| 19582 | 400 error: reasoning item provided without required following item (OpenAI Responses API) | | | Pending | |
| 19583 | [Bug]: Using Open AI GPT-5 API Key Gives Error | | bug | Pending | |
| 19575 | Fallback chain exits early after 2 model failures | | | Pending | |
| 19577 | Add Chinese (Simplified) language support for Control UI | | enhancement | Pending | |
| 19458 | Feature Request: Markdown rendering in kimi-claw terminal output | | | Pending | |
| 19460 | subagents.model override ignored — sub-agents always use primary model | | | Pending | |
| 19465 | [Bug]: memory-lancedb auto-capture stores assistant/process text and pollutes memory | | | Pending | |
| 19466 | [Bug]: Docker build can miss memory-lancedb runtime dependencies | | | Pending | |
| 19467 | [Bug]: exec tool stdout always empty without PTY on Windows ARM64 | | bug | Pending | |
| 19469 | [Bug]: Cron jobs with sessionTarget "isolated" still never fire on v2026.2.16 (Docker/WSL2) — regression from #10788 | | bug | Pending | |
| 19470 | [Feature]: allow memory_search and memory_get for sub-agents (remove from SUBAGENT_TOOL_DENY_ALWAYS) | | enhancement | Pending | |
| 19472 | Feature: Named agent groups in config for bulk spawning | | | Pending | |
| 19474 | Bug: MiniMax M2.5 always shows as 'configured,missing' | | | Pending | |
| 19475 | Feature Request: Token-free statusline output (like Claude Code) | | | Pending | |
| 19478 | [Bug]: Telegram DM `updateLastRoute` poisons `agent:<id>:main` deliveryContext/origin (raw chatId vs `telegram:` prefix) → WebChat attribution/bleed | | bug | Pending | |
| 19480 | Telegram forward-burst coalescing: add safety caps + stable abuse accounting | | | Pending | |
| 19481 | Normalize inbound origin class for burst grouping (Telegram first, cross-channel ready) | | | Pending | |
| 19488 | compaction.memoryFlush.enabled doesn't trigger memory flush before compaction | | bug | Pending | |
| 19493 | [Bug]: QMD embed spawns multiple concurrent processes, causing extreme CPU usage | | | Pending | |
| 19494 | suppressToolErrors: true doesn't suppress exec tool failure notifications | | | Pending | |
| 19496 | [Bug]: exec and write tools corrupt any input containing [] or {} (shell glob / extra brackets) | | bug | Pending | |
| 19498 | Transient rate-limit errors shown to user before retry succeeds | | | Pending | |
| 19499 | Feature request: outbound message text replacements / transform hook | | | Pending | |
| 19501 | Feature: compaction.model — configurable model override for auto-compaction | | | Pending | |
| 19503 | issue(mattermost): Enabling/Installing Mattermost not working as intended | | bug | Pending | |
| 19506 | Feature Request: sessions_spawn_and_wait for reliable subagent execution | | | Pending | |
| 19509 | WhatsApp Group Messages Not Received - Inbound Broken in 2026.2.15 | | | Pending | |
| 19511 | Signal: support receive-mode on-connection to prevent message loss on startup | | | Pending | |
| 19517 | [Bug]: Agent-Gateway WebSocket Auth Fails After Config Change | | bug | Pending | |
| 19520 | Bug: Telegram streaming (streamMode: partial) broken with MiniMax Portal | | | Pending | |
| 19521 | Gateway plist missing LAUNCH_JOB_LABEL — causes orphan process on restart | | | Pending | |
| 19523 | [Bug]: MiniMax API Malformed Response Corrupts Session - "400 invalid params, tool result's tool id() not found" | | bug | Pending | |
| 19524 | [Bug] Session reset when thinking blocks present in long sessions (thinkingSignature modified during compaction) | | bug | Pending | |
| 19527 | Add Preflight Prompt Budget Enforcement at Model Call Boundary | | | Pending | |
| 19529 | Support Claude Sonnet 4.6 (claude-sonnet-4-6) — released today | | | Pending | |
| 19530 | resolveChannelGroupToolsPolicy receives undefined sender fields in group chats — toolsBySender never matches | | | Pending | |
| 19531 | [Bug] WhatsApp polls sent but not received | | | Pending | |
| 19534 | [Bug]: Anthropic prompt caching broken - Cache Read always 0" | | bug | Pending | |
| 19535 | BlueBubbles extension: always use Private API when available for sends | | | Pending | |
| 19537 | isSilentReplyText regex too permissive — suppresses substantive replies ending with NO_REPLY | | | Pending | |
| 19550 | [Bug]: Context-overflow recovery can finalize empty instead of deterministic fallback | | | Pending | |
| 19552 | Subagent completion notification loop — duplicate deliveries burn tokens indefinitely | | | Pending | |
| 19554 | Docs: web_search supports grok provider but Web Tools docs/config help only mention brave/perplexity | | | Pending | |
| 19564 | Feature: Per-agent session history visibility (granular tools.sessions.visibility) | | | Pending | |
| 19569 | feat(workspace): core document change tracking — git-backed history & rollback for AGENTS.md, SOUL.md, etc. | | | Pending | |
| 19571 | Access Qwen3.5 & Latest Models Same-Day via Higress | OpenClaw 当天接入 Qwen3.5 等最新模型（Higress 集成） | | | Pending | |
| 19572 | [Enhancement] Add caller attribution for Restarted LaunchAgent: gui/501/ai.openclaw.gateway attempts (pid/ppid/parent cmdline) | | | Pending | |
| 19574 | [Bug]: tools.elevated.enabled: true causes elevatedDefault to default to "on", silently breaking all non-PTY exec calls | | bug | Pending | |