What's breaking AI agent harnesses right now, with quoted workarounds where one exists.
Pulled from public GitHub issues updated in the last 45 days and filtered for always-on problems: install and update breakage, crashes and hangs, gateway/daemon trouble, auth, and headless setups. Ranked by engagement, weighted toward always-on, self-hosting, install/update, crash, auth, MCP and daemon problems (cosmetic desktop-UI issues are down-weighted). Workarounds and "fixed in" versions are quoted only when the issue or its top comments state them. Severity is an automatic estimate from labels, keywords, and engagement, not an official rating. Last refresh: Oct 26, 2026 6:18 AM ET.
Claude Code versions 2.1.100 and 2.1.101 consume ~20,000 more cachecreationinputtokens per request than v2.1.98, despite sending fewer bytes in the request payload. The inflation is server-side and version-specific (likely User-Agent…
Downgrade to v2.1.98 or earlier. Check `~/.local/share/claude/versions/` for available binaries, or use `npx claude-code@2.1.98`.
Codex Desktop: 26.527.60818 Codex build: 3437 Darwin 25.5.0 arm64 arm Codex Desktop for macOS triggers a persistent macOS Gatekeeper/SystemPolicy loop after launch. With Codex closed, the system is quiet. As soon as Codex Desktop is…
Is there plane to add mcp server call?
Starting ~June 16, my ChatGPT Plus Codex budget drains in 2–3 prompts on gpt-5.5, where the same model/plan/app previously gave 20+ prompts. My Codex session logs (tokencount / ratelimits events) show the limit-% consumed per token has…
When Claude Code generates a compound bash command using && (e.g.: cd /some/path && git add file && git commit -m "msg"), the permission approval prompt incorrectly identifies cd as the action requiring approval, showing a cd: permission…
Upstream main after #26210 (Encrypt multi-agent v2 message payloads, merged 2026-06-05). This appears to affect versions that include that change and enable MultiAgentV2 (post-0.137.0). Not subscription-specific. Not model-specific. This…
codex-cli 0.157.0 win terminal cmd Description: After installing the Codex daemon, several terminal windows briefly open and close during every request. This happens repeatedly and is distracting. Codex daemon version: 0.157.0 Client…
Codex websocket transport to wss://chatgpt.com/backend-api/codex/responses upgrades successfully and is then immediately closed by server policy (1008), leading to reconnect loop and fallback to HTTPS. Codex CLI: 0.106.0 (also reproduced…
reverted to the previous version
Codex persists a refreshtoken for a routed MCP server in ~/.codex/.credentials.json, but it does not automatically refresh the access token when the token expires. As a result, once the access token expires: MCP tool calls start failing…
unexpected status 401 Unauthorized: Incorrect API key provided: sk-svcacfvMA. You can find your API key at https://platform.openai.com/account/api-keys., url: https://chatgpt.com/backend-api/codex/responses, cf-ray: a40d95aa6bfcb8db-SEA…
Codex CLI 0.157.0 fails to start on Windows with the following error: Downgrading to version 0.156.1 resolves the issue. Install Codex CLI 0.157.0 on Windows: Launch Codex CLI: Observe the daemon privilege error. Codex CLI should start…
Downgrading to version 0.156.1 restores normal operation:
On Windows 11, the Claude Desktop window behaves as a topmost window: it stays drawn above other applications even after another window is focused and brought over it. There is no in-app setting to turn this off. This is the Windows…
On a Max subscription, the same workload uses about 1.8× more of the 5-hour window when it runs headless (claude -p --input-format stream-json, recorded as entrypoint: "sdk-cli" in the transcripts) than when it runs in the interactive TUI…
Bug Description VS Code extension update ~13 hours ago (Jan 23-24 2026) causes "API Error: 400 due to tool use concurrency issues" on every tool call. CLI works fine, only VS Code affected Error persists through extension reinstall, VS…
downgrade to vscode extension version 2.1.17 fixed the issue `uninstall the extension --> restart vscode --> install extension with 2.1.17 version -> fixed`
26.825.31414 Microsoft Windows NT 10.0.26200.0 x64 ChatGPT / Powered by Codex & OWL Release date shown in About: August 28, 2026 Windows 11 x64 with WSL2 / Ubuntu. Project management works correctly when the Codex Agent Environment is set…
I want to use a remote Hermes Agent instance while keeping tool execution on my local machine. Machine A: local client / workspace machine Machine B: remote Hermes Agent with my existing skills, memory, sessions, and model/provider config…
ChatGPT desktop on Windows, with an About-dialog release date of September 29, 2026. The user confirmed the app is up to date. Installed Windows package version: 26.928.1915.0 (read from local package metadata, not the About-dialog display…
After a desktop app was signed into account B and then back into account A on the same machine, pairing an Android phone with account A loops forever. After scanning the QR code, tapping Authorize this phone and completing the browser…
26.908.40834 Darwin 25.6.0 arm64 arm Trying to queue task after editing it once. So I had a queued task, then decided to edit it, and then I start to get this error. Submit queue, or submit queue edit it and try to submit again Queue being…
Installed Codex Desktop Windows package: 26.903.9818.0, verified locally with Get-AppxPackage. The executable used by the currently running Codex backend reports codex-cli 0.153.4 via --version. These were collected on September 11, 2026…
Codex Remote pairing fails on Android. ChatGPT mobile version 1.2026.265 and the Windows desktop app use the same verified personal account. “Allow connections” is enabled on the PC. After scanning a fresh QR code, the phone shows…
On Windows, the managed remote-control app-server daemon launches tool-runtime and MCP child processes in visible console windows. Starting or opening a session through codex agents produces a cascade of black windows for noderepl.exe…
Is there a way to restrict terminal command from a agent to access/edit file outside the current directory ? I see that gemini-cli or codex-cli use seatbelt on macOS. I don't see any equivalent for copencode.
26.924.1866.0 (Windows package version; the app does not load reliably enough to open About Codex) ChatGPT Plus Microsoft Windows NT 10.0.19045.0 x64 Codex Desktop remains indefinitely on the startup spinner before showing projects or…
Title: Windows app 26.1002.51308: sandbox setup fails with sharing violation when validating its own active runtime After updating the Windows Codex app to 26.1002.51308 (build 13417), all command execution fails before the requested…
Claude Desktop 2.37337.1, Windows Store/MSIX package (Claudepzs8sxrjxfjjc), self-updated on 2026-10-13 15:11 local time Windows 11 Pro 10.0.26200 Code tab, sidebar in sections layout (sidebarMode: epitaxy) After the update, the dedicated…
I cannot finish pairing my Android phone with Codex running in WSL. The phone opens the browser login, recognizes the correct ChatGPT account, and asks me to authorize access. After I authorize, it briefly shows “Pairing…” and returns to…
Since 0.157 the TUI talks to a shared codex app-server --managed-daemon (one per CODEXHOME), and lifecycle hooks (UserPromptSubmit, Stop, PostToolUse) run inside that daemon. The daemon inherits the environment of whichever TUI first…
What changed. On 14 October, #106742 (the unified gateway) moved every local conversation into one gateway process. The CLI, the terminal UI, the local Desktop app, the API server, editors, messaging bots, cron jobs, Kanban cards and Group…
The only workaround is config (`ws_orphan_reap_grace_s=0` plus a higher `HERMES_TUI_SESSION_TTL_S`). "I expected orchestration to continue with the laptop off." - [https://discord.com/channels/1053877538025386074/1563120848624164905](https://
A dot can start a local Codex task on a connected desktop, but its task-creation tool cannot select an existing saved Codex App project. A local Codex task can create a thread in that project, yet the dot then cannot read or message that…
26.924.20706 Linux Mint, X11, amd64 After updating from 26.917.71314 to 26.924.20706, every local Codex task hangs indefinitely at "Starting your task". This reproduces with both an existing project task and a brand-new trivial task…
Downgrade to 26.917.71314.
This has been reported but was closed for being stale. Reopening in a new ticket. The most recent upgrade to Claude Code regressed on https://github.com/anthropics/claude-code/issues/48550 which was auto closed for inactivity. I cannot…
26.1021.22027 ChatGPT Plus Visual Studio Code 1.143.1 (Windows) Microsoft Windows NT 10.0.19045.0 x64 Codex IDE extension version 26.1021.22027 fails to initialize its user interface in Visual Studio Code 1.143.1 on Windows. When opening…
Confirmed workaround: Downgrading the Codex extension from `26.1021.22027` to `26.1007.21434` immediately resolved the issue.
On Omarchy master, Hyprland 0.54.2, kernel 6.19.8-arch1-1 After updating Omarchy and rebooting my laptop, my browser logins/sessions were cleared and my GitHub CLI auth was also gone I have not fully diagnosed the root cause, but this…
I applied the documented workaround: blanked the `Default` keyring password via seahorse.
The Monitor tool now hard-rejects persistent and timeoutms above 3600000 (1 hour) even with persistent set, it rejects with a schema validation error, and even a Monitor started with the maximum now allowed (3600000) is reported as…
An active Codex session suddenly failed with: I did not update Codex or change feature settings while the session was running. Afterwards, attempting to resume the session produced: Importantly, this is not specific to the crashed session…
ChatGPT desktop 26.1007.21159, build 20052. The app’s updater reports that this version is current. ChatGPT Pro. macOS 26.6.2, build 25G83. After the latest update, the contextual menu shown when selecting assistant response text no longer…
Since 2026-10-12 at about 18:18 UTC, every newly started interactive session writes its main-conversation prompt cache under the 5-minute TTL. Headless claude -p runs on the same account, machine, version and model still write under 1…
copilot plugin update (and the same flow triggered from the GitHub Copilot desktop app) fails on Windows with: The git fetch/checkout of the marketplace clone succeeds every time (the clone under %LOCALAPPDATA%\copilot\marketplaces\…
On Windows 11, Computer Use cannot initialize and ordinary sandboxed shell commands fail before execution with helperunknownerror: setup refresh had errors. The sandbox log identifies an error 32 (sharing violation) when opening the…
On Windows, authenticating to an Entra ID–protected MCP server (Microsoft's hosted Azure DevOps MCP server, https://mcp.dev.azure.com/{org} , server type http ) fails on every fresh interactive sign-in attempt with: Authentication failed…
Claude Desktop 2.2553.1 (macOS 26.6.2, Apple Silicon) Bundled Claude Code 2.1.275 Remote Control in use daily from claude.ai/code while away from the machine When an update is staged, the app waits for a short idle window and then quits…
Claude desktop app: 2.37337.1 Code tab sidebar After a recent update, the Code tab sidebar no longer groups sessions by project/repository. Sessions are now grouped only by environment: "This computer", "Cloud", SSH host (e.g. ubuntu@ )…
I had Claude create a custom sidebar group for each of my 15 project folders and move ~127 sessions into them.
On Claude Max (20x), the weekly "all models" meter went from 0% to 54% in the first 22 hours after the weekly reset. My Claude Code token volume in those 22 hours was nearly identical to the same 22 hours of the previous week. The previous…
Remote compaction fails with a 401. The rejected key is an OpenAI service-account key (sk-svcacct-…), but that key isn't configured anywhere on the client. The session shows "Compaction interrupted" followed by "Reconnecting 2/…". The…
What I expect I run local scheduled tasks (routines) in Claude Desktop on my Mac. Each run used to show up as a normal session: in the desktop app's left sidebar, and in Remote Control on my phone, so I could open a run and continue it…
26.924.20706 Linux 6.6.141-09476-g954adab60416 x8664 unknown After updating ChatGPT Desktop on Linux to 26.924.20706, all local Codex tasks remain indefinitely on “Starting your task”. The standalone Codex CLI works correctly with the same…
After updating the Claude desktop app, the Routines section is gone from the Code tab sidebar. My local routines still exist and still run on schedule; only the section is missing. The sidebar has no control to show it again. What I found…
After upgrading OpenCode to v1.17.10, OpenCode started crashing with a native Bun segmentation fault on Windows. Downgrading to v1.17.9 appears to stop the fatal crash under the same repository/configuration/session pattern. This looks…
Downgrading to `opencode-ai@1.17.9` appears to avoid the fatal crash under the same environment and workflow.
After auto-updating to 2.1.269, text inserted by a voice dictation tool (Wispr Flow — clipboard + simulated Ctrl+V) is no longer inserted into the Claude Code prompt when Claude Code runs in the VS Code integrated terminal (Remote-WSL)…
Pin 2.1.268 (`ln -sfn ~/.local/share/claude/versions/2.1.268 ~/.local/bin/claude` + `DISABLE_AUTOUPDATER=1`), or run Claude Code in Windows Terminal instead of the VS Code integrated terminal.
In 0.0.389 we added support for Remote MCP Servers protected by OAuth that support the DCR (Dynamic Client Registration) standard. DCR allows an OAuth client to register itself with an OAuth server "just in time" instead of requiring the…
After updating Codex on September 25, 2026, a blank Windows Terminal window for codex-code-mode-host.exe repeatedly opens and closes while I use Codex. This did not happen before the update. The window title points to the app-server-daemon…
My Prometheus always stuck at “Preparing write...”。
A mod's ui.render output (e.g. an AbovePrompt band with a Button) is drawn in the terminal but never appears in the Claude desktop app or the Claude iOS app when they view the same session through Remote Control. Only mod-initiated…
Codex CLI: codex-cli 0.157.0 (global npm installation) Managed app-server daemon: 0.157.0-x8664-pc-windows-msvc Codex Desktop 26.917.9434.0 is also installed. The process chain captured below belongs to the active CLI session; I have not…
The schema for mcp.servers[].env (and mcp.servers[].headers) is: This means SecretRef objects are rejected — operators following security best practices around secret indirection (openclaw config set --ref-source env --ref-id ...) hit a…
The only documented workaround is the global `$OPENCLAW_STATE_DIR/.env`.
