openclaw/openclaw v2026.6.8

openclaw 2026.6.8

on GitHub

2026.6.8

Highlights

• Telegram and WhatsApp channel delivery are richer and less brittle: Telegram can send structured rich text with tables, lists, expandable blockquotes, preserved intentional line breaks, prompt-preserving CLI backend delivery, retired native draft migration, and safer rich-media boundaries, while WhatsApp now honors configured ACP bindings. (#92679, #93164, #84082, #89421, #92513) Thanks @obviyus, @jzakirov, @spacegeologist, and @TurboTheTurtle.
• Agent and Gateway recovery is sharper across account-scoped DM sends, generated media completions, auto-reply message-tool final replies, reset archive fallback reads, restart shutdown aborts, yielded subagent pauses, trusted subagent thinking override fallback, yielded cron media, heartbeat dedupe, session identity prompts, and unknown OpenAI agent selector rejection. (#92788, #91246, #92879, #91357, #92631, #92412, #92146, #91287, #92468, #92510) Thanks @yetval, @TurboTheTurtle, @masatohoshino, @CadanHu, @ooiuuii, @openperf, @IWhatsskill, @ZengWen-DT, and @zhangguiping-xydt.
• Provider/model handling expands and tightens with GLM-5.2, Claude Haiku 4.5 catalog rows, OpenRouter and Google Vertex provider-prefix normalization, managed SecretRef auth, OAuth image-default routing through Codex, bounded model browse discovery, LM Studio binary thinking-off delivery, storeless OpenAI Responses replay gating, invalid OpenAI reasoning-signature and genericized Anthropic thinking-signature recovery, Claude 4.5 Copilot tool-streaming safety, and OpenAI/Anthropic-family payload quarantine for unreadable or post-hook tool schemas. (#92796, #90116, #92627, #91218, #90686, #92824, #92247, #92002, #90706, #92941, #92201, #92916, #75393, #92908, #92921, #92928) Thanks @arkyu2077, @liuhao1024, @bymle, @rohitjavvadi, @nxmxbbd, @bek91, @samson910022, @mmyzwl, @CarlCapital, @snowzlm, @Kailigithub, and @vincentkoc.
• /usage and reply payload hooks now have a native full footer renderer, default template, fixed-decimal formatting, credential-aware limits, better partial-count handling, and warnings for broken templates instead of silent bad output. (#92657, #89835, #89629) Thanks @Marvinthebored.
• UI and mobile flows are steadier: workspace files can collapse and start collapsed, WebChat backscroll survives streaming, the sidebar session picker remains interactive above the desktop workbench, reset soft args survive UI dispatch, stale dashboard session parent lineage is preserved, and iOS reconnects stale foreground gateways. (#92779, #92622, #92705, #91353, #90658, #92552) Thanks @shakkernerd, @TurboTheTurtle, @NianJiuZst, @zhouhe-xydt, @luoyanglang, and @Solvely-Colin.
• Memory, state, and diagnostics recover cleaner: oversized OpenAI embedding batches split before 431s, QMD memory search stays available in transient mode, SQLite avoids WAL on NFS state volumes, stuck-session recovery scheduling no longer resets warning backoff, full memory reindexes preserve rollback/cache recovery, raw Memory Wiki source pages stop looking malformed, and Infinity chunk limits stay genuinely unbounded. (#92650, #92618, #92639, #91247, #92752, #92881, #59137, #92876, #69700, #92735) Thanks @mushuiyu886, @TurboTheTurtle, @849261680, @gnanam1990, @TSHOGX, @arlen8411, and @yhterrance.

Changes

• Providers/models: add GLM-5.2 support and Claude Haiku 4.5 catalog entries while keeping provider-qualified model IDs normalized across OpenRouter and Google Vertex paths. (#92796, #90116, #92627, #91218) Thanks @arkyu2077, @liuhao1024, and @bymle.
• Web search: keep key-free providers such as Parallel Free, DuckDuckGo, Ollama, and Codex Hosted Search as explicit opt-ins instead of selecting them automatically when no API-backed provider is configured. (#93616) Thanks @davemorin and @vincentkoc.
• Channel plugins: ship Telegram rich-message delivery and WhatsApp ACP binding support, including preserved intentional line breaks, rich prompt handoff to CLI backends, and transport fixtures for richer drafts. (#92679, #93164, #92513) Thanks @obviyus and @TurboTheTurtle.
• Agent commands: support /btw in CLI-backed sessions and keep CLI usage-error exits classified as usage failures instead of successful runs. (#92669, #92162) Thanks @joshavant and @Pandah97.
• Usage hooks: add built-in full footer rendering, default footer templates, per-turn usage state, credential-aware limits, and fixed-decimal formatting for usage-bar templates. (#92657, #89835, #89629) Thanks @Marvinthebored.
• Docs and operator guidance: document node config examples, clarify before-install hook scope, correct agent default concurrency comments, refresh ZAI provider docs, and update channel/group docs for current Telegram and WhatsApp behavior. (#92677, #92766, #92695) Thanks @liuhao1024, @sallyom, and @ArielSmoliar.

