permission prompts. ### Steps to reproduce 1. On the local machine, start a remote-control

importance 6@yury-opolevclosedduplicateplatform:windowsarea:permissionsstaleOriginal evidence ↗

issue2026-08-12T22:18:46Z

[Bug] Anthropic API Error: Server rate limit exceeded (temporary)

**Bug Description** API Error: Server is temporarily limiting requests (not your usage limit) · Rate limited **Environment Info** - Platform: win32 - Terminal: windows-terminal - Version: 2.1.114 - Feedback ID: 53ecaef4-9d15-41db-b9ce-69ddd3395141 **Errors** ```json [{"error":"Error: 429 {\"type\":\"error\",\"error\":{\"type\":\"rate_limit_error\",\"message\

importance 6@Sahasraraclosedbugplatform:windowsexternalarea:apiapi:anthropicOriginal evidence ↗

Release, roadmap, and discussion

Publicly indicated next

Observation trail