Everytime after upgrading copilot cli through HomeBrew the following message appears: Apple could not verify “copilot” is free of malware that may harm your Mac or compromise your privacy. Afterwards, to get it working correctly, I need to…
I believe this is a bug. I'll try to join the Continue Discord for questions I'm not able to find an open issue that reports the same bug I've seen the troubleshooting guide on the Continue Docs After using the Continue VSCode extension…
Codex Desktop and Codex CLI both began failing suddenly with 401 Unauthorized after working normally all day. Re-authenticating in both clients does not resolve the issue. Codex CLI: 0.155.1 Login status: Logged in using ChatGPT Remote…
When an HTTP MCP server is configured in .mcp.json with "type": "http" and "headers": { "Authorization": "Bearer " }, the CLI attempts OAuth discovery (/.well-known/oauth-authorization-server ) and fails with: MCPOAuthError: Failed to…
Manually enabling the server via /mcp enable bypasses OAuth discovery and connects successfully using the Bearer token.
Multiplayer agentic AI is the future. Hermes can and should lead. TL;DR: Memory operations bypass the hook system entirely, making tenant isolation impossible without forking core — we've been running a fix in production for months with…
Regression / performance On OpenClaw 2026.9.6 the Gateway exhibits a repeating memory "sawtooth": process RSS climbs steadily for several minutes, trips a memory pressure: level=critical event, an idle worker is retired under pressure, RSS…
VSCode Extension The "Thinking" section (extended thinking / chain-of-thought display) is no longer visible in Cline v3.50.0. This feature was working correctly in v3.45.1. Install Cline v3.50.0 Start a new conversation with Claude Ask any…
A /model pick made in any gateway chat (Telegram, Bot Chat, TUI) on a session that follows the profile config fails with an internal error ("Model switch failed") whenever the session policy carries private config secrets…
I'd like to contribute a Nextcloud Talk messaging-platform adapter for the gateway — the same shape as the existing Mattermost / Signal / Telegram adapters: Hermes both receives messages from and sends replies to a Talk conversation, fully…
Crash (process/app exits or hangs) I genuinely regret upgrading to OpenClaw 2026.9.5. Before this update, my environment was stable. After installing 9.5, I spent approximately 8 continuous hours investigating problems that simply did not…
After updating the Codex VS Code extension on October 1, 2026, pressing Enter frequently clears the composer, but the message does not appear in the conversation and Codex does not respond. This happens even when Codex is idle. Repeating…
Downgrading to extension version 26.917.62051 stops the problem.
Windows 11 (NT 10.0.26300) Kiro: 1.4.13 · Windows 11 (NT 10.0.26300) Server: https://mcp.atlassian.com/v2/mcp (remote, OAuth via DCR, no custom config) What happens Click Retry → browser opens Atlassian consent → "Authorization Successful"…
OpenCode Desktop exposes the "Custom OpenAI-compatible provider" form in Settings → Providers, but the save handler unconditionally throws provider.custom.unavailable, so the flow can never succeed on any server — including the app's own…
After the most recent Codex/ChatGPT desktop update on Ubuntu, the Codex desktop UI opens but a Codex chat remains stuck on the loading spinner and cannot be interacted with. This started immediately after the update. Bundled Codex CLI…
Omarchy: 3.8.3 | Arch Linux | Waybar custom/update module: "exec": "omarchy-update-available" On Omarchy, Waybar shows the "updates available" icon when the system is not up to date. This icon is controlled by bin/omarchy-update-available…
On startup, this warning is printed (twice in my case) before the agent begins processing: The messages render over already-drawn TUI content, garbling the screen. After the warnings, processing starts and works normally — solstice is a…
Linux Mint 22.1 Cinnamon When attempting to sign in to Kiro IDE (via Google, GitHub, or any other authentication provider), the IDE displays a message indicating it will open a browser for sign-in, but no browser window actually launches…
After updating past slack-sdk 3.44.0, the Slack working-status indicator ("is thinking…", live per-tool status via livestatus / typingstatustext) silently disappears on every turn. All status calls fail with invalidarguments; status…
After a recovery that moved the old session store aside, openclaw doctor stops at: The only holder of the agent DB is the running gateway itself. Deterministic on our host: Run OpenClaw 2026.10.5 as a systemd user service (gateway…
`P2`: A large legacy-history installation cannot reliably complete offline repair at default limits, with an advanced-operator workaround and a focused repair already proposed. - `impact:ux-friction`: Repeated offline repair failures require downtime and manual runtime tuning before the operator…
Codex 0.155.0-alpha.15 fails before executing any filesystem-restricted command on a native Ubuntu host when /proc/self/mountinfo contains valid nsfs entries created by snapd. The failure is: The affected mountinfo entries have an nsfs…
Today, Hermes Agent supports exactly one terminal backend at a time — set globally via terminal.backend in config.yaml (mapped to TERMINALENV). You pick local, ssh, docker, modal, daytona, or singularity, and every terminal() call routes…
Regression (worked before, now fails) After updating to 2026.3.2, any message causes embedded agent to fail with "Cannot convert undefined or null to object". Update to 2026.3.2 (npm install -g openclaw@latest) Set model to…
With discord.autothread: true, every automation wake that targets a session whose thread was auto-created is refused at admission: The drain requeues the event and retries forever, so parent agents never resume after their children finish…
Omarchy 3.8.3 PC: Framework Desktop (AMD Ryzen AI Max 300 Series) CPU/APU: AMD Ryzen AI Max 385 w/ Radeon 8050S (Strix Halo, gfx11) GPU: amdgpu 0000:c2:00.0, VRAM 16384M (UMA), ATOM BIOS 113-STRXLGEN-001 SMU firmware: 0x0a640600 (100.6.0)…
I changed only `blankProcess` in `Service.qml` from:
In an interactive VSCode-extension session, Claude Code declares the MCP elicitation capability during initialize, but auto-declines every elicitation/create request without showing any UI. The same session can render permission prompts…
Am I the only one no longer able to figure out how to switch off automatic work tree creation in the newest version. Updated to the latest claude code to do development on the desktop app and now it seems I'm forced into using worktrees?…
/home/ /.openclaw/tmp/plugin-captures/ accumulates multi-GB session directories from plugin-build and model-catalog operations. These directories are never cleaned up after the operation completes, so disk usage grows permanently with each…
Manually delete orphaned dirs where `owner.sqlite` is empty and no process holds a file handle:
26.917.51856 Ubuntu 24.04.5 LTS (Noble) After the ChatGPT Linux application updated from 26.915.31029 to 26.917.51856, Codex can no longer execute commands using its normal Linux sandbox. The failure is reproducible directly using the…
Windows package version observed in app logs: 26.928.1915.0. Desktop client version reported in runtime logs: 26.928.20755. Bundled WSL CLI verified with --version: codex-cli 0.159.0. Not provided. Codex Desktop on Windows, with the agent…
After updating Claude Desktop, the Routines section disappeared from the Code tab sidebar. Previously every scheduled task run appeared there, so I could see at a glance which tasks ran, when, and whether they succeeded. Now there is no…
This broke overnight. We've been using the same Azure MCP registry for months fine. Copilot CLI Rust runtime fails with BrokenPipe when validating Azure API Center MCP registry. GitHub Copilot CLI 1.0.83 cannot validate MCP servers against…
codex-cli 0.157.1 Microsoft Windows NT 10.0.26200.0 x64 Windows Terminal + Powershell On Windows, a batch of visible console windows flashes for git.exe / pwsh.exe (and other helper) child processes: several git.exe windows on every new…
After updating to Copilot CLI 1.0.100 (which also pulled a new voice runtime), voice input no longer works on WSL2. The voice server process segfaults (signal 11) a few seconds after starting, so /voice stays on "Fetching voice models" and…
26.1007.71326 (build 20398), production channel Updater reports plantype=promax Darwin 27.2.0 arm64 arm — macOS 27.2 (26B5112a), Apple Silicon Selecting Check for Updates shows: The failure happens while fetching the update feed, before an…
Regression (worked before, now fails) Since OC v2026.5.12 + codex-cli 0.130.0 (first observed 2026-05-15), the Codex app-server rewrites the full prompt between a user turn and its tool-call continuation turn instead of appending only the…
On macOS, MCP OAuth entries in the shared Keychain item Claude Code-credentials get rewritten with empty accessToken/refreshToken (and no clientId) while the refresh tokens are still valid, when several claude processes touch the store…
The gateway's prepared-model-catalog.worker.js worker thread leaks memory continuously and monotonically, independent of workload volume, model provider, or plugin configuration. On an idle-to-light-traffic single-user install, RSS grows…
The master branch in this repo was deleted apparently around September 14th. This breaks the omarchy-update-git step for any system where the local git copy is set up to track the master branch. Since the documented flow for upgrading to…
With Allow network egress = ON and Domain allowlist = "All domains" (Settings -> Capabilities -> Code execution), the session egress proxy still rejects every host outside a small built-in allowlist with 403 on CONNECT. This is a…
MCP servers can expose data through the resources primitive (resources/list, resources/read). This is one of the three core MCP primitives alongside tools and prompts. Currently, Copilot CLI only supports MCP tools. When an MCP tool…
Published artifacts no longer show a "Version history" option anywhere: not in the claude.ai web viewer's title menu, and not in the Claude desktop app's artifact side panel. Earlier versions are still being saved (each publish returns a…
Plugin tools can't raise their own execution deadline. PluginContext.registertool() (hermescli/plugins.py) does not accept or forward executiontimeout, but tools.registry.register() supports it and…
`timeouts.tools.sequential_call: 960` in `config.yaml`. This raises the ceiling for every tool, so a wedged tool now blocks for up to 16 minutes instead of 7. It also leaves the concurrent-batch path at 420 s.
Powered by Codex & OWL Version 26.928.21956 Microsoft Windows NT 10.0.26200.0 x64 no-active-thread-01a0f3f0-94ff-7451-83bf-3238585cc8fe Subject: Codex desktop on Windows with "Run agent in WSL" enabled: every command fails with "Failed to…
ubuntu 24.04 using ssh to it ubuntu default terminal i cant fucking copy text, the reason codex-cli is the best was because i could fucking copy text normally from my cli , and it was feeling like a cli not like a fucking GUI application…
Hey! Whenever gemini-cli defers to the generalist agent it hangs forever. Simple changes like folder creation hang. I've let it wait for up to an hour before cancelling it. Instructing the model not to defer to sub agents resolves this…
I’m experiencing a Windows ↔ Android Remote pairing failure. Windows ChatGPT desktop: 26.904.11930 Android ChatGPT: 1.2026.265 ChatGPT Plus / Personal workspace Same account and workspace on both devices Windows generates the QR normally…
A user-configured stdio MCP server connects successfully (status: connected) but always reports toolcount: 0, across brand-new sessions, app restarts, and even a full OS reboot. The problem persists indefinitely — no retry ever recovers…
Regression (worked before, now fails) On 2026.9.5 the prepared model-catalog worker keeps creating plugin source captures under $OPENCLAWHOME/tmp/openclaw-model-catalog-/openclaw-plugin-build- and never deletes them. On our gateway this…
a cron every minute that, inside each `openclaw-model-catalog-` folder, keeps the 10 oldest captures (the ones created at worker start) and deletes the rest older than 1 minute, by creation time (mtime keeps changing).
GitHub Copilot desktop app 1.1.22 (commit c3daa01), macOS arm64, bundled CLI 1.0.84-5 (server mode, stdio, Rust JSON-RPC engine) Single github.com account signed in to the app; gh auth status (bundled gh 2.98.0, using the app-exported…
On Windows, the Claude Desktop app's stealth auto-update deadlocks. After an idle timeout, the app tears down its window and any running Claude Code sessions and logs beforeQuitForUpdate handler fired, going down for update, but the main…
`tasklist /apps | findstr /i claude ; Stop-Process -Id -Force` Then relaunch.
its cli is thinking and thinking and loop donot breaks event it took 2 hours for completing a task which was just be done in 2 minutes in previous version, its started happening in this latest version. it also sometimes in thinking loop…
extensions/openai/openai-provider.ts always queries the Codex models endpoint with the managed runtime's version: However, the codex plugin officially supports external app-server binaries via plugins.entries.codex.config.appServer.command…
Forbidden, OpenCode Free Model Gateway Detect Abuse Traffic
On v0.21.6 the version line still carries the release date of v0.21.5: The version is right (it comes from the install stamp), but the date in parentheses is hermescli.releasedate, which is still "2026.9.24" in the v0.21.6 tag and on main…
Whenever I'm signing in with Google - the Gemini CLI is again asking for authentication and this is never ending One time authentication only Client Information Run gemini to enter the interactive CLI, then run the /about command.
The security-guidance plugin (2.0.8) runs hooks/sg-python.sh on SessionStart, every UserPromptSubmit, every PostToolUse on Edit|Write|MultiEdit|NotebookEdit, PostToolUse on Bash, every Stop and every SubagentStop. On every one of those…
After updating to opencode 1.18.30 via pacman, every single prompt fails immediately with: Failed to send prompt. Unexpected server error. Check server logs for details. This happens on a brand new session, on the very first message, with…
Downgrading to 1.18.18 (via the cached pacman package) fixes it immediately — same machine, same everything else.