Fixes

• Channels and delivery: preserve account-scoped DM channel send policy, intentional rich-message line breaks in Telegram and status output, rich Telegram final replies, rich Telegram tables and lists, Telegram thread-create CLI remapping, Feishu dynamic-agent routes after persisted binding reuse, Slack outbound message_sent hooks, contributed message-tool schema optionality, same-channel generated media completions, and channel chunking around surrogate pairs and Infinity limits. (#92788, #93164, #92679, #89421, #89943, #42837, #92814, #91137, #91246, #92735) Thanks @yetval, @obviyus, @spacegeologist, @rishitamrakar, @liuhao1024, @lundog, @TurboTheTurtle, and @yhterrance.
• Discord: give generated auto-thread titles a 60-second timeout and 4,096-token reasoning-model output budget, clamped to the selected model output cap. (#64734) Thanks @hanamizuki.
• Agent, cron, and Gateway runtime: mark active main sessions before restart shutdown aborts, pause yielded subagent runs whose terminal also signals abort, clamp trusted subagent thinking overrides through provider/model fallback, preserve yielded media completions, deliver channel message-tool final replies through auto-reply while hiding internal delivery hints, restore reset archive fallback reads when active async transcripts are missing, de-duplicate main-session heartbeat events, expose session identity in runtime prompts, reject unknown OpenAI agent selectors, keep generated media completions, slash-command block replies, and trajectory export commands in WebChat, and require admin privileges for HTTP session/model override surfaces. (#91357, #92631, #92412, #92146, #92879, #91287, #92468, #92510, #91246, #92651, #92646) Thanks @ooiuuii, @openperf, @IWhatsskill, @masatohoshino, @CadanHu, @ZengWen-DT, @zhangguiping-xydt, and @TurboTheTurtle.
• Providers and model replay: preserve storeless OpenAI Responses replay compatibility, recover invalid OpenAI reasoning signatures and genericized Anthropic thinking-signature replay errors, route OAuth image defaults through Codex for eligible OpenAI profiles, avoid eager tool streaming for Claude 4.5 in Copilot, quarantine unreadable and post-hook OpenAI/Anthropic-family tool schemas without broadening allowed tool choices, deliver explicit thinking-off requests to LM Studio binary-thinking models, honor profile auth for SecretRef model entries, bound model browsing, strip provider prefixes where runtimes need bare IDs, and surface nested embedding fetch failures. (#90706, #92941, #92201, #92916, #92824, #75393, #92908, #92921, #92928, #92002, #90686, #92247, #92627, #91218, #92628) Thanks @snowzlm, @mmyzwl, @CarlCapital, @bek91, @Kailigithub, @vincentkoc, @rohitjavvadi, @samson910022, @nxmxbbd, @liuhao1024, @bymle, and @mushuiyu886.
• Memory, state, diagnostics, and config: split header-too-large embedding batches, keep QMD memory search enabled in transient mode, avoid SQLite WAL on NFS volumes, preserve recovery scheduling outside stuck-session warning backoff, preserve full-reindex rollback/cache recovery, treat raw Memory Wiki source pages as source evidence, and keep shell environment fallbacks contained in config write tests. (#92650, #92618, #92639, #91247, #92752, #92881, #59137, #92876, #69700) Thanks @mushuiyu886, @TurboTheTurtle, @849261680, @gnanam1990, @TSHOGX, and @arlen8411.
• UI/mobile/TUI: preserve dashboard session parent lineage, WebChat backscroll, reset soft command args, sidebar session picker interactivity, collapsed workspace files, resolved /model confirmation refs, stale foreground iOS Gateway reconnects, and paused setup-parent stdin after inherited-stdio child exit. (#90658, #92622, #91353, #92705, #92779, #92773, #92552, #93159) Thanks @luoyanglang, @TurboTheTurtle, @zhouhe-xydt, @NianJiuZst, @shakkernerd, @NarahariRaghava, @Solvely-Colin, and @fuller-stack-dev.
• Plugins and updates: repair missing required platform packages during managed plugin installs and updates, including omitted Codex platform binaries.
• Dependencies: update Hono to 4.12.25 so published OpenClaw and ACPX packages use the patched runtime.
• Release and test reliability: extend slow Gateway/full-suite watchdogs, split local full-suite shards when throttled, stabilize plugin auth marker fixtures, avoid brittle provider-ref error text, fold Telegram RTT sampling into live QA evidence, simplify QA scorecard mappings around canonical coverage IDs, keep QA Lab bootstrap selection assertions aligned with flow-only scenarios, skip QA coverage artifact consumers when runtime parity producer status is not green, keep Feishu lifecycle release checks pointed at the active fixture config, isolate trajectory-export live seed turns from Codex-native shell approvals, preserve release-check child refs while pinning expected SHAs, widen live OpenAI TTS budgets for slower provider responses, and avoid false downgrade prompts for unresolved latest-tag updates. (#92652, #92550, #92558, #92911) Thanks @RomneyDa and @Andy312432.