Unable to install OpenCode on windows with winget command I follow the instruction described in Changelog: https://github.com/anomalyco/opencode/blob/dev/CHANGELOG.md#highlights-3 The root cause…
Track update, upgrade, Doctor, migration, rollback and restart reliability around OpenClaw 2026.9.3 / 2026.9.4. This coordination index does not replace the linked reports or their evidence, and does not authorize closing, merging or…
Hermes v0.14.0 (latest, updated 2026-05-27 with 471 commits pulled) crashes on every hermes chat call when using openai-codex provider with gpt-5.5. The ChatGPT Codex backend (chatgpt.com/backend-api/codex) returns output: null in…
After upgrading from V1.18.31 to V2.0.8, sessions that V1 stored in the global project (sessions created in non-git directories) become invisible in the TUI session picker, because V2 creates a per-directory project and the picker filters…
Before repair: 104 sessions total, 93 with `project_id = 'global'` (81 in `/home/wjf`, 8 in `/tmp/opencode`, 3 in `/home/wjf/novel`, 1 in `/home/wjf/videos`). - Sessions in a git repository were unaffected: `/home/wjf/fadebreak` kept a stable project id across the upgrade and stayed visible. -…
Claude Desktop (enterprise 3p deployment, deploymentMode: "3p"), macOS (Apple Silicon), app 1.24012.1 Embedded Claude Code: 2.1.236 (second machine: 2.1.270), claude-desktop-3p entrypoint Inference: custom gateway via…
A single transient 429 / modelentitlement cooldown on one credential benches that credential (and often the whole provider row, or every auxiliary call) for hours to days, with no reset-time probe, no visible "in cooldown until T because…
Proposal: goose doesn't support A2UI today. This issue proposes adding it to goose desktop, starting with rendering A2UI surfaces returned by MCP servers. The first version would be off by default and limited to the A2UI basic catalog…
gnome-keyring-daemon crashed with SIGABRT on 2026-09-13 at 01:38:02 CEST while handling D-Bus requests. Signal: SIGABRT (from glogdefaulthandler calling abort()) Root cause: NULL client pointer in gkdsecretservicegetpkcs11session cascading…
When hermes update / the Desktop updater fails, the install is left half-applied (Desktop launches then closes, gateway venv missing pydanticcore, git checkout carrying partial-clone config) and the product prints the raw underlying error…
In a forum-enabled bot DM (Telegram "Threads" feature), when the user creates a new topic via the client's "New thread" composer, Telegram now delivers the opening message without messagethreadid (thread-less), while the forumtopiccreated…
Hi! I'm developing a skill that uses the following stack: Kimi K3; from OpenCode Go Bitwarden; for secrets (using the "official" resource from Hermes docs) Browserbase headless; integrated in Hermes (free) without API key - as is. Hermes…
Telegram init hangs pre-socket inside the gateway process after a routine network blip, while the identical stack (same venv, same token, same transports, same SSL context) connects in 0.3s standalone. Appears to be a new variant of the…
Since upgrading to 2026.9.6 (eb377ac) the prepared-model-catalog.worker.js thread grows by roughly 1 GiB every 5 minutes with no user action, reaches 8 to 13 GiB, and the gateway's memory-pressure reclamation then resets it to under 2 GiB…
A session served by a long-running claude remote-control host (server mode, same-dir spawn) can't be used again after it's archived and then unarchived from claude.ai. The web UI accepts the unarchive, but a message sent to the session…
`claude remote-control --session-id ` in the same directory. This starts a separate single-session host. - Calling `bridge/reconnect` for the session, as above. This puts it back into the existing host.
This is not a cosmetic redesign. My classic claude.ai chat projects, which I used as sealed folders of separate private conversations with their own instructions, were converted to the coordinator-based Projects without my consent, without…
Windows 11 Business 25H2 Summary Over the past several weeks Kiro has become unreliable across core operations. File edits fail, the IDE becomes unresponsive, and spec (design/task) operations fail repeatedly with a backend network error —…
Running Gemini CLI with MacOS seatbelt enabled: And then starting Gemini CLI in a folder that contains a Git repository instantly crashes the CLI. Folder without a Git repository: Initialising this folder with a Git repo: Not to crash…
On Linux, launching Hermes Desktop again while it is already running (the launcher, gtk-launch, or a hermes:// deeplink that starts the binary) can leave windows-sandbox-fallback.json stuck at {"state":"fallback","reason":"boot-loop"}…
Regression (worked before, now fails) Heartbeat IsolatedSessions is in the live docs but not in the latest version 2026.3.13 Add IsolatedSessions to your heartbeat config as per the docs…
AMD Ryzen 9 5900X, NVIDIA RTX 2080 Ti, Omarchy 4.0.4 Today's update pulls Qt 6.12.0 from extra, but quickshell 0.3.1-1 is still built against Qt 6.11.2. After the update and a shell restart: The bar is gone. Every menu (Super+Space…
downgrade the qt6- packages plus fcitx5-qt, python-pyqt6, qt6ct and telegram-desktop from the pacman cache, then add them to IgnorePkg until quickshell is rebuilt.
On OpenClaw 2026.9.7 (c074824), an existing Codex-backed conversation intermittently fails before a reply with: The first failure followed an in-place session model switch from openai/gpt-6-astra to openai/gpt-6-sol. A…
⚠ ⚠️ CRITICAL SECURITY WARNING: Untrusted Command Flags Detected The following flags/arguments were sourced from external untrusted input (e.g., Buganizer ticket, external document): Executing commands with flags sourced from untrusted…
then rolled back to v0.62.0 via npm install -g @google/gemini-cli@v0.62 which works.
Omarchy: 4.0.0rc5-1 Kernel: 7.1.8-arch1-3 CPU: AMD Ryzen AI MAX+ PRO 395 GPU: AMD Strix Halo / Radeon 8060S Migration 1786643346 repeatedly asks me to close the browser windows and continue even after Chromium is completely stopped…
Window 10/11 Kiro CLI 2.31.1 Kiro CLI 2.31.1 V2 engine: Working V3 engine: Affected Custom agent configuration: JSON (resources) Kiro CLI V3 appears to have two regressions in custom agent resource handling compared with V2. In V2, skills…
Using `file://` as a workaround is not equivalent to registering a skill for progressive loading.
Immediately after installing the latest macOS security update and rebooting, all Copilot CLI sessions—both newly started sessions and resumed existing sessions—became unable to process prompts. GitHub Copilot CLI 1.0.90-3. macOS Golden…
Regression (worked before, now fails) The buttons are overlapping and you can't use the sidebar buttons Start v2026.9.7 on Linux, I'm using Ubuntu 26.04, installed as a .deb Do something to open the side panel See the icons all…
What happens: hermes -z "…" (and any entrypoint that ensures the gateway) fails with Error: Gateway inaccessible: servicemanagerunavailable when run without a GUI login session — e.g. from a system LaunchDaemon with UserName= , or via sudo…
Attempting to use the official HTTP MCP server from github: https://github.com/github/github-mcp-server I was able to get it to work with a Github Personal Access Token(PAT), with the following config: but /mcp auth githubCopilot results…
hermes kanban swarm's verifier role hardcodes the skill requesting-code-review (hermescli/kanbanswarm.py:238). This skill ships bundled (skills/software-development/requesting-code-review/), so it normally exists — but the Curator's…
In the Claude desktop app (Code tab), Remote Control only stays connected while a session's process is running. Once a session finishes its turn and goes idle, the app shuts its process down, and the Remote Control link drops with it. The…
The filter button next to "+" at the top of the session list used to let you show archived sessions. After the update it doesn't. The only way to them now is "View All", which opens a whole separate panel listing every session. You archive…
On a self-managed (shell-installer / PM) Hermes install, every scheduled cron job fails before it can run. The restart-safe external worker is spawned as sys.executable -m cron.scheduler with a sanitized environment whose PYTHONPATH…
Any agent defined via .kiro/agents/.json does not receive MCP tools in its tool list. Only the built-in kirodefault agent receives MCP tools. MCP servers initialize successfully in all cases, but tools are only injected for the built-in…
`kiro-cli chat --agent kiro_default` The built-in `kiro_default` agent correctly receives MCP tools from global `~/.kiro/settings/mcp.json`. Custom agent prompt/resources must be loaded manually at session start.
A Windows install that moved to the PM runtime (managed Python 3.14, dependencies under installs/ /environments/ /venv) still has the old pre-PM \venv (Python 3.11) on disk. After hermes update --gateway respawned the messaging gateway…
On macOS 12 (Monterey, Intel), the Desktop app's model picker shows Opus 5.5 greyed out with the tooltip "Update Claude to use this model on this computer", while the app's own update check reports "You're using the latest version." The…
(Note: this is copy/pasted from google-internal bug https://buganizer.corp.google.com/issues/429600201#comment13, with internal links removed... they're still available on buganzier, though.) ------------- We are seeing a large number of…
Regression (worked before, now fails) On the Codex app-server (ChatGPT Plus sub), a multi-tool agent turn reliably fails with Codex stopped before confirming the turn was complete starting in 2026.5.27; 2026.5.26 completes the same turns…
Only found workaround is pinning 5.26.
The Figma hosted MCP server (https://mcp.figma.com/mcp) authenticates and initializes successfully, but the CLI never registers its tools. The CLI's discovery probe receives -32601 Method not found, records it as a fatal server failure…
Update to a newer CLI build that supports Figma's `2026-07-28` MCP protocol handshake.
We use OpenClaw daily — multiple agents, scheduled cron jobs, and a long-lived state directory on Windows. We just spent two months fighting it, and only after rolling back from 2026.9.6 to 2026.7.35 (Extended-Stable LTS) did we get a…
The Omarchy shell (quickshell) segfaulted at 19:57:33 while I clicked on its UI (bar/panel element). Quickshell's in-process crash recovery restarted the engine ("Quickshell has been restarted") and the bar came back by itself. No data…
Dell Pro Max 16 Worked well up until 3.8 upgrade today, now the screen wont turn back on after suspend (have to hard reset): Anyone else seeing this?/
MacBook Pro 13" (2020) with an Intel chip (1,4 GHz Quad-Core Intel Core i5) I downloaded the Omarchy (v3.0.2) ISO from the website and created a bootable USB using this guide. It boots successfully — I see the Omarchy logo when the…
When updated to lastest veriosn of Omarchy (3.4). The native laptop speaker is not working, which was working fine pre-update. Bluetooth works fine, and other external audio peripheral works good
openclaw-2026-07-15.log Regression (worked before, now fails) gateway doesn't start with manual launch Error: gateway did not start on 127.0.0.1:18789 Pls have a look! update to openclaw 2026.7.1 openclaw gateway run --force CLI is…
pm/workspace.py::membersstamp() hashes every file in each enabled plugin member directory to decide whether the recorded dependency environment is current (pm/install.py::venviscurrent). The only exclusions are the hard-coded MEMBEREXCLUDE…
In ACP mode (copilot --acp), the OAuth token for a remote HTTP MCP server is never written to disk or the OS keychain after a successful sign-in. Each new ACP session (for example, each new thread in an ACP client) starts unauthenticated…
I believe this is a bug. I'll try to join the Continue Discord for questions I'm not able to find an open issue that reports the same bug I've seen the troubleshooting guide on the Continue Docs When I open a project through Jetbrains…
A git-installed Hermes update on Ubuntu 24.04 failed while preparing the new isolated dependency environment during the historical PM takeover. The selected features included matrix; mautrix[encryption] pulled python-olm==3.2.16, which…
This is about the Claude in Chrome browser extension, not the Claude Code CLI. In Vivaldi, the extension's new side panel (the embedded claude.ai Cowork interface) never works. It only shows: Up to extension 1.0.93, you could work around…
What problem would this solve? MCP has no standard way to ship workflow instructions alongside the tools they describe. Server instructions are size-bounded, so complex workflows do not fit. Today a server and the skill that teaches how to…
A long-lived Copilot CLI session becomes largely unresponsive. Opening/refreshing tasks shows: The process stays alive, and some nested tool activity continues, but the UI/event reader repeatedly fails to acquire the session lock. macOS…
When authenticating the built-in Slack MCP server ( mcp.slack.com ) via Copilot CLI, the generated OAuth consent URL requests the full superset of Slack scopes — including write scopes ( chat:write , files:write , channels:write …
After upgrading claude-desktop from 2.31226.1 to 2.37337.1 (apt, stable) on Ubuntu 26.04 (GNOME/Wayland), every click on the "+" next to a project group in the Code tab sidebar shows a toast: "Couldn't switch environment: the Local…
Linux (Ubuntu) cloud server in Tokyo kiro-cli 2.27.1 Account: Kiro Pro, individual account, signed up with Google (Gmail) sign-in. No organization or administrator. kiro-cli 2.27.1, Linux (Ubuntu) cloud server in Tokyo Headless mode, auth…
Opening a conversation that belongs to a messaging platform (Telegram DM/group, WeChat, Feishu, …) in Hermes Desktop leaves the composer's reasoning-effort pill in its loading state indefinitely. It never paints the chat's actual level…
Regression (worked before, now fails) OpenClaw: 2026.9.7 (c074824) Node.js: v24.21.0 Linux [vps].de 5.2.0 #1 SMP Mon Sep 30 15:36:27 MSK 2024 x8664 GNU/Linux Gateway: system-wide systemd service Service file…
JetBrains Plugin I am using a beta version of Cline Today after IDEA update Cline plugin open with empty window. IDEA version: 2026.2 (Build #IU-262.8665.258, built on July 15, 2026) Open plugin window IntelliJ IDEA 2026.2 Build…
OpenCode fails before sending any prompt to the model. The UI shows: The server log shows: The error occurs repeatedly for new sessions and different working directories. Observed output: macOS 27.0(Apple Silicon)
openclaw plugins reload --wait parks indefinitely when the named plugins hold long-lived retained work. While it is parked, no isolated agent run on the Gateway can start — automations for every agent fail with cron: isolated agent setup…
Cline Desktop 0.0.38 (core 0.0.88) cannot connect to a Linux aarch64 SSH remote. The bundled cline-remote-helper-aarch64-unknown-linux-gnu (sha256 93444d160eed1dc9…) exits 1 on --remote-hub-ensure with a Bun SyntaxError: Invalid character…
`mkdir -p ~/.cline/remote-helpers ; scp :~/.cline/remote/cline-remote-helper-linux-arm64-8b51183e615e7442 \`
After the 2026-10-10 omarchy update (which pulled Arch's Qt 6.11.2 → 6.12.0 transition together with quickshell 0.3.1 → 0.3.2), pressing SUPER+SPACE shows a fullscreen opaque black rectangle instead of the menu. The menu itself is…
I would like to report a security-relevant issue in Aider's code generation behavior around sensitive file access outside the project directory. In a validated retest, Aider was gradually steered through a multi-turn conversation and…
Hi maintainers, Opening an issue first to gauge appetite before any code work, in the same shape as #12423. Would maintainers accept a third-party integration entry for Vaara (AGPL-3.0-or-later, https://github.com/vaaraio/vaara) used as a…
Regression (worked before, now fails) After upgrading from 2026.7.2-beta.1 to 2026.7.2-beta.2, the shared SQLite migration creates indexes referencing managedoutgoingimagerecords.agentid before adding that column, so doctor --fix fails…
During installation I selected Swiss German as the keyboard layout. The installer used that layout correctly. After the first boot into the desktop, though, the keyboard was US English everywhere. The layout chosen in the installer should…
`~/.config/hypr/input.lua`: `hl.config({ input = { kb_layout = "ch", kb_variant = "de" } })` - `~/.config/fcitx5/profile` with `DefaultIM=keyboard-ch-de`, then restart fcitx5.
What problem would this solve? goose can already ask before a tool runs (approve, smartapprove) and can deny a tool entirely. That is a permission policy, not a sandbox. Once a command, MCP server, or extension subprocess is allowed, it…
After installing Ghostty (ghostty 1.3.1, Arch extra), clicking Ghostty in the app menu does nothing — no window, no error. Running ghostty (or /usr/bin/ghostty --gtk-single-instance=true) from another terminal opens a window fine. The…
OpenCode Desktop 2.0.28 fails to complete ChatGPT OAuth authorization on Windows. After selecting the ChatGPT sign-in option and proceeding through the browser authorization flow, the browser displays Authorization failed with the message…
Describe the bug: On Windows, a remote MCP extension fails to load after a successful authorization. The browser shows Goose's own "Authorization Success" page, then the extension errors. The real cause is that saving the credential fails…
Additional context: GOOSE_DISABLE_KEYRING=1 is a complete workaround, which confirms the diagnosis.
Continue plugin crash : IntelliJ IDEA 2026.2.1 Build #IU-262.9437.185, built on August 9, 2026 Source revision: 5ced7b0cd2861 Licensed to CNRS / DSI / David Marques You have a perpetual fallback license for this version. Subscription is…
An OAuth-protected HTTP MCP server that has a valid, unexpired access token in ~/.copilot/mcp-oauth-config is answered with HTTP 401 on every new --acp session, and the CLI responds by opening the OAuth authorization page again. Because…
Omarchy 4.0.4-1 (updated from 4.0.3-1 earlier on the day of the incident) Kernel currently running after recovery: 7.2.5-3-omarchy; linux-omarchy 7.2.5-3 and headers installed during that update Previous stock kernel linux 7.2.3.arch1-3…
3.0.69 (npm cline, Windows x64); same code on main (7a75e99) Tools that an MCP server adds during a session never reach the model in the CLI. The server sends notifications/tools/listchanged, but the CLI keeps the tool list it built when…
After upgrading to an environment that has httpx2==2.12.0 (the pin Hermes carries for the mcp extra since bc8545bd1a, "bump httpx2/httpcore2 2.7.0 -> 2.12.0 for the audit advisories"), any HTTP MCP server whose tools/list response arrives…
Long-running work is a first-class citizen in agent-driven development — builds, test suites, dev servers, watchers, interactive programs (debuggers, REPLs), and scheduled follow-ups. Today goose has no first-class way to run that work in…
Intel i9-9880H, Intel UHD 630 + NVIDIA Quadro T2000 Mobile, Omarchy 4.0.4-1, mise-bin 2026.9.12-1 A wrapper generated by omarchy-mise-install can call itself forever. It never exits, and it pins a CPU core by running mise use -g about…
A long-running Copilot CLI process permanently loses authentication. After the failure, every normal prompt and /ask immediately returns an authorization error. Running /login does not recover the process. Restarting and resuming the same…
This machine's fingerprint sensor — a Dell XPS 15 7590 with the Goodix 27c6:5395 ("Goodix HTK32") — has no driver in the pinned libfprint-git snapshot. The community fork libfprint-goodix53x5 (upstream, built on libfprint tag v1.94.10…
The gateway-owned group chat workspace (canonical-group, Desktop v0.22.2) lost several capabilities the legacy Desktop room UI had. Five gaps, all in the same surface: Composer: no clipboard image paste, no drag-and-drop — a bare in…
kiro-cli acp does not answer session/new until it has downloaded the 83 MB all-MiniLM-L6-v2 semantic-search model. On a fresh machine that took about 66 s here, with no output, no progress notification and nothing on stderr. An ACP client…
After updating a git/source installation to Hermes Agent v0.21.5+2959.ga342e65 (commit a342e6563d), the automatic migration of the Hindsight memory provider from core to the plugin catalog gets stuck while preparing the plugin's Python…
The Gateway exits with SIGBUS (code=dumped, status=7/BUS) when a read-only session worker shrinks an agent SQLite WAL-index file (openclaw-agent.sqlite-shm) to 3 bytes while the Gateway still has it memory-mapped, during concurrent agent…
I am using a beta version of Cline When configuring the OpenAI Compatible provider in Cline Desktop, clicking on or interacting with the Base URL and/or API Key input fields automatically overwrites any custom input with the default URL…
On macOS, every update started from the Desktop app's Update button fails with exit code 2 because hermes update refuses the lock held by its own hand-off: The PID named in the error is the custodian that scripts/desktop-update/posix.sh…
quit the Desktop app and run `~/.hermes/hermes-agent/.hermes/bin/hermes update` in a terminal.
On the official Docker image (linux-arm64, v0.21.6) with HERMESHOME on a case-insensitive volume (macOS/APFS default — a plain Docker Desktop bind mount under /Users/...), no dependency environment can ever be committed for the install…
The OpenCode Go gateway (https://opencode.ai/zen/go/v1) enforces a 30-image-per-request cap on only some routes: enforcement follows the x-opencode-session routing value and the model. Identical requests (same API key, same model, same…
Workaround used: temporarily switch the model/provider, compact there, switch back. - The same request works or fails purely based on the session header, which is extremely hard for users to diagnose. - Compounds #47493 (no client-side media strip) and #39677 (image-count 400 not classified as…
Windows 11 Pro 10.0.26100 Claude desktop app, UA: Claude/2.7032.0 Chrome/152.0.7977.130 (MSIX) Hostname mapped to 127.0.0.1 via the Windows hosts file, e.g. dev.example.test; IIS over HTTPS on 443 Repro (run in the pane's page context…
In a long-running Copilot CLI session connected to Visual Studio Code, an automatic managed-settings refresh can fail while reloading the dynamically contributed IDE MCP server: After this failure, the session enters a fail-closed…
goose has its own repository for MCP servers but given that there is an MCP server for everything now, this concept is outdated. goose has deeplinks to install remote MCP servers but uses its own format for that. Meanwhile the world has…
0.0.26 (Desktop; CLI cline.exe bundles Bun 1.3.13) I am using a beta version of Cline Cline Desktop (and the cline CLI) are unusable on CPUs without AVX support because both bundle the standard (AVX2) build of Bun v1.3.13 instead of the…
https://geminicli.com/extensions/ https://geminicli.com/extensions.json The gallery entry for logospell/logospell-mcp-server still shows the extension as it was on October 7, 2026, and hasn't picked up any manifest change since. In…
Every MCP tool call that omits an optional parameter is rejected with a Zod-style invalidtype / expected: "nonoptional", received: "undefined" error for that parameter — even though the tool's own JSON Schema declares a default for it…
explicitly pass `fields` and `attribute_keys` (e.g. as a real list, or as the literal string `null` for list-typed fields — note this is not JSON `null`, it becomes the one-element list `["null"]` server-side, which is itself a separate minor bug but harmless for read calls).
hermescli/kanbandbdispatch.py::pidrecycled has a string branch for the composed worker fingerprint (" | ") that compares the recorded value to a freshly computed one with EXACT equality. On hosts without /proc (macOS) the start-time half…
kiro-cli 2.24.0 on macOS (Apple Silicon) Remote MCP server over streamable HTTP, OAuth 2.1 with dynamic client registration and PKCE (S256) The authorization server's metadata has "authorizationresponseissparametersupported": true The same…
Component: Workflow engine / orchestration (background workflow runs) Severity: Low functional impact (run is inert), but it's an unkillable orphaned run that clutters state and can't be cleaned up through any available in-product…
Omarchy 4.0.3-1 → 4.0.4-1 (update ran 2026-09-16; 4.0.4 released 2026-09-15) - Lenovo laptop, hybrid graphics: Intel iGPU + NVIDIA RTX 4070 Max-Q (AD106M, 10DE:2860), supergfxctl/supergfxd + nvidia-powerd active, Hyprland/Wayland - Driver…
On Gemini CLI v0.54.0, selecting "Sign in with Google" and completing the browser flow with a personal Google account that has an active Google AI Pro subscription returns the CLI to the "Enter Gemini API Key" screen. The CLI silently…
incorrectly binds the OpenAI provider to the OpenCode Zen integration, causing a Zen API key to be sent to ChatGPT’s OAuth-only Codex endpoint. OpenAI rejects the key with “API keys are not supported by this endpoint.” The correct OpenAI…
On OpenClaw 2026.9.5 (build ec9c1a1, installed via npm on 2026-09-18), the gateway node process exhibits a sawtooth RSS pattern: process RSS jumps by 1–2 GiB within 30-second windows while V8 heapUsed stays between 786 MB and 1 GB. The…
After upgrading to OpenClaw 2026.7.2-beta.2, openclaw gateway restart (or openclaw doctor --fix) fails with: followed by a chain of noncanonical X schema that cannot be repaired errors. Gateway cannot start, systemd enters a failed state…
resolvehermesargv() in hermescli/kanbandbdispatch.py decides the kanban worker's spawn argv using an importability check evaluated in the gateway (parent) process, but that result does not hold for the freshly-spawned child. On installs…
Pin `HERMES_BIN` to the shim for the gateway via a systemd drop-in (survives self-update, unlike a source patch):
Hermes Agent intermittently fails with APIError and servererror when using the openai-codex provider with gpt-6.1-sol. The issue occurs in both Hermes Desktop and Discord Gateway on Windows 11. Some requests complete successfully…
Desktop (Windows, 1.53.0): turn-final assistant messages ([thinking + text], no tool call) missing from chat view, while fully stored and re-played by the server Goose Desktop 1.53.0, Windows 11 Pro; provider ollamacloud Models…
In Telegram DM topic mode, when a topic is actively running, typing a new message in the All lane to start a new topic: creates a new (empty, dead) topic, and the message text lands in the actively running topic/session instead — the new…
On an idle-ish 2026.9.6 gateway, SQLite store/read workers are created and torn down about 7 times per minute. Each one boots the full worker runtime (about 2,000 module files, all plugin manifests) with a 512 MB heap limit and exits after…
After updating VS Code, the Claude Code panel is blank when connected to WSL (Ubuntu). The extension activates normally and authentication works (Output log shows no errors), but the webview fails to load its files: GET…
After I sign in through the Microsoft Entra account broker for an HTTP MCP server, the token cache file stays empty. Every new process shows the account picker again. A /restart in the same process does not prompt…
After a config hot reload of an agents.entries. change, the plugin setup inventory was retired but several callers kept using it. For about 1.5 hours, until a Gateway restart, they failed with Plugin setup inventory has retired / Plugin…
macOS 27.0.1 (arm64) kiro-cli 2.32.0 With the local command sandbox enabled, ssh and ssh-keygen exit immediately with No user exists for uid 501. OpenSSH calls getpwuid() on startup and gives up when it fails, so the Seatbelt profile…
I am use Gemini Pro account login, but keep met the following "Resource Exhausted" issue. if i use API key login, then dont have this issue. How to check if my account be flagged abuse? no 429 isse. Client Information Run gemini to enter…
What happens: During a spec task-execution session, the task-status tools and the spec-task-execution subagent dispatch repeatedly become unavailable ("Tool X is not available"), then come back after a restart, then drop again later in the…
Add harness-level progressive tool discovery so Goose can keep most MCP tool schemas out of the model context until they are relevant. A small always-visible discovery tool could search across configured MCP servers using server/tool…
In ACP mode, a runshellcommand call that prints output ends with a toolcallupdate whose content is empty, so an ACP client never sees the command's output. toToolCallContent in packages/cli/src/acp/acpUtils.ts forwards returnDisplay when…
AdminSettingsChangedDialog.test.tsx tests the lowercase restart key twice: The standalone test at lines 26–41 renders the dialog, writes 'r' to stdin, and checks that handleRestartMock was called. The parameterized test at lines 43–58 uses…
Intel Core i5-1335U / Iris Xe (HP Pavilion 15-eg3xxx), Omarchy 4.0.5, Hyprland 0.56.2, Wayland, gtk4 1:4.24.1 After the 2026-10-18 update, Hyprland stopped blurring the background behind Ghostty. The window is still translucent and the…
Behavior bug (incorrect output/state without crash) MCP tool schemas registered via mcp.servers are not injected into sessions spawned by sessionsspawn, regardless of how the documented exposure mechanism is configured. The documented…
Copilot CLI displays organization-policy warnings at startup even though every configured third-party MCP server is explicitly disabled through disabledMcpServers in the user settings. The same pair of warnings appears twice during…
Asked to verify my pro account to continue, clicked link, verified, got taken to page that states: Authentication successful The authentication was successful, and the following products are now authorized to access your account: Gemini…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again macOS 26.3.1 (25D771280a) After logging in with Google (SocialGoogle), kiro-cli whoami should include a resolved Profile section…
A remote MCP server that speaks streamable HTTP (or SSE) is replaced under a live Cline session — a restart, a deploy/rollout, a session-TTL sweep. From that moment on, every tool call to that server fails, for the rest of the session…
The only workaround is restarting Cline (in the CLI/TUI, restarting the hub daemon, which is what actually owns the connection).