Complete contribution ledger

This audited record covers the complete v2026.6.6..v2026.6.8 history: 167 PRs and 67 linked issues. The grouped notes above prioritize user impact; this ledger preserves every contribution reference and eligible human credit.

Pull requests

• fix(cron): report SQLite storage path in cron.status instead of legacy jobs.json (#92144). Thanks @liuhao1024.
• fix(channel): harden local setup trust (#92175). Thanks @hxy91819.
• fix: handle explicit silent assistant replies (#92073). Thanks @sallyom.
• fix(docker): bundle QA Lab runtime in the image (#92087). Thanks @jesse-merhi.
• fix(anthropic-vertex): stop re-marking cache_control on transport-budgeted payloads (#92387). Thanks @openperf and @Takhoffman.
• Fix doctor preview channel SecretRef resolution (#92229). Thanks @joshavant.
• Fix disabled heartbeat one-shot cron retries (#92225). Thanks @joshavant.
• Fix configured DeepSeek model transport inheritance (#92265). Thanks @joshavant.
• Fail closed for CLI-backed /btw fallback (#92226). Thanks @joshavant.
• Fix suppressed heartbeat commitment delivery (#92231). Thanks @joshavant.
• fix(agents): classify structured unsupported model errors (#92280). Thanks @joshavant.
• Fix OTLP log trace correlation (#92276). Thanks @joshavant.
• fix(update): hand off Linux service auto-updates (#92282). Thanks @joshavant.
• fix: resolve managed SecretRef provider auth (#92235). Thanks @joshavant.
• Fix provider static model fallback resolution (#92293). Thanks @joshavant.
• fix(agent): continue after source message tool replies (#92343). Thanks @joshavant.
• fix(codex): preserve memory prompt registration (#92350). Thanks @rubencu.
• fix: clarify gateway SecretRef auth diagnostics (#92290). Thanks @joshavant.
• fix: repair rejected Anthropic thinking replay (#92286). Thanks @joshavant.
• Fix Telegram spooled buffered replay (#92281). Thanks @joshavant.
• fix(outbound): honor top-level image param as send media source (issue 92407) (#92416). Thanks @xydigit-sj.
• fix(sandbox): render CLI skill prompts from materialized paths (#92508). Thanks @brokemac79.
• chore: fix esbuild production audit failure (#92540). Thanks @RomneyDa.
• Add QA evidence artifact output (#91484). Thanks @RomneyDa.
• Add QA scorecard taxonomy validation (#91500). Thanks @RomneyDa.
• feat(moonshot): add Kimi K2.7 Code support (#92554).
• fix(moonshot): backfill reasoning_content on assistant tool-call replay messages (#92396). Thanks @xialonglee.
• Fix lifecycle timeout cleanup after leader exit (#92566). Thanks @RomneyDa.
• Expose paged channel action results (#88993). Thanks @fuller-stack-dev.
• fix(fireworks): resolve catalog model params from plugin.json via core (#90326). Thanks @obuchowski.
• fix(doctor): warn for untrusted external Discord plugin (#86629). Thanks @brokemac79.
• fix(providers): skip unreadable Mistral tool schemas (#90242). Thanks @vincentkoc.
• fix(reply): mirror same-channel Slack final replies (#92498). Thanks @TurboTheTurtle.
• fix(channels): default boundary logger for swallowed progress-draft start errors (#92083). Thanks @hansraj316.
• fix(channels): make timer-fired progress-draft start errors observable (#92031). Thanks @hansraj316.
• fix(agents): isolate invalid plugin model catalogs [AI-assisted] (#92564). Thanks @tangtaizong666.
• docs: UX-013 — design system documentation (#89827). Thanks @BunsDev.
• feat(ui): hide empty workboard columns (#89615). Thanks @BunsDev.
• fix(a11y): B-1+B-2+B-3 — contrast, focus states, minimum font sizes (#89822). Thanks @BunsDev.
• fix issue 92218: memory_search tool disabled with QMD backend (#92618). Thanks @mushuiyu886.
• docs(gateway): add uptime monitoring guidance to health check docs (fixes issue 55768) (#92608). Thanks @liuhao1024.
• fix(docs): pin Windows Hub download links to v2026.6.5 (#92605). Thanks @lzyyzznl.
• issue 92589: fix(internal-runtime-context): wrap prompt-preface runtime context body in delimiters (#92593). Thanks @zhangqueping.
• Run Vitest and Playwright scenarios from qa suite (#92606). Thanks @RomneyDa.
• feat(hooks): per-turn usageState on reply_payload_sending (#89629). Thanks @Marvinthebored.
• feat(usage): native templated /usage full footer renderer (#89835). Thanks @Marvinthebored.
• fix(models): bound /models and models list catalog loading (#92247). Thanks @samson910022.
• fix(gateway): honor profile auth for SecretRef model entries (#90686). Thanks @rohitjavvadi.
• fix: require admin for HTTP session kills (#92651). Thanks @steipete-oai.
• test(models): stabilize plugin auth marker fixtures (#92652).
• fix(slack): warn when channels map is keyed by name instead of channel ID (#89438). Thanks @Alix-007.
• fix(agents): pause yielded subagent runs whose terminal also signals abort (#92631). Thanks @openperf.
• fix(ui): preserve WebChat backscroll during streaming (#92622). Thanks @TurboTheTurtle.
• fix(openrouter): strip openrouter/ prefix from model ID in normalizeResolvedModel hook (fixes issue 92611) (#92627). Thanks @liuhao1024.
• fix(cron): preserve yielded media completions (#92146). Thanks @IWhatsskill.
• fix: add Claude Haiku 4.5 static catalog entries (#90116). Thanks @arkyu2077.
• fix(channels): keep contributed message-tool schema properties optional (#91137). Thanks @lundog.