Omarchy 4.0.4-1; Hyprland 0.56.2-2 (Wayland) Kernel: 7.2.5-3-omarchy CPU: AMD Ryzen 9 7950X3D; motherboard: ASUS TUF GAMING X670E-PLUS WIFI GPU driving both monitors: NVIDIA GeForce RTX 5080 (MSI), nvidia-open-dkms / nvidia-utils…
macOS 15.7.9 (24G830) Kiro CLI version: 2.28.0 OS: macOS 15.7.9 (24G830) Server: Google Workspace first-party MCP (.mcp.googleapis.com, e.g. people.googleapis.com/mcp/v1), configured as a Desktop-app OAuth client with clientId +…
JetBrains Plugin 4.1.23 (JetBrains plugin core 2.0.7, extensionVariant: next) I am using a beta version of Cline cline-core running inside Android Studio grows until it hits the 8 GB heap limit and dies, killing the session mid-task. The…
its working like about 10 min and its stuck then i was cancel the converstation and type continue and its stuck on working.... and i try this more then 10 time still stuck its working like about 10 min and its stuck then i was cancel the…
macOS 26.7 (arm64, Apple Silicon) kiro-cli 2.26.0 — KAS 2.26.0, bundled node v22.22.2 Summary. A remote HTTP MCP server is configured with an explicit oauthScopes list that includes an optional provider scope. Kiro completes the OAuth…
After updating Claude Code to the latest version in order to use Opus 5.5, Auto Mode started blocking legitimate, user-directed actions in an existing multi-session repository workflow. This workflow had been used repeatedly before the…
Disable Auto Mode for the affected sessions. - Use `acceptEdits`. - Add explicit permission allow rules for the required commands.
Installed python from Microsoft Store using Python Install Manager. Installed aider-install according to instructions: python -m aider-install But then aider-install is not found. Pip says that it is installed: Package Version…
macOS 26.6.2 (25G83), arm64 kiro-cli 2.32.0 (macOS app Kiro CLI.app, dashboard bundle dated 2026-10-15). Docker Desktop installed and running. The macOS app's Help & support → Automated checks keeps reporting Shell integrations ✘ with an…
Describe the bug 💡 Before filing, please check common issues: https://goose-docs.ai/docs/troubleshooting 📦 To help us debug faster, attach your diagnostics zip if possible. 👉 How to capture it…
Crash (process/app exits or hangs) On a macOS npm-global installation, updating Gateway 2026.10.5 to 2026.10.6 stalls in candidate Doctor before activation, with a waiting Doctor process and a waiting SQLite read-only worker; interrupting…
This report was explicitly reviewed and confirmed in OpenClaw. OpenClaw version: 2026.9.6 Node version: 24.18.0 Update action: Gateway RPC: update.run (control-ui) Installation method: npm-global Update target: 2026.10.2 Failed phase…
Since 1.0.102, interactive autopilot stops whenever a background subagent runs for more than 600 seconds. The subagent is still working normally when this happens. The CLI shows: A common pattern is for the main agent to start a background…
In 1.0.102 the native runtime adds new autopilot continuation logic, and the stop message and `autopilot_background_timeout` warning first appear there. - Workaround: `export COPILOT_TASK_WAIT_TIMEOUT_SECONDS=3600` (or higher) before launching. - The limit can't be turned off.
Copilot CLI v0.0.421 hangs indefinitely when launched from directories with Nix flake-based development environments managed by direnv. The bash tool fails to execute any commands, timing out with Invalid shell ID errors. Analysis of the…
Hermes Agent: v0.21.3 Server: macOS (Mac mini) Surfaces: Desktop app + mobile dashboard When a session is open on the Desktop app (surface=desktop), a second surface (mobile dashboard, surface=tui) cannot submit messages to that same…
In a Cowork conversation on Claude Desktop for Windows, tools work normally in text mode (including a local MCP server — Obsidian — that reads/writes my vault). As soon as I switch the same conversation to voice mode, the tools become…
tools.lazydeps.installspecs is now a shim that calls hermescli.oldupdater.stopforrelaunch(). That function starts the full historical update takeover (updatetakeover.py → updatefinish.py), which syncs PM and rebuilds Node deps, the TUI…
Setting `memory.provider: ''` stops the loop. After Desktop relaunched, agent init finished normally and no takeover started. Hindsight memory is unavailable while this setting is in place.
The Gemini CLI process stopped with a fatal error after approximately 9.3 hours of one interactive session. The workspace had Playwright reports, Allure results, and JSON files. The heap increased to more than 8.29 GB. The V8 garbage…
I purchased a Gemini Code Assist Standard subscription and successfully assigned a license to my account. Gemini CLI correctly recognizes my tier as "Gemini Code Assist Standard". However, after assigning the license, every request returns…
Describe the bug Goose Desktop's ACP chat path drops the MCP server's non-routing CallToolResult.meta before delivering ui/notifications/tool-result to an MCP App. The app receives content and structuredContent, but meta is reconstructed…
AMD Ryzen AI 5 340 w/ Radeon 840M (Framework), Omarchy 4.0.4-1, Quickshell 0.3.1-1, Hyprland 0.56.2, kernel 7.2.5-3-omarchy, Wayland. After cloning the menu plugin with the documented flow (omarchy plugin clone omarchy.menu), installed…
I've tried finding an answer on the Continue docs site I'm not able to find an open issue that reports the same bug I've seen the troubleshooting guide on the Continue Docs When using an OpenAI-compatible provider whose SSE stream includes…
Cline CLI 3.0.73 on macOS reuses an existing local Hub. A shell tool invoked by the model then sees the terminal context from whichever client started that Hub. Launch Cline in terminal A and leave its Hub running. In terminal B export a…
The nousresearch/hermes-agent:latest Docker image on Docker Hub does not include the dashboard (or web) subcommand, making it impossible to run the web UI in containerized environments like Coolify. The dashboard command should be…
On Windows, the first Microsoft Entra account broker (WAM) sign-in for a remote MCP server on the built-in Entra fast-path list (here https://sentinel.microsoft.com/mcp/data-exploration) hard-crashes the CLI process with an access…
What problem would this solve? PR #12869 (closes #11987) added a signed, per-user NSIS installer (Goose-win32-x64-setup.exe) for goose Desktop on Windows. From the next release it will be published alongside the existing ZIP, but the…
On Windows, the daily gateway log at %TEMP%\openclaw\openclaw- .log is truncated and rebuilt every time a CLI command or the gateway itself starts. After a gateway crash, the first thing that destroys the crash evidence is the operator's…
I see that on the Kiro IDE 1.3.0 - where Tool Search was recently added and is on by default - I still get the warning: You have N MCP tools enabled. We recommend disabling servers or tools to keep this below 50. At the same time, the IDE…
Describe the bug After upgrading from 1.54.0 to 1.55.0, goose doctor panics immediately in a background sqlx worker thread: Exit code: 1. This looks like a regression of #8719 (fix: avoid sqlite session creation panic on legacy dbs, closed…
Roll back to 1.54.0, which works against the same database:
Omarchy 4.0.5-1, linux-omarchy 7.2.8-9 (regression from 7.2.5-3) Dell XPS 13 9340, BIOS 1.29.0, Meteor Lake, IPU6-v4, Intel IVSC Webcam: OV02C10 (OVTI02C1, \SB.PC00.LNK1) via IVSC → IPU6 CSI2 4 IR camera: Himax HM1092 (HIMX1092…
Hi i tried to update the hermes client on MacOS but it failed. Error message is only Update failed so no info there :). Hermes Agent v0.21.5+8076.g85db7c3 (2026.9.24) · upstream 85db7c3a Install directory…
After updating to the latest main (24fd22b94d, v0.21.3 / 2026.9.14), the interactive CLI crashes immediately on startup with: This is a regression from 182ec5c28d (#111408, "widen the remaining config.yaml stat caches to filesignature")…
Behavior bug (incorrect output/state without crash) On an npm-global 2026.9.7 Gateway running inside an unprivileged Proxmox LXC container, openclaw update to the next candidate (2026.9.8) fails deterministically in phase validating at…
On 2026.9.6 the Gateway's main-isolate heap grows steadily at rest, about 45–100 MiB/h, until it reaches Node's default V8 ceiling (~4.2 GiB on this host) and aborts with Ineffective mark-compacts near heap limit. It happened once…
`--max-old-space-size=10240` as a direct `node` argument in the service's `ExecStart` (not `NODE_OPTIONS`). That buys days; it does not stop the growth.
11th Gen Intel(R) Core(TM) i5-1145G7 (8) @ 4.40 GHz, Intel Iris Xe Graphics @ 1.30 GHz [Integrated] Whenever a mise python version is installed, packages that depend on system python, like anki, will break builds and won't install: build…
With @google/gemini-cli 0.64.0 in headless mode (-p, --output-format stream-json), API-key auth and gemini-4-argon at thinkingLevel: HIGH, a turn in which the model writes one large tool-call argument (for example a whole document passed…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again https://kiro.dev/docs/cli/acp/ acp integration used to work as per above instructions until update this week get an error in intellij…
Behavior bug (incorrect output/state without crash) When running the Gateway inside a Docker container and using Docker-outside-of-Docker for Sandboxing, /workspace can't binds properly. The Gateway containder attempts to mount the agent…
`"sandbox": { ; "mode": "all",`
Windows absolute paths are classified as remote package slugs by decodePackageIdentifier. On main (5522c6f), C:\project\rule.yaml, C:/project/rule.yaml, and \\server\share\rule.yaml all return uriType: "slug". The forward-slash drive path…
I am using the free tier of OpenCode. I received the "Free usage exceeded, subscribe to Go [retrying in XXXXXs]" message. The problem is that the limit does NOT reset even after waiting for more than 24 hours and doing nothing in the app…
Architecture: x8664 - CPU: Intel Pentium N3710 @ 1.60GHz - GPU: Intel Atom/Celeron/Pentium Processor x5-E8000/J3xxx/N3xxx Integrated Graphics Controller (rev 35) - Bootloader: Limine / UKI - Root filesystem: encrypted, Btrfs After updating…
Kiro Crew v0.7.1 After a KiroCrew auto-update installed today, every Crew MCP tool call is refused with identityunattested. The gateway’s session-claim handshake reports connections=0 on every claim-push — for dashboard sessions, scheduled…
I subscribed to OpenCode Go today. My Go subscription is active, but the console does not show any personal API key or any option to create/copy one. /connect → OpenCode Go requires an API key. The Keys page only allows creating Service…
OpenCode returns an OpenAI 401 when sending a prompt. The error is: `Incorrect API key provided: sk-svcac OpenCode’s bot is asking you to edit the issue body (not just comment). The numbered fields were left as No response. You have about…
When an API call fails in headless mode, the last line I get is: The actual error (in my case a Vertex ModelNotFoundError: Publisher model ... was not found) doesn't make it to that line. This looks like the same thing as #13553, which was…
`export GOOGLE_GENAI_USE_VERTEXAI=true GOOGLE_CLOUD_PROJECT= GOOGLE_CLOUD_LOCATION=us-central1` 2.
OpenClaw: 2026.9.9 (bcfc888) → target 2026.10.3 Install: npm global package (installKind: package, npm 11.14.1) Node: v24.21.0 Gateway: systemd user service, channel stable openclaw update --tag 2026.10.3 --yes --timeout 600 installs the…
Microsoft Edge segfaults when hovering over images on certain sites (e.g. omarchythemes.com theme detail pages). The crash is caused by the interaction between Edge's native Wayland popup/overlay handling and the tile on window rule in…
Create `~/.config/microsoft-edge-stable-flags.conf`:
I'm running Gemini CLI locally on arch linux with the podman sandbox enabled: GEMINISANDBOX=podman When i run npm start, the CLI gets to the point where it rebuilds one of the native dependencies inside the sandbox, but the rebuild fails…
This is closely related to `https://github.com/google-gemini/gemini-cli/issues/15376` where `--userns=keep-id` was suggested as a workaround. the workaround does fixes the underlying UID mapping problem, but requiring users to manually add a podman specific flag feels like something the CLI could…
On Omarchy 4.0.0-1, SteamVR's vrmonitor aborts during Qt GUI initialization because the session-wide environment sets: SteamVR's bundled Qt 5 installation provides only the xcb platform plugin. It treats wayland;xcb as one literal plugin…
What happens Add a remote OAuth MCP server and sign in. ~/.aws/sso/cache/ .token.json and .registration.json are written. Keep the session running past the access token's lifetime (5 minutes in our case). Kiro refreshes successfully in the…
I have an active OpenCode Go subscription, and Go models work normally in OpenCode CLI. However, I cannot find any way to create, view, or copy a Go API key in the new v2 Console. The current Go documentation still says users can copy an…
Running copilot interactively on Windows causes repeated OAuth browser sessions during initialization. Browser sessions keep opening until the user terminates Copilot. This prevents normal startup and authentication. Existing CLI logs…
When running kiro-cli acp --agent-engine v3 --auth-method cli, the --auth-method cli flag is ignored. The agent hardcodes acp-callback authentication regardless of the flag, causing an infinite retry loop when the ACP host doesn't…
Use --agent-engine v2 instead, which respects CLI credentials. However, this loses access to MCP tool integration which requires v3.