• fix(copilot): disable eager tool streaming for Claude 4.5 (#75393). Thanks @Kailigithub.
• fix issue 73713: surface nested embedding fetch failures (#92628). Thanks @mushuiyu886.
• fix(slack): emit message_sent hook on outbound delivery (mirror Telegram) (#89943). Thanks @rishitamrakar.
• fix(docs): finalize i18n postprocess before skip (#92668). Thanks @hxy91819.
• fix: split image setup and request timeout semantics (#92673). Thanks @hxy91819.
• fix(memory): keep memory_search in transient qmd mode (#92639). Thanks @TurboTheTurtle and @Takhoffman.
• fix(ui): restore sidebar session picker interactivity above desktop workbench (#92705). Thanks @NianJiuZst.
• feat: support /btw in CLI-backed sessions (#92669). Thanks @joshavant.
• fix(gateway): mark active main sessions before restart shutdown aborts (#91357). Thanks @ooiuuii.
• fix(ios): force stale foreground gateway reconnects (#92552). Thanks @Solvely-Colin.
• fix(diagnostics): keep recovery scheduling out of the stuck-session warning backoff (#92752). Thanks @gnanam1990 and @Takhoffman.
• clarify before_install hook scope (#92766). Thanks @sallyom.
• Honor WhatsApp configured ACP bindings (#92513). Thanks @TurboTheTurtle and @mcaxtr.
• feat(providers): add GLM-5.2 support (#92796).
• fix(heartbeat): route outbound mirror to isolated session key (#92807). Thanks @agent-merkava.
• fix(memory): explain skipped short-term recall hits (#92745). Thanks @mushuiyu886.
• fix(gateway): forward image-only input on /v1/responses (parity with chat completions) (#92488). Thanks @s554097550.
• fix(status): avoid cumulative usage for context percent (#92604). Thanks @ashishpatel26.
• fix(nodes): surface pending reapproval diagnostics (#92547). Thanks @fuller-stack-dev.
• fix(doctor): avoid false-positive legacy cron store warning when store was already migrated (fixes issue 92683) (#92690). Thanks @liuhao1024.
• fix(telegram): skip IPv4 fallback when user explicitly configures non-ipv4first dnsResultOrder (fixes issue 41671) (#92806). Thanks @liuhao1024 and @vincentkoc.
• fix(macos): defer isOverflowing mutation to break SwiftUI render loop (fixes issue 43480) (#92778). Thanks @liuhao1024 and @vincentkoc.
• fix(gateway): use resolveNonNegativeNumber for totalTokens to display 0 instead of ? (fixes issue 43009) (#92795). Thanks @liuhao1024 and @vincentkoc.
• fix(gateway): preserve active runs during plugin finalization (#92746). Thanks @scotthuang and @vincentkoc.
• UI: localize Logs tab labels (#92820). Thanks @rubensfox20.
• UI: localize logs hardcoded labels (#61080). Thanks @rubensfox20.
• fix(telegram): preserve command callbacks while prefixing generic callback data (#92825). Thanks @hnshah.
• fix(telegram): add 'callback_data:' prefix to inline button callbacks (#54962). Thanks @hnshah.
• fix(copilot): drop thinking blocks from replay (#87060). Thanks @giodl73-repo.
• fix(github-copilot): strip thinking blocks from latest assistant turn (issue 81520) (#81534). Thanks @SymbolStar.
• feat(browser): extend --labels overlay to full-page and element captures (#92834). Thanks @hxy91819.
• fix issue 92039: [Bug]: WhatsApp login reports success before auth is durably persisted, so Docker rebuilds/upgrades can force relink (#92095). Thanks @zhangguiping-xydt.
• fix(stale): exempt ClawSweeper actionable labels from stale lifecycle (fixes issue 89564) (#92801). Thanks @liuhao1024.
• fix(status): render sub-1000 token counts as plain integers (#89736). Thanks @jbetala7 and @vincentkoc.
• fix(agents): catch malformed image blocks in sanitizeContentBlocksImages (#92792). Thanks @LowCode191 and @vincentkoc.
• ci: gate stable releases on Windows companion assets (#92555). Thanks @fuller-stack-dev.
• fix(agents): add usage guidance to sessions_spawn tool description (fixes issue 91814) (#91824). Thanks @zenglingbiao.
• fix(qqbot): surface failed media sends (#92823). Thanks @zhangguiping-xydt.
• Fix diagnostics OTEL runtime install trust (#92045). Thanks @efpiva.
• fix(update): continue after package doctor warnings (#91586). Thanks @fuller-stack-dev.
• fix(feishu): target typing reaction on inbound (#67783). Thanks @huiwen01.
• fix(feishu): preserve root_id thread routing without thread_id forcing (#73958). Thanks @huiwen01.
• fix(lobster): surface workflow path errors (#68106). Thanks @vvitovec.
• fix(openai): preserve opaque reasoning transcript fields (#90682). Thanks @toruvieI.
• fix(anthropic): strip thinking blocks from history when thinking is disabled (fixes issue 92360) (#92373). Thanks @liuhao1024.
• fix(anthropic): merge consecutive assistant turns in turn validation (#87346). Thanks @Jefsky.
• fix(anthropic): quarantine invalid direct tool schemas (#92896).
• fix(anthropic): quarantine invalid projected tools (#89418). Thanks @vincentkoc.
• fix(agents): guard Anthropic tool descriptors (#89221). Thanks @vincentkoc.
• fix(agents): guard Anthropic tool schema conversion (#90228). Thanks @vincentkoc.
• fix(agents): skip unreadable Anthropic tool schemas (#89622). Thanks @vincentkoc.
• fix(llm): guard Anthropic provider tool descriptors (#89229). Thanks @vincentkoc.
• fix(providers): skip unreadable Anthropic tool schemas (#90278). Thanks @vincentkoc.
• fix(active-memory): preserve verbose recall summaries (#90739). Thanks @brokemac79.
• Simplify QA scorecard mapping shape (#92558). Thanks @RomneyDa.
• fix(memory-wiki): stop flagging raw source pages as malformed (#92876). Thanks @vincentkoc.
• fix(providers): quarantine unreadable Anthropic payload tools (#92908).
• fix(memory): preserve reindex rollback recovery (#92881). Thanks @TSHOGX and @vincentkoc.
• fix(openai): quarantine unreadable tool schemas (#92921).
• fix(openai): quarantine unreadable projected tools (#89413). Thanks @vincentkoc.
• fix(agents): materialize OpenAI tool schemas (#89013). Thanks @vincentkoc.
• fix(agents): guard OpenAI transport tool descriptors (#89016). Thanks @vincentkoc.
• fix(agents): guard OpenAI tool schema conversion (#89378). Thanks @vincentkoc.
• fix(agents): harden OpenAI strict schema inspection (#89543). Thanks @vincentkoc.
• fix(agents): guard OpenAI strict tool diagnostics (#90200). Thanks @vincentkoc.
• fix(providers): skip unreadable OpenAI completion tool schemas (#90283). Thanks @vincentkoc.
• fix(providers): skip unreadable OpenAI responses tool schemas (#90286). Thanks @vincentkoc.
• fix(openai): skip unreadable responses tool schemas (#90397). Thanks @vincentkoc.
• Fold Telegram RTT sampling into live QA evidence (#92550). Thanks @RomneyDa.
• fix(media): route OAuth image defaults through Codex (#92824). Thanks @bek91.
• fix(cli): avoid false downgrade prompt for latest tag (#92911). Thanks @Andy312432 and @vincentkoc.
• fix(openai): guard post-hook tool payloads (#92928).
• fix(openai): guard responses tool payload names (#89703). Thanks @vincentkoc.
• fix(sessions): fall back to reset archive for missing async transcripts (#92879). Thanks @masatohoshino and @CadanHu and @vincentkoc.
• fix(feishu): re-resolve route when dynamic agent binding already exists in runtime config (fixes issue 42837) (#92814). Thanks @liuhao1024 and @vincentkoc.
• fix(openai): omit gpt-5.5 tool reasoning effort (#90574). Thanks @BSG2000.
• fix(openai): recover invalid reasoning signatures (#92941).
• fix(lmstudio): deliver thinking "off" to binary-thinking models (#92002). Thanks @nxmxbbd.
• issue 92201: Embedded runner: freshly streamed thinking signatures intermittently invalid on replay (Anthropic); recovery wrapper never fires because error text is genericized (#92916). Thanks @mmyzwl.
• fix(agents): do not misclassify client-disconnect abort as run timeout (#90936). Thanks @openperf.
• fix(agents): make wrapToolWithBeforeToolCallHook idempotent to prevent double hook execution (fixes issue 92973) (#93009). Thanks @zenglingbiao.
• fix(cron): require explicit message target proof (#92318). Thanks @hxy91819.
• fix(gateway): repair usage cost aggregation across agents (#93022). Thanks @luke-skywalker-open-claw and @stablegenius49.
• fix(tui): keep parent stdin paused after exit (#93159). Thanks @fuller-stack-dev.
• Keep key-free web search providers opt-in (#93616). Thanks @davemorin and @vincentkoc.
• feat(telegram): send rich message text (#92679). Thanks @obviyus.
• fix(telegram): preserve rich markdown line breaks (#93164). Thanks @vincentkoc.
• fix(telegram): allow expandable blockquotes (#84082). Thanks @jzakirov.
• fix(telegram): expose thread create CLI remap (#89421). Thanks @spacegeologist.
• fix(sessions): derive channel from account-scoped DM session keys in send-policy (#92788). Thanks @yetval.
• Fix webchat media completion handoff (#91246). Thanks @TurboTheTurtle.
• fix(cron): de-duplicate main-session systemEvent in heartbeat model input (#91287). Thanks @ZengWen-DT.
• fix issue 92453: add session identity to runtime prompt (#92468). Thanks @zhangguiping-xydt.
• fix(gateway): reject unknown OpenAI agent selectors (#92510). Thanks @zhangguiping-xydt.
• fix(google): strip provider prefix from Vertex model path (#91218). Thanks @bymle.
• fix(OpenAI Responses): disable item id replay for storeless providers (#90706). Thanks @snowzlm.
• feat(usage): ship built-in /usage full footer (#92657). Thanks @Marvinthebored.
• fix: start workspace files collapsed (#92779). Thanks @shakkernerd.
• fix(ui): preserve /reset soft args in Control UI dispatch (#91353). Thanks @zhouhe-xydt.
• fix(ui): preserve dashboard session parent lineage when session list is stale (#90658). Thanks @luoyanglang.
• fix issue 92465: split OpenAI 431 embedding batches (#92650). Thanks @mushuiyu886.
• fix(state): avoid sqlite wal on nfs state volumes (#91247). Thanks @849261680.
• fix(memory): preserve retry state and embedding cache across reindex rollback (#59137). Thanks @TSHOGX.
• fix(markdown-core): treat Infinity chunk limit as unbounded, not 1 (#92735). Thanks @yhterrance.
• issue 92069: fix(cli): usage errors exit 0 (#92162). Thanks @Pandah97.
• docs(nodes): add openclaw.json config example to Nodes overview (#92677). Thanks @liuhao1024.
• docs(config): correct maxConcurrent default in agent-defaults type comments (AI-assisted) (#92695). Thanks @ArielSmoliar.
• fix(discord): raise thread title timeout and tokens to fit reasoning models (#64734). Thanks @hanamizuki.
• fix: require admin for HTTP model overrides (#92646). Thanks @steipete-oai.
• fix(tui): show resolved canonical model ref in /model confirmation (#92773). Thanks @NarahariRaghava.