Apple MacBookAir5,2, Intel Core i5-3427U, Intel HD Graphics 4000, Omarchy 4.0.4-1, Hyprland 0.56.2, Linux 7.2.5-3-omarchy Upon opening the laptop lid and resuming from suspend (S3), the Omarchy lock screen briefly displayed Hyprland's…
After upgrading to Qt 6.12.0, Super + Space opens the Omarchy menu as a fully black screen. The shell's Color singleton sub-objects (Color.menu, Color.tooltip, Color.bar, ...) evaluate to undefined, so every themed surface (menu, tooltips…
Intel i7-12800H, Intel Iris Xe + NVIDIA RTX A1000 Laptop GPU, Omarchy 4.0.4-1, Quickshell 0.3.1-1, Hyprland 0.56.2 If idle.screensaver or idle.lock in ~/.config/omarchy/shell.json differs from the built-in defaults so that the first idle…
`quickshell -p /usr/share/omarchy/shell ipc call idle disable && quickshell -p /usr/share/omarchy/shell ipc call idle enable` after the shell has loaded its config (and after each `shell.json` idle change).
I'm on the latest version as of 9/26/2026. I constantly deal with an issue where the taskupdate tools are not functioning, which causes my sessions to not be able to update the status of tasks in the tasks.md file of my specs. Task tools…
Since #127201 (d428df2a65, "fix(desktop): restore YouTube embeds in the packaged app"), the packaged Desktop renderer is served from http://127.0.0.1:47891. Its WebSocket upgrades therefore carry Origin: http://127.0.0.1:47891. When…
When starting the CLI with --agent , the custom agent is selected and then silently deselected ~1 s later, before any user input. The banner shows Selected custom agent: my-agent, then MCP Servers reloaded: 1 server connected, and the…
Frequency (automated pty launches, no input, 15 s each): dropped 5/6 with a user-level agent (`--agent my-agent`), 4/6 with a plugin agent (`--agent myplugin:my-agent`). - When MCP startup failed (`Failed to start MCP Servers: session not found`), the agent was never dropped, which points to the…
When using a BYOK provider with a short-lived bearer credential (e.g. an Entra ID / Azure AD OAuth access token, an AWS STS token, or any OIDC JWT), the Copilot CLI reads COPILOTPROVIDERAPIKEY (and any custom headers) once at process…
export COPILOT_PROVIDER_TYPE=azure export COPILOT_PROVIDER_BASE_URL=https://my-apim.azure-api.us/openai export COPILOT_MODEL=gpt-5.1 ### CLI runs this to get a token, and re-runs it automatically before expiry / on 401: export COPILOT_PROVIDER_API_KEY_COMMAND='az account get-access-token --scope…
Thinkpad T480, Omarchy 3.3.0 I upgraded to 3.3.0 today, and after I enter the login credentials, it waits for 2 minutes showing the login window, then reboots. I pressed Ctrl-Alt-7 to see the terminal while it was waiting, here is the…
Arch Linux - kernel version: 7.2.4-arch1-2 After upgrading kiro-ide from 1.0.437 to 1.1.14 on Arch Linux, sign-in is broken. Clicking "Sign in" shows "We will take you to your default web browser to sign in and bring you back when it's…
Downgrading to 1.0.437 resolves the issue.
When the TUI reconnects after a brief network interruption, it can replace the composer with the "Session location unavailable — Choose another directory to continue this session" recovery prompt even though the session's directory is…
I am using a beta version of Cline During a single prompt session using the Cline CLI with a reasoning model (z-ai/glm-5.3-flash), the Cline daemon experienced an extreme memory allocation explosion and disk write churn that completely…
With the recent deprecation of Gemini CLI OAuth for individual, Google AI Pro, and Ultra accounts ("Google no longer supports Gemini CLI OAuth for individual, AI Pro, or Ultra accounts. Enable CodexBar's Antigravity..."), individual paying…
OpenCode 2.0.8 removed --command from opencode run, and run does not expand slash commands in the message, so there is no CLI way to invoke a command headlessly. --command was the supported headless path in v1 — see #7345, where the…
Ubuntu 24.04 (WSL) and MacOS 15.7.3 Error 'Mcp error: -32002: Client error: HTTP status client error (405 Method Not Allowed) for url (https://my-gw-rrin.gateway.bedrock-agentcore.us-west-1.amazonaws.com/mcp) Before running kiro-cli, env…
Behavior bug (incorrect output/state without crash) openclaw secrets audit reports PLAINTEXTFOUND for a provider API key held in a SQLite auth profile, but that plaintext is written by the official openclaw configure --section model flow…
AMD Ryzen 7 6800U, Radeon 680M (integrated), Omarchy 4.0.0.r6807.g902fd8a-1 (edge channel; now rolled back to r6720.g8e02fc8-1) Since around 2026-10-08 09:04 UTC, the edge repo database at https://pkgs.omarchy.org/edge/x8664/omarchy.db…
Source: [[P3] [Triggered] container-runtime/copilot-runtime-bazel-cache-memory-usage-vs-limits-dotcom on…
I've tried finding an answer on the Continue docs site I'm not able to find an open issue that reports the same bug I've seen the troubleshooting guide on the Continue Docs When using the Continue extension in the primary/secondary side…
What happened? The Gemini CLI is consistently throwing a Tool execution denied by policy error. This is a generalized blockade affecting all tool invocations, but it severely impacts fundamental I/O operations (file reading, file writing…
macOS 26.7 (arm64) A remote MCP server that authenticates via OAuth connects reliably in interactive kiro-cli chat, but always fails with Unauthorized in kiro-cli chat --no-interactive — even when a valid, unexpired access token (plus a…
Omarchy dev channel (dev-linked checkout, integration build db439093), Hyprland 0.56.2, AMD Ryzen AI 9 HX 370 w/ Radeon 890M (GPD G1628-04). Two displays configured: internal eDP-1 panel plus a Vestel TV on DP-1. Session uses a plain…
https://geminicli.com/extensions/ The made-in-show extension is not listed in the Extensions Gallery, nor present in https://geminicli.com/extensions.json, although the repository meets every documented publishing requirement and the…
Cloud Platform N/A - ClinePass web billing issue I am using a beta version of Cline I am reporting a subscription billing issue and requesting a refund for an accidental annual upgrade. On October 21, 2026, I accidentally selected the…
I've tried using the "Ask AI" feature on the Continue docs site to see if the docs have an answer I'm not able to find a related conversation on GitHub discussions that reports the same bug I'm not able to find an open issue that reports…
Title: Remote Control sessions not accessible from claude.ai/code without session-specific URL The claude.ai/code landing page and the Code button in claude.ai both redirect to GitHub onboarding with no way to skip. This makes it…
Since v0.60.0, running Gemini CLI with sandboxing asks for authentication every time. The OAuth tokens are not saved persistently. What is worse, all sessions are lost: it is not possible to resume previous sessions, and new sessions are…
An ACP client starts a session and supplies its own MCP server in session/new's mcpServers (ours serves a shared task board to each agent). Every first call of every tool on that server then stops on a session/requestpermission, once per…
AMD 5900x, RX9070XT, Omarchy 4.0.0 Omarchy: 4.0.0 ISO: omarchy-4.0.0.iso, SHA256 verified against the published checksum Boot media: SanDisk 64 GB USB written with balenaEtcher, Secure Boot disabled Machine: two NVMe drives — Samsung SSD…
Source: [[P3] [Triggered] container-runtime/copilot-runtime-bazel-cache-memory-usage-vs-limits-dotcom on…
Source: [[P3] [Triggered] container-runtime/copilot-runtime-bazel-cache-memory-usage-vs-limits-dotcom on…
VSCode Extension I am using a beta version of Cline I ran into a weird routing issue while using Cline with deepseek/deepseek-v4.1-flash. The request started failing with a 429: What caught my attention was the routing metadata. The…
This report was explicitly reviewed and confirmed in OpenClaw. OpenClaw version: 2026.10.3 Node version: 24.21.0 Update action: CLI command: openclaw update Update target: exact target unavailable; mode: pnpm Failed phase…
During an interactive video localization and YouTube packaging workflow, the agent (Gemini 3.8 Flash) exhibited complete breakdown of agentic initiative, active instruction non-compliance, and degenerative conversational behavior. Passive…
Omarchy 3.8.2 (updated 2026-05-27) Hyprland 0.55.2 (was 0.54.3) hyprlock 0.9.5-4 (was 0.9.5-2) SDDM 0.21.0-7 (was 0.21.0-6) SDDM greeter rejects correct password after 3.8.x update — greeter falls back to us keymap because installer only…
Added the following to `/usr/share/sddm/hyprland.conf`:
When resuming an existing conversation in the Claude Code desktop app, the session gets stuck with: "Can't reach your computer — It may be asleep or offline. This session will reconnect when it's back." The message never clears while…
Recently assisted permissions mode has been requiring my approval for too many commands. This is mostly just a feeling that something has changed so sorry for being vague, but for example this powershell command to find files in the…
CLI (cline from npm, 3.0.71), headless: cline --json -P cline-pass -m cline-pass/deepseek-v4.1-flash -t 7200 --retries 5 " " We run the Cline CLI unattended to build apps through an MCP server (long tasks of 20 to 200 requests, mostly…
I pay for Cline Pass. sub1ULOU3JvJ1E14BGMWjDRrINh. You shipped a broken product, and now you've stopped answering. Somewhere between 2026-10-02 19:17Z and 2026-10-03 05:35Z, deepseek-v4.1-flash was moved off the DeepSeek channel: Same…
<h2 id="summary" style="color: rgb(204, 204, 204); font-family: -apple-system,BlinkMacSystemFont,Segoe UI,Roboto,Helvetica Neue,Arial,sans-serif; font-size: 28.8px; font-style: normal; font-variant: normal; font-weight: 400…
In non-interactive Copilot CLI, an already-discovered Azure MCP tool call can hang after delivering its initial progress notification. No tool result is delivered, and the MCP WebSocket closes five minutes later with idle timeout. The…
Tools with no input parameters use an empty JSON Schema for their Input Schema. This causes Copilot CLI to reject the tool entirely and fail on any prompt when the MCP server is connected. 0.0.365 Commit: 76d0881 To have an MCP tool with…
I am using a beta version of Cline I updated the latest version 0.0.42 of cline from 0.0.37, but I can't open my app because it always Tauri invoke failed for get desktopbackendendpoint: desktop backend endpoint not ready and I can't open…
Behavior bug (incorrect output/state without crash) The stock signed macOS app cannot inspect a managed Bun Gateway command containing --no-install because installedServiceCLI only skips memory flags; the native update/recovery screen…
Issue #411 was closed as "notplanned" in Oct 2025, but the bug is still present and getting worse — it now blocks the CLI from running any shell command at all on Windows machines that ship only with Windows PowerShell 5.1…
Describe the bug A GDK state machine using InferenceRunner can keep calling the provider indefinitely when the provider returns empty responses. The runner makes an initial attempt and three retries, then stores a hidden fallback message…
What problem would this solve? In many enterprise and team environments, access to LLM providers—whether through unified gateways like LiteLLM Proxy, corporate API gateways (Kong, Envoy, Apigee), or direct enterprise endpoints for OpenAI…
PENDINGDELIVERYNOTICE ("I couldn't confirm whether my previous reply reached this chat...") is being sent in place of the real agent reply, repeatedly, across many separate turns over 24+ hours — not just once after a restart as the…
I use the TUI over SSH and can't paste images. Unsure if it's a bug or a gap in the design: the clipboard I want to read is on my laptop, but the TUI is running on the remote box. Add a read path to the terminal clipboard backend using…
Since upgrading Ubuntu from 24.10 to 25.04 (and thus, from Python 3.12 to 3.13.3) , pipx install aider-chat fails with: /home/user/.local/state/pipx/log/cmd2025-07-1213.40.46piperrors.log contains this: I tried importing…
As a workaround, you can try installing `aider-chat` using an earlier Python version managed by `pyenv`:
After upgrading, the gateway crash-looped (5+ boots). The SQLite store worker hit its V8 heap cap during startup maintenance and took the gateway down each time. Observed on a long-running install; not yet reproduced on a clean install…
agy -p "Hi" hangs indefinitely (5+ minutes, no response, no error, no timeout message) on all inference calls in print mode. The CLI starts, accepts the prompt, then never returns. This happens on the latest version (1.2.2) and was…
Split out of #4378, which carries two faults. The registry policy 401 in that issue has a cause and a workaround. This one does not, and it survives that workaround. In a Copilot cloud agent session on a GHEC data residency tenant (…
macOS 26.6.2 Kiro CLI 2.21.4 (first noticed after updating today) KiroCrew v0.6.0 (also updated today) The shell tool available to agents in the KiroCrew dashboard does not inherit SSHAUTHSOCK from the parent process environment. This…
Amazon Linux 2023.12.20260930 (x8664), kernel 6.12.110-135.202.amzn2023.x8664. Accessed over SSH, running inside tmux (TERM=screen-256color). kiro-cli 2.28.0 The V3 TUI is not usable for me in a plain remote terminal. I run kiro-cli over…
Describe the bug I'm testing the pre-registered OAuth client support for streamablehttp extensions introduced in Goose v1.47.0. PR #11182: https://github.com/aaif-goose/goose/pull/11182 Commit: ea6dbf30b8e53ac5a623320eab1b77f61b604166 My…
After upgrading to qt6-declarative 6.12.0-1 (and quickshell 0.3.2-1), opening the Omarchy menu (omarchy-menu toggle via SUPER + SPACE or SUPER + ESCAPE) causes a pitch-black screen with locked focus, and volume/brightness key adjustments…
Downgrade `qt6-` to `6.11.2` and `quickshell` to `0.3.1-1`.