Linked issues

• Reported: [Bug]: openclaw cron status reports legacy storePath (#91766). Thanks @AaronFaby.
• Reported: [Bug]: Node.js auto-installer fails silently with ioctl errors then falsely reports success before crashing (#73837). Thanks @ItsMeForLua.
• Reported: [Bug]: Missing SQLite perf and query-plan harness (#91616). Thanks @galiniliev.
• Reported: [Feature]: Error: Gateway service install not supported on openbsd (#25621). Thanks @kucharskim.
• Reported: [Bug]: cron edit --cron silently strips schedule.tz and staggerMs (direct path replaces schedule without merging) (#92291). Thanks @dcapclaw.
• Reported: message tool: image param silently dropped on send — delivers text without attachment but returns ok:true (#92407). Thanks @ichirokyoto.
• Reported: Kimi K2.6 reasoning_content 400 regression in long conversations after LCM compaction (follow-up issue 70392) (#71491). Thanks @RoseKongPS.
• Reported: [Bug]: Moonshot/Kimi duplicate tool-call IDs in replay, exposed by WhatsApp group chats (#51593). Thanks @Faaab84 and @Pluviobyte.
• Reported: Discord channel stays disabled with no warning unless plugins.entries.discord.enabled is set (#83212). Thanks @cdeyoung67.
• Reported: [Bug]: Slack channel/thread sessions never persist assistant replies to the session transcript → total context loss when the CLI session binding is invalidated (#92489). Thanks @TalkingHeadsJed and @TurboTheTurtle.
• Reported: ModelRegistry: a single invalid plugin catalog aborts the entire custom-models load, leaving zero models and an unlogged error (#92553). Thanks @fxstein.
• Reported: Health check bloat: uptime monitors must use /health, not /v1/chat/completions (#55768). Thanks @faahim.
• Reported: [Bug]: Windows Hub download link is not working (#92470). Thanks @arjkul.
• Reported: Feishu channel leaks system runtime context (relevant-memories, sender metadata) into user-visible reply (#92589). Thanks @jovi2014-cyber.
• Reported: [Performance] /models command slow in v2026.6.1 — catalog loading regression (#91809). Thanks @syfvb.
• Reported: models.list marks auth-profile-backed SecretRef providers unavailable (#90685). Thanks @rohitjavvadi.
• Reported: Name-keyed entries in channels.slack.channels silently dropped under groupPolicy: "allowlist" (#81665). Thanks @cjalden.
• Reported: [Bug]: sessions_yield in a depth-1 subagent settles its background task as "cancelled" (operator-reserved status) and delivers a false "Background task cancelled" notice to the requester session (#92448). Thanks @aleps001.
• Reported: [Bug]: OpenRouter: Anthropic models send wrong model ID to API (includes openrouter/ prefix) (#92611). Thanks @lijenhsin.
• Reported: anthropic (api_key) provider: Claude Haiku 4.5 missing from static model catalog → "Unknown model" (model_not_found) (#90088). Thanks @maaron34.
• Reported: github-copilot: tools[].eager_input_streaming still rejected on v2026.4.29 (re: issue 72183) (#75348). Thanks @finchinslc and @Kailigithub.
• Reported: openclaw infer embedding create fails with TypeError: fetch failed on Node 24 despite valid Voyage credential; underlying cause is swallowed (#73713). Thanks @crsnpalmer-art.
• Reported: Gateway runs well-formed-but-unknown agent slug under agents.defaults instead of 4xx (no roster validation in resolveAgentIdForRequest; x-openclaw-agent-id header never roster-validated) (#92504). Thanks @ryanhelms.
• Reported: [Bug]: Delivery retry loop corrupts active sessions (R-004) — retry selector bypasses delivery.mode=none (#91420). Thanks @CarotaWealth.
• Reported: Agent runtime header lacks session identity, causing misleading self-references (#92453). Thanks @QQSHI13.
• Reported: fix(slack): emit message_sent hook on outbound delivery (mirror Telegram) (#89942). Thanks @rishitamrakar.
• Reported: bug(cli): usage errors exit 0 (#92069). Thanks @marcospaulo.
• Reported: void requireRef silences dead-import lint via side-effect expression (#83878). Thanks @davinci282828.
• Reported: formatDiskSpaceBytes emits "1024 MiB" instead of "1.0 GiB" at the GiB boundary (#90245). Thanks @jbetala7.
• Reported: setFeishuClientRuntimeForTest resets the SDK without clearing the client cache (#83911). Thanks @davinci282828.
• Reported: [Bug]: Cron job with sessionTarget: "main" triggers both systemEvent and reminder despite delivery.mode: "none" (#44922). Thanks @GSL-R.
• Reported: music_generate background task completion delivery consistently fails (completion wake + fallback both fail) (#91003). Thanks @kumaxs.
• Reported: [Bug]: In the Control UI, /reset soft is truncated to /reset when executed, and the args are lost (#91316). Thanks @MaBeitian.
• Reported: Bug: dashboard child sessions record assistant replies but do not display them (#90623). Thanks @lily-oc.
• Reported: openclaw message thread create for Telegram: thread-create → topic-create remap not happening; gateway rejects with Unsupported Telegram action (#81581). Thanks @myrzka.
• Reported: [Bug]: GatewayRequestError: Error: file is not a database: code=ERR_SQLITE_ERROR (#90491). Thanks @AFabyTWE.
• Reported: Signal image captions truncated to first character (Infinity chunk limit normalizes to 1) (#92734). Thanks @yhterrance.
• Reported: Docs feedback: /nodes (#92662). Thanks @Casper-Mars.
• Reported: Bulk memory import can hit OpenAI 431; chunked indexing avoids it (#92465). Thanks @BrettHamlin.
• Reported: Docker image ships an extraneous stale openclaw in /app/node_modules (extensions pin the published release) (#92551). Thanks @fxstein.
• Reported: Telegram callback queries time out when agent turn is queued behind sequentialize (#42156). Thanks @Diaspar4u.
• Reported: doctor + cron status still report the retired cron/jobs.json store after the SQLite migration (2026.6.5) (#92683). Thanks @motteman.
• Reported: [Bug]: Telegram media download fails on IPv4-broken / IPv6-working hosts because runtime IPv4 fallback overrides config (#41671). Thanks @leandroirani933-ctrl and @vincentkoc.
• Reported: macOS app pinwheels due to SwiftUI infinite render loop in VoiceWakeOverlay (#43480). Thanks @gdiab and @vincentkoc.
• Reported: TUI displays Context Tokens as ?/200k instead of actual value (#43009). Thanks @ltxy12138-ai and @vincentkoc.
• Reported: [Bug]: Telegram inline button callback_query not routed to agent — hallucination instead of tool call (#54909). Thanks @timt80 and @hnshah.
• Reported: Status/session context window can over-report the selected model's actual window (#39857). Thanks @xdanger.
• Reported: [Bug]: Reasoning model thinking blocks ( tags) in conversation history cause HTTP 400 on GitHub Copilot provider (#81520). Thanks @warcold.
• Reported: fix(memory): EPERM on Windows persists after 64187 retry — needs copyFile/unlink fallback (was in closed PR 71611) (#78640). Thanks @MilleniumGenAI.
• Reported: [Bug]: WhatsApp block streaming can suppress complete final replies after partial stream delivery (#81078). Thanks @Jackten.
• Reported: Bug: Twilio voice-call can get stuck in hold music after failed/no-stream call (#81122). Thanks @donkeykong91.
• Reported: [Bug]: Stale workflow does not exempt ClawSweeper queueable issues (#89564). Thanks @brokemac79.
• Reported: [Bug]: openclaw status renders sub-1000 token counts as misleading fractional k (999 → "1.0k") (#89735). Thanks @jbetala7 and @vincentkoc.
• Reported: [Feature]: sessions_spawn tool description lacks usage guidance, causing agents to not use sub-agents when appropriate (#91814). Thanks @cattails-lgao.
• Reported: [Bug]: Lobster tool falls back to pipeline parsing for relative workflow file paths (#68101). Thanks @MPC7500 and @vvitovec.
• Reported: openai-chatgpt-responses native replay sends encrypted reasoning and breaks next turn with invalid_encrypted_content (#90093). Thanks @richardmqq.
• Reported: [Bug]: Persistent sessions corrupted by stale thinking blocks — provider rejects all subsequent turns (#92360). Thanks @notnaji.
• Reported: Bug: Subagent announce-delivery echo messages inherit wrong provider/model metadata, causing persistent "thinking blocks cannot be modified" errors after gateway restart (#87329). Thanks @travellingsoldier85.
• Reported: Codex-authenticated installs can auto-select direct OpenAI for image media understanding without OPENAI_API_KEY (#87168). Thanks @bek91.
• Reported: [Bug]: Feishu dynamicAgentCreation feature not working (#42837). Thanks @cwlong163-afk and @vincentkoc.
• Reported: BUG: sessions_spawn silently half-fails when thinking level is unsupported — fan-out spawns produce non-deterministic survivors, no signal to orchestrator (fix: symmetrize CLI-launch fallback with embedded path) (#92412). Thanks @oiGaDio.
• Reported: Config hot-reload permanently disabled when inotify watches exhausted (no polling fallback) (#92851). Thanks @danbao.
• Reported: Gateway becomes slow or times out under multi-session / multi-agent load (#92057). Thanks @xiaopings.
• Reported: fix(memory-wiki): guard against missing agentIds (#92207). Thanks @qq230849622-a11y.
• Reported: Embedded runner: freshly streamed thinking signatures intermittently invalid on replay (Anthropic); recovery wrapper never fires because error text is genericized (#92201). Thanks @CarlCapital.
• Reported: before_tool_call hook fires twice: tools double-wrapped after normalizeToolParameters strips the wrap marker (#92973). Thanks @dertbv.
• Reported: memory-wiki lint: sources/ directory requires frontmatter on raw imported files (#69700). Thanks @arlen8411.