Starting with 1.0.93, Copilot CLI on macOS can't connect to its API when it runs inside a sandbox that denies the Mach service com.apple.SecurityServer. Sandboxes such as nono deny this service so the agent can't read keychain secrets…
MCP tool discovery ignores nextCursor from a tools/list response. When a server advertises tools across multiple pages, Gemini CLI discovers only the first page, with no diagnostic that additional tools were omitted. On main at…
Describe the bug Since v1.39.0, goose no longer attaches an image when a message references it by a Windows path such as C:\Users\me\Pictures\shot.png (or C:/Users/me/Pictures/shot.png). The request sent to the provider only contains the…
just looking up running extensions, this one is on top, taking quite a bit of loading time and... really not running at all anywhere i am currently accessing Extension version: 0.20.0 VS Code version: Code 1.139.1…
Launching the CLI with --new-session-id and then running /restart fails every time: By the time this surfaces the foreground session has already been shut down, so the restart is not recoverable. copilot --new-session-id…
Updating an existing schedule with an invalid cron expression disables its timer even though the update returns an error. The schedule remains in the in-memory job list, but its previous cron expression has already been replaced and its…
What problem would this solve? Enterprise deployments of LiteLLM proxy often enforce OAuth 2.0 / OIDC JWT authentication (enablejwtauth: true) backed by an identity provider (Keycloak, Okta, Entra ID / Azure AD, Auth0, Dex) rather than…
VSCode Extension 4.1.24 (also reproduced on 4.1.23 and 4.0.12) On this machine, Cline's MCP servers have never connected at startup — this is 100% reproducible, not intermittent. Every window reload ends the same way: Zero MCP-related…
Kiro CLI 2.22.0 WSL2 (Ubuntu 24.04), Windows 11 (build 26200) wslu/wslview installed; wslview opens the Windows default browser correctly Identity Center (org / --license pro) login When running kiro-cli login inside WSL, the CLI detects a…
After the lock screen blanks the display, waking it does not restore focus to the password field. Keystrokes go nowhere. The only way to recover is to click with a mouse or tap the touchpad, which defeats keyboard-only operation. Omarchy…
https://geminicli.com/extensions/ Our extension https://github.com/olostep/olostep-mcp-server (olostep) is not listed in the Extensions Gallery, although it meets all documented requirements: Public repository gemini-cli-extension topic…
When Copilot CLI runs without a usable GitHub credential, the background auth probe copilotruntime::auth::gh (i.e. gh auth token) is retried in a tight loop with no backoff. I hit this in a normal interactive session using a BYOK provider…
Signing in (`/login`) or exporting `COPILOT_GITHUB_TOKEN` / a fine-grained PAT stops the loop; installing `gh` only lowers the severity.
What happens @continuedev/cli 1.5.47 is the current release on npm. When isAuthError classifies an error as authentication, an interactive session falls back to npx -y mcp-remote in MCPService.ts. The predicate covers HTTP 401 and 405 and…
On 2026.9.8, a ChatGPT Responses WebSocket that closes with code 1000 after dispatch ends the run with "Agent run failed". Neither the same-model retry nor the settled-batch continuation from #130721 runs, and the session idles until the…
Cloud Platform N/A - ClinePass web billing issue I am using a beta version of Cline Cloud Platform — ClinePass web billing (app.cline.bot) I am reporting a ClinePass billing issue and requesting a refund of a USD 77.22 credit balance that…
Every hour or so I'm getting Authorization error. Your credentials may be expired or invalid. (Request ID: ) Run /login to sign in again. Running /login does finish successfully but does not remedy the issue. Nor does mcp reload. GitHub…
After using OpenCode for a few hours, any model I try gets stuck indefinitely on "Thinking". There is no error message and the request never finishes. Switching to another model does not help. I have not reached any usage or rate limits…
Since 1.0.89, every new interactive session shows this error twice right after startup: Failed to read model provider attribution: Error: Not authenticated Sign-in completes about 3 seconds later and chat/model requests work normally. It…
0.0.39 (previously 0.0.32 → 0.0.37) I am using a beta version of Cline Connecting to a remote environment fails with "Unable to connect to SSH host". The remote helper daemon (--cline-hub-daemon) is leaked and never reaped after failed…
When an MCP server passed in ACP session/new fails to start, the ACP client is never told. The session/new response is a normal success, no session/update is sent, and nothing appears on stdout or stderr. The host has no way to learn that…
On Omarchy 4.0.3-1 with kernel 7.2.3-arch1-3, this laptop intermittently hard-freezes ~20-50 s after a successful s2idle resume. Suspend and resume both complete and log normally (PM: suspend entry and PM: suspend exit); the machine then…
Pin/downgrade to kernel 7.1.9-arch1-2 (or use `linux-lts`) until a fixed kernel is available. - Try `i915.enable_psr=0` / `i915.enable_dc=0`, since the failure is in the post-resume display re-initialisation path. - Try `mem_sleep_default=deep` to move off s2idle.
Laptop: Lenovo ThinkPad T460, CPU: Intel(R) Core(TM) i7-6600U (4)z, Graphics: Intel HD Graphics 520 / Skylake-U GT2, GPU driver: i915, Omarchy: 4.0.4-1 When I shut down the laptop: The screen turns off. The power button light turns off…
The context usage indicator in the chat panel always displays 0%, regardless of how many files have been read or how long the conversation is. When the context window is actually full (after reading many files or having a long…
Product: Kiro IDE Version: 1.0.0 (updated morning of June 17, 2026) OS: Windows Trigger: Occurred immediately after updating Kiro to version 1.0.0 and enabling the Agent Focus experimental feature. Issue: Previous session chat history is…
What problem would this solve? Goose users can't easily see how AI agents behave on everyday business jobs with planted traps, test their own agent against those traps, or check a skill, MCP server or package before trusting it. The Agent…
Crash happened using model gemini-3-flash-preview trying to do the ingest phase of a LLM wiki. I uploaded 2 80-slides pdfs and one short md and asked LLM for ingest. The LLM thought for 18-min reaching the end of the free-tokens , switched…
What problem would this solve? This will prevent context of the model from being set incorrectly and allow user to correct the model context size. What would a good outcome look like? when clicking on the model context token, user can…
GitHub Copilot OAuth login succeeds and the account is stored, but no Copilot models ever appear in /models or opencode models. The model-list sync is never attempted after login: the server log contains no sync success or failure for the…
The ripgrep binary bundled with the Copilot CLI ( /ripgrep/bin/linux-arm64/rg) statically links jemalloc built assuming a 4KB page size. On a kernel configured with a 16KB page size — the default on Asahi Linux (Linux-on-Apple-Silicon) —…
Replacing the bundled `rg` at each cached version's `ripgrep/bin/linux-arm64/rg` with the system's dynamically-linked `rg` resolves the crash locally, but is undone on every CLI update since it re-fetches the broken bundle.
I'm hitting random failures with deepseek-v4.1-flash on the opencode-go provider. The session just stops with: Bad Request: {"model":"deepseek-v4.1-flash"} Sometimes I also see it come back as a 403 with error code: 1010 when I replay the…
Updated after the discussion below. The blank-to-link-drop mechanism v1b3m describes is confirmed on this machine too and belongs with #9078 and #10691, so I have split it out of this report. What remains here are the two findings that the…
On start it use opencode --port 16955 Unrecognized flag: --port in command opencode but correct syntax is opencode serve --port 16955 opencode v2.0.10 Install opencode & vscode plugin Use button with black square on white "Open opencode in…
i5-10400F, RTX 2080 Ti, Omarchy 4.0.0.r1787.g9455496-1 quickshell: symbol lookup error: quickshell: undefined symbol: ZN23QUntypedPropertyBindingC1EP23QPropertyBindingPrivate, version Qt6PRIVATEAPI 1) Tried reboot 2) Re-tried Omarchy…
I am using a beta version of Cline I'm having problems getting agents to see installed MCPs in Cline Desktop on Windows. Whether they are manually installed or installed via the marketplace, when I ask the model to list the tools available…
Cline Desktop 0.0.38 (core 0.0.88) cannot connect to a Linux aarch64 SSH remote. The bundled cline-remote-helper-aarch64-unknown-linux-gnu (sha256 93444d160eed1dc9…) exits 1 on --remote-hub-ensure with a Bun SyntaxError: Invalid character…
`mkdir -p ~/.cline/remote-helpers ; scp :~/.cline/remote/cline-remote-helper-linux-arm64-8b51183e615e7442 \`
Describe the bug MCP servers that authenticate via OAuth demand a fresh interactive browser authorization on (nearly) every new goose session, even though valid credentials were previously stored. Once it starts happening, the stored…
https://geminicli.com/extensions/ The extension in https://github.com/ANTEVO-CH/antevo-mcp (manifest name antevo) is missing from the Extensions Gallery. It is not in https://geminicli.com/extensions.json (last-modified 16 Oct 2026 02:36…
Copilot CLI 1.0.98 (installed via official gh.io/copilot-install script, Linux x64, headless — no system keychain, token stored in plaintext ~/.copilot/config.json) Auth via copilot login --device-code on github.com (not a GHEC…
A setting (or environment variable) for the API request headers timeout, honoured with every auth type, including Vertex AI. For example model.requestTimeoutMs in settings.json, or GEMINIAPIHEADERSTIMEOUTMS. Since #26191 the global undici…
Our workaround is to preload a module (`NODE_OPTIONS=--import`) that re-creates the global `Agent` with a longer `headersTimeout` when the CLI installs it.
I am using a beta version of Cline Cline CLI starts up normally — the ASCII logo, "What can I do for you?", the input box, model line (MiMo-V2.6-Flash (free) (medium)), git status (ai agent (master) | 51 files +1375) and "Auto-approve all…
Debian 13 (trixie), x8664. The mechanism is OS-independent: the status is a string literal in the agent runtime's own JavaScript, written before the command result is looked at. kiro-cli 2.21.4, @kiro/agent 0.63.3, engine v3 (KAS). Not a…
I am using a beta version of Cline After updating to version 0.0.34 Beta, the application stopped working properly. Before the update everything was fine, but now the software does not start or respond as expected. I did not change…
Can't use MCP anymore. Tool Permission executebash trust read-only commands fsread trusted fswrite not trusted introspect trusted reportissue trusted useaws trust read-only commands Servers still loading version = "1.15.0" hash =…
I am using a beta version of Cline I'm using a large-context thinking LLM, which outputs a lot of code-blocks and patches. When starting a new session, cline is snappy / responsive. But as the context / history gets longer, the TUI begins…
I am using a beta version of Cline The Cline CLI (v3.0.65, also reproduced on 3.0.38) hangs indefinitely at startup when run in non-interactive/headless mode with the -c flag. The interactive TUI mode (cline without flags) works fine on…
Gemini CLI 0.62.0 on macOS retains one distinct /dev/ptmx master per completed ordinary shell command in a minimal token-free reproduction. This looks like a recurrence of #15945 after #25079. It exhausts the host PTY limit and prevents…
VSCode Extension I am using a beta version of Cline When configuring the OpenCode provider in Cline Desktop and selecting any model from the dropdown (such as LongCat 2.5 Preview Free): Incorrect Model Execution: Even though…
On a Linux box without xdg-utils (SSH, containers), opencode auth login prints the authorization URL and then exits with Error: Executable not found in $PATH: "xdg-open" and a stack trace, so the login can't be finished. openUrl in…
I am using a beta version of Cline In the CLI's hub mode, a session's stdio MCP server processes are never stopped when you leave that session, whether by /clear or by quitting the TUI. Each new session starts its own set of MCP servers…
I am using a beta version of Cline Leaving the Cline Desktop App running overnight causes the WebSocket connection between the sidecar and the local hub daemon to drop immediately on the next launch. Every prompt then fails with a hard…
What problem would this solve? The backend currently treats the extensions selected for a session and the MCP clients running for them as one mutable ExtensionManager. The same component resolves configuration and secrets, starts…
VSCode Extension I am using a beta version of Cline Happens while displaying the stream of thinking tokens. Display slows down, then stops (but I know the model is actually producing tokens cause it is internally hosted), then bottom…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again macOS 26.7.1 (25G313) Kiro points to the updated Version of note Kiro points to the unsafe Version of note
0.0.45 (extensionversion 0.0.45, clinetype: desktop, platformversion 0.0.45) A long-running agent session (hundreds of tool calls; tool results carry multi-KB JSON including full Python source; no images at all) accumulated a 23.1 MB…
I am using a beta version of Cline i'm install github mcp via marketplace and it cant authorize on github why it not able to authorize in github like cloudflare-api mcp ? install github mcp via cline marketplace try to use github mcp via…
I am using a beta version of Cline macOS (Cline Desktop app, not the VS Code extension) Cline version: Mac: 16 GB RAM, up 5 days, load average ~4.5 MCP servers enabled: 5 (aiemaily, narrateai, zapier, browseros, buildin) The app closes…
`pkill -f "cline-hub-daemon"; sleep 1; open -a Cline` before each session.