Release verification

• npm package: https://www.npmjs.com/package/openclaw/v/2026.6.8

• registry tarball: https://registry.npmjs.org/openclaw/-/openclaw-2026.6.8.tgz

• integrity: sha512-iziR8fi69+ojrtX7FYYvTpkGcVnmyLpIhvchgt5LFkkdHVWw973XAAekKVZ3/xQJ5FG4NwgHkXL0LLTrgsNOSQ==

• release SHA: 844f405ac1be805d5c598922a37254f12ab6d765

• full release CI report: https://github.com/openclaw/releases/blob/main/evidence/2026.6.8/release-evidence.md

• release publish: https://github.com/openclaw/openclaw/actions/runs/27631170936

• npm preflight: https://github.com/openclaw/openclaw/actions/runs/27627935264

• full release validation: https://github.com/openclaw/openclaw/actions/runs/27627935341

• plugin npm publish: https://github.com/openclaw/openclaw/actions/runs/27631412538

• plugin ClawHub publish: https://github.com/openclaw/openclaw/actions/runs/27631418543

• plugin ClawHub bootstrap: not needed

• OpenClaw npm publish: https://github.com/openclaw/openclaw/actions/runs/27632229765

• npm Telegram beta E2E: not supplied

• Windows Hub promotion: https://github.com/openclaw/openclaw/actions/runs/27632639125 from openclaw/openclaw-windows-node@v0.6.3

• macOS signed/notarized preflight: https://github.com/openclaw/releases/actions/runs/27632793120

• macOS Swift validation: https://github.com/openclaw/releases/actions/runs/27632792831

• macOS asset promotion: https://github.com/openclaw/releases/actions/runs/27635471007

• stable appcast: #93722