OpenCode Version: 1.18.29 Model / Provider: Mistral (mistral/zai-glm-5-2) When using Mistral models (specifically zai-glm-5-2) with tool calling enabled, streaming responses crash with an AITypeValidationError. The error trace indicates…
What problem would this solve? MCP App guest content is served from a stable loopback web origin. Without host-enforced origin isolation, separately supplied apps can share browser origin state and may interact with host cookies or storage…
Stop letting @thdxr push :) Running the built-in /btw slash command crashes the TUI after the generated answer is returned. The answer dialog attempts to access PluginProvider outside its context boundary. opencode version: 2.0.8 Shell…
STATUS: FURIOUS. ABSOLUTELY LIVID. Kiro decided to shit the bed mid-session Extension host processes spawning like rabbits (PIDs 45736, 46296) Grammar scope overwrites flying everywhere (SQL, Java, YAML all fighting) Deprecation warnings…
Aider version: 0.86.2 Python version: 3.12.15 Python implementation: CPython Virtual environment: Yes Git version: git version 2.50.1 (Apple Git-155) An uncaught exception occurred:
CLI 2.24.1: KIROAPIKEY works with agent-engine v1/v2 but V3/KAS returns 403 invalid bearer token export KIROAPIKEY='ksk...' kiro-cli chat --no-interactive --agent-engine v1 "Reply with OK" kiro-cli chat --no-interactive --agent-engine v2…
export KIRO_API_KEY='ksk_...'
Describe the bug Cannot connect to the Datadog MCP server via Oauth Expected behavior When connecting to the Datadog MCP server ( https://mcp.datadoghq.com/api/unstable/mcp-server/mcp?toolsets=core with a custom query string parameter to…
Downgrading to `1.44.0` restores connectivity. https://github.com/aaif-goose/goose/releases#release-v1.44.0
I am using a beta version of Cline When user logout from the Cline Desktop app, and navigate to Providers > Configures section, the session still shows that the user is logged in and will not update to logout until user restart the app…
VSCode Extension 4.1.23 (saoudrizwan.claude-dev; official v4.1.23 release). After updating an existing Cline installation, the extension returned to onboarding. I configured Bring my own API key → OpenAI Compatible for a self-hosted…
VSCode Extension While running Cline workflow, Cline extension is restarting very often with VSCode + Remote SSH Connection to Linux VM. Also as per the log analysis of https://github.com/cline/cline/issues/8778 we tried to downgrade to VS…
This is about the Kiro CLI Windows installer script, not the IDE — routing note for triage. The documented install command is: Under iex the script body executes in the caller's scope. There is no script to exit, so each of the five exit 1…
What problem would this solve? MCP Apps can declare CSP destinations for network connections, resources, frames, and base URIs. Goose currently accepts loopback IPs, localhost, and other IP literals in those declarations while serving app…
Stored OAuth credentials for a streamable HTTP extension can be erased by Goose itself when multiple sessions start concurrently, producing an endless re-authentication loop: the user completes the vendor's browser auth, and within…
I've tried finding an answer on the Continue docs site I'm not able to find an open issue that reports the same bug I've seen the troubleshooting guide on the Continue Docs I tried to let my agent use tool editexistingfile but no matter…
Describe the bug The Goose desktop app shows a "Honk! An error occurred" screen with the message: This crashes on startup for users who have never saved a theme preference (i.e., getSetting('theme') returns undefined). The undefined value…
Desktop App + CLI/SDK (affects any test suite whose module graph reaches @cline/llms) A corrupt bun install produced a partially-extracted @sap/xssec@4.15.0 in nodemodules — src/token/ was missing entirely. Because…
I put together a small open dataset of real coding-agent tool-call failures, built from public GitHub issues: https://github.com/faddacake/agent-failure-evals (v0.1). 1 of the 25 cases comes from an issue in this repo. I'm sharing it in…
Describe the bug In agent sessions running under the goose Desktop app (server-managed backend), results of subagents launched via the delegate tool with async: true cannot be retrieved with load(source: " "). The call returns Error: Task…
0.0.38 (/Applications/Cline.app, bot.cline.app; hub daemon: code-sidecar --cline-hub-daemon --port 25463 --pathname /hub). Verified against tag desktop-v0.0.38; still present on main as of 2026-09-30. Adding a custom OpenAI-compatible…
continue lets users wire in MCP servers as tools, which raises a before-you-connect question: can we show how trustworthy a given MCP server is first? there's a public TrustScore for ~19.7K MCP servers/agents today. an optional check could…
Describe the bug 💡 Before filing, please check common issues: https://goose-docs.ai/docs/troubleshooting 📦 To help us debug faster, attach your diagnostics zip if possible. 👉 How to capture it…
### Cline Surface 3.0.66 (cline), @cline/core 0.0.88, @cline/cli-windows-x64 3.0.48 I am using a beta version of Cline The plugin-sandbox subprocess fails on every launch and all plugin commands are silently disabled. The failure is…
Creating `C:\Users\ \AppData\Local\Cline\node_modules\` containing `jiti` and `@cline/` makes both errors disappear (verified working).
Since 1.54.0, a subagent spawned while Code Mode (codeexecution) is enabled can never terminate. Its system prompt advertises recipefinaloutput and instructs it that calling that tool is mandatory, but under Code Mode the tool is…
Describe the bug Since v1.48.0, goose doesn't send notifications/cancelled when a tool call to a stdio MCP server on protocol 2026-07-28 is stopped (Stop in the desktop, or ACP session/cancel) or hits the extension timeout. goose reports…
Oauth login(Sign in with google) api key login I'd expect to have the model I've selected Client Information Run gemini to enter the interactive CLI, then run the /about command.
Remote SSH is completely non-functional due to a chain of three cascading bugs. The server binary cannot be downloaded, and even with manual installation, version mismatch causes silent fallback to the local filesystem while the status bar…
Kiro is down from morning? I have to followup more than 200 times. WTH can't you fallback to available model, if you have availability issues. If none of the model is available, why can't you scale your systems? Kiro is down from morning?…
A long-running Goose Desktop conversation disappeared from Goose history after upgrading from v1.51.0 to v1.52.0 and subsequently restarting the application. The recurring user-visible behavior was first observed after the v1.51.0 →…
Cline Desktop 0.0.43 (darwin/arm64) I am using a beta version of Cline The Linux x64 remote helper shipped in Cline Desktop 0.0.43 crashes immediately on every invocation. Remote environments are completely unusable. The desktop app…
VSCode Extension I am using a beta version of Cline Related: #13487 (same symptom family — hang after runcommands, resumed by a user message; CLI, no diagnosis), #13814 + PR #13848 (approval-state recovery, awaitingapproval phase — active…
JetBrains Plugin The tips text updates way too frequently, causing the screen to flash. This is annoying and distracting. The user expects either a slower update rate or an option to completely turn off tips. Open JetBrains Rider with the…
After upgrading from Goose 1.30.0 to 1.31.0, starting goose session panics with: The Goose session never starts — the process exits with an error. Schema migration v10 (applied on first run of 1.31.0) introduces a new threads table and…
After upgrading from 1.49.0 to 1.50.0, all streamablehttp MCP extensions that worked fine on 1.49.0 fail to connect on 1.50.0, with: and in the CLI log: Bisected to #11827 ("prefer latest MCP version"), which changed…
In packages/cli/src/gemini.test.tsx, the test 'should NOT load project hooks when workspace is not trusted' fails with FatalUntrustedWorkspaceError whenever tests run in a non-TTY or automated environment (e.g. piped test runners, CI…
Date/Time: September 11, 2026, ~04:15 UTC Error Code: 10737 (user-assigned) Task Being Executed: Discussing remaining Blade/Vue fixes after Agent removal What Was Completed Before Crash: Removed Jenssegers\Agent dependency from 4 files…
What problem would this solve? Desktop supports externally managed backends through settings and environment configuration, including local development services. Users and backend authors need a clear contract for how the client…
Please explain the motivation behind the feature request. Identity/credential brokers (e.g. Warden, and similar projects) sit between agents and external services. At session start, the agent calls the broker's discovery endpoint and gets…
I've tried using the "Ask AI" feature on the Continue docs site to see if the docs have an answer I'm not able to find a related conversation on GitHub discussions that reports the same bug I'm not able to find an open issue that reports…
Description: When attempting to update the goose Desktop application on Windows 11 via the in-app settings (Settings > App > Updates), clicking the Download button works successfully. However, when clicking the subsequent Install & Restart…
Since v0.61.0 (#29117), /mcp auth fails for any MCP server whose authorization server publishes RFC 8414 metadata with an issuer but does not return iss in the authorization response: The same server authenticated fine on v0.56.0. This…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again macOS 26.5.2 (25F84) Installer / updater image disks are unmounted after an update. I get a growing list of mounted disk images over…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again macOS 15.3.2 (24D81) look like the last release introduced a bug in login session not being persisted. i keep getting asked about to…
Totally not AI-generated OpenCode 1.18.31 on Android/Bionic crashes before the provider call when compiled with Bun 1.4.1 and splitting:true. Observed failure: TypeError: undefined is not an object (evaluating 'a.name') Affected endpoints…
`minify:true, splitting:false`
When agent tools execute commands asynchronously in background task mode (runcommand -> managetask), tasks frequently enter an indefinite "Forever Running" hang where task status remains permanently stuck on RUNNING. Even after the…
A plugin PreToolUse hook is discovered and loaded ("Loaded plugin hooks" appears in the CLI log), but never executes for any tool call made by the main agent in a headless goose run session. The tool runs as if no hook existed. This makes…
Commands.cmdtest supports a callable test command (testcmd), and Coder.create(..., testcmd=...) accepts one. But Coder.getplatforminfo() builds the system prompt with self.testcmd + "\n", so any coder whose testcmd is a callable raises…
InputOutput.validatecolorsettings() handles an invalid assistantoutputcolor by warning "Disabling this color." and setting the attribute to None. The assistant reply is then rendered with rich.markdown.Markdown(..., style=None), and Rich…
When a Gemini CLI session is finished and displays the termination card (To resume this session: gemini --resume ), the underlying Node.js process hangs and does not exit. This leaves zombie terminal panes in terminal multiplexers like…
Windows with WSL2 (Ubuntu) VS Code opened from a WSL directory using code . Continue VS Code extension installed under both Local and WSL: Ubuntu Continue version: 2.0.0 VS Code version: 1.140.0 When Continue’s agent runs a terminal…
packages/core/src/scheduler/scheduler.ts:927-934 recursively calls this.execute() when a tool returns sandboxexpansionrequired, with no counter or depth limit: A malicious or buggy tool returning sandboxexpansionrequired on every call…
Users look for Suno, Seedance, HappyHorse, GPT Image and Nano Banana by service name. Current Continue source removed Hub package/slug loading, so a local MCP file is not an official searchable listing. The existing official Example MCP…
[ACTION REQUIRED] 📎 PLEASE ATTACH THE EXPORTED CHAT HISTORY JSON FILE TO THIS ISSUE IF YOU FEEL COMFORTABLE SHARING IT Root causes: interactiveCli captures 0x03 via Ink keybindings with exitOnCtrlC: false and calls void…
When the agent invokes the built-in GoogleSearch tool, the CLI frequently gets stuck in an infinite Thinking... state. The tool execution never times out, never throws an error, and never yields control back to the prompt, keeping the…
I probed every remote endpoint advertised in the official MCP registry (10,513 of 10,542 reached). The dev.continue/docs entry came back in the broken column and the reason looks like a straightforward mix-up rather than an outage. The…
Windows 11 Pro, version 10.0.26200 (please check Help > About; not captured at report time) fswrite (and any operation that produces a session snapshot) fails with ENOENT whenever the target is an absolute Windows path outside ~/.kiro…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again Linux 6.6.87.2-microsoft-standard-WSL2 - Ubuntu 20.04.6 LTS kiro cli should start ● failed to retrieve governance settings — MCP and…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again macOS 26.4.0 (25E246) Windows 客户端升级至 commit a9f293869e317e8f9e195227dffceb57fe52a847 后，远程 SSH 连接 macOS (darwin-arm64) 下载缺失…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again macOS 15.6.1 If I pass in the --identity-provider or --region flag, I'd expect a browser window or automatic login. If I pass in the…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again macOS 15.7.9 (24G830) which kiro-cli kiro-cli 2.29.0 /Users/username/.local/bin/kiro-cli it should properly start with connecting mcp…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again Linux 6.6.87.2-microsoft-standard-WSL2 - Ubuntu 24.04.4 LTS kiro-cli 2.23.0 WSL2 (Ubuntu) on a corporate/enterprise network (MDLZ)…
Would a concise Arcmira remote-MCP configuration example fit the existing Example MCP Servers page? I am Arcmira's founder. Following the contribution guide, I wanted to ask about scope before preparing a docs PR. The example would cover…
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again Linux Red Hat Enterprise Linux 8.10 (Ootpa) kiro-cli chat launch doesn't take super long kiro-cli chat launch takes super long and is…
aider 0.86.2 installed freshly via pip in a venv
I have run q doctor in the affected terminal session I have run q restart and replicated the issue again Linux 6.18.33.2-microsoft-standard-WSL2 - Debian GNU/Linux 13 (trixie) Description: After upgrading kiro-cli to 2.21.4, every session…
Hi Aider maintainers — Hexum offers a hosted MCP server at https://hexum.dev/mcp that shrinks coding-agent turn context, with optional architecture.hexum checks. Would you consider listing/linking it for users, or would a short integration…
I've tried finding an answer on the Continue docs site I'm not able to find an open issue that reports the same bug I've seen the troubleshooting guide on the Continue Docs src/tools/runTerminalCommand.test.ts > runTerminalCommandTool >…
Incorrect or outdated information https://docs.continue.dev/customize/models It's already very old information.
I've tried finding an answer on the Continue docs site I'm not able to find an open issue that reports the same bug I've seen the troubleshooting guide on the Continue Docs
No issues match those filters.