### Changes
- Control UI/dashboard-v2: refresh the gateway dashboard with modular overview, chat, config, agent, and session views, plus a command palette, mobile bottom tabs, and richer chat tools like slash commands, search, export, and pinned messages. (#41503) Thanks @BunsDev.
- OpenAI/GPT-5.4 fast mode: add configurable session-level fast toggles across `/fast`, TUI, Control UI, and ACP, with per-model config defaults and OpenAI/Codex request shaping.
- Anthropic/Claude fast mode: map the shared `/fast` toggle and `params.fastMode` to direct Anthropic API-key `service_tier` requests, with live verification for both Anthropic and OpenAI fast-mode tiers.
- Models/plugins: move Ollama, vLLM, and SGLang onto the provider-plugin architecture, with provider-owned onboarding, discovery, model-picker setup, and post-selection hooks so core provider wiring is more modular.
- Docs/Kubernetes: Add a starter K8s install path with raw manifests, Kind setup, and deployment docs. Thanks @sallyom @dzianisv @egkristi
- Agents/subagents: add `sessions_yield` so orchestrators can end the current turn immediately, skip queued tool work, and carry a hidden follow-up payload into the next session turn. (#36537) thanks @jriff
- Slack/agent replies: support `channelData.slack.blocks` in the shared reply delivery path so agents can send Block Kit messages through standard Slack outbound delivery. (#44592) Thanks @vincentkoc.
### Fixes
- Security/device pairing: switch `/pair` and `openclaw qr` setup codes to short-lived bootstrap tokens so the next release no longer embeds shared gateway credentials in chat or QR pairing payloads. Thanks @lintsinghua.
- Security/plugins: disable implicit workspace plugin auto-load so cloned repositories cannot execute workspace plugin code without an explicit trust decision. (`GHSA-99qw-6mr3-36qr`)(#44174) Thanks @lintsinghua and @vincentkoc.
- Models/Kimi Coding: send `anthropic-messages` tools in native Anthropic format again so `kimi-coding` stops degrading tool calls into XML/plain-text pseudo invocations instead of real `tool_use` blocks. (#38669, #39907, #40552) Thanks @opriz.
- TUI/chat log: reuse the active assistant message component for the same streaming run so `openclaw tui` no longer renders duplicate assistant replies. (#35364) Thanks @lisitan.
- Telegram/model picker: make inline model button selections persist the chosen session model correctly, clear overrides when selecting the configured default, and include effective fallback models in `/models` button validation. (#40105) Thanks @avirweb.
- Cron/proactive delivery: keep isolated direct cron sends out of the write-ahead resend queue so transient-send retries do not replay duplicate proactive messages after restart. (#40646) Thanks @openperf and @vincentkoc.
- Models/Kimi Coding: send the built-in `User-Agent: claude-code/0.1.0` header by default for `kimi-coding` while still allowing explicit provider headers to override it, so Kimi Code subscription auth can work without a local header-injection proxy. (#30099) Thanks @Amineelfarssi and @vincentkoc.
- Models/OpenAI Codex Spark: keep `gpt-5.3-codex-spark` working on the `openai-codex/*` path via resolver fallbacks and clearer Codex-only handling, while continuing to suppress the stale direct `openai/*` Spark row that OpenAI rejects live.
- Ollama/Kimi Cloud: apply the Moonshot Kimi payload compatibility wrapper to Ollama-hosted Kimi models like `kimi-k2.5:cloud`, so tool routing no longer breaks when thinking is enabled. (#41519) Thanks @vincentkoc.
- Moonshot CN API: respect explicit `baseUrl` (api.moonshot.cn) in implicit provider resolution so platform.moonshot.cn API keys authenticate correctly instead of returning HTTP 401. (#33637) Thanks @chengzhichao-xydt.
- Kimi Coding/provider config: respect explicit `models.providers["kimi-coding"].baseUrl` when resolving the implicit provider so custom Kimi Coding endpoints no longer get overwritten by the built-in default. (#36353) Thanks @2233admin.
- Gateway/main-session routing: keep TUI and other `mode:UI` main-session sends on the internal surface when `deliver` is enabled, so replies no longer inherit the session's persisted Telegram/WhatsApp route. (#43918) Thanks @obviyus.
- BlueBubbles/self-chat echo dedupe: drop reflected duplicate webhook copies only when a matching `fromMe` event was just seen for the same chat, body, and timestamp, preventing self-chat loops without broad webhook suppression. Related to #32166. (#38442) Thanks @vincentkoc.
- iMessage/self-chat echo dedupe: drop reflected duplicate copies only when a matching `is_from_me` event was just seen for the same chat, text, and `created_at`, preventing self-chat loops without broad text-only suppression. Related to #32166. (#38440) Thanks @vincentkoc.
- Subagents/completion announce retries: raise the default announce timeout to 90 seconds and stop retrying gateway-timeout failures for externally delivered completion announces, preventing duplicate user-facing completion messages after slow gateway responses. Fixes #41235. Thanks @vasujain00 and @vincentkoc.
- Mattermost/block streaming: fix duplicate message delivery (one threaded, one top-level) when block streaming is active by excluding `replyToId` from the block reply dedup key and adding an explicit `threading` dock to the Mattermost plugin. (#41362) Thanks @mathiasnagler and @vincentkoc.
- Mattermost/reply media delivery: pass agent-scoped `mediaLocalRoots` through shared reply delivery so allowed local files upload correctly from button, slash-command, and model-picker replies. (#44021) Thanks @LyleLiu666.
- macOS/Reminders: add the missing `NSRemindersUsageDescription` to the bundled app so `apple-reminders` can trigger the system permission prompt from OpenClaw.app. (#8559) Thanks @dinakars777.
- Gateway/session discovery: discover disk-only and retired ACP session stores under custom templated `session.store` roots so ACP reconciliation, session-id/session-label targeting, and run-id fallback keep working after restart. (#44176) thanks @gumadeiras.
- Plugins/env-scoped roots: fix plugin discovery/load caches and provenance tracking so same-process `HOME`/`OPENCLAW_HOME` changes no longer reuse stale plugin state or misreport `~/...` plugins as untracked. (#44046) thanks @gumadeiras.
- Models/OpenRouter native ids: canonicalize native OpenRouter model keys across config writes, runtime lookups, fallback management, and `models list --plain`, and migrate legacy duplicated `openrouter/openrouter/...` config entries forward on write.
- Windows/native update: make package installs use the npm update path instead of the git path, carry portable Git into native Windows updates, and mirror the installer's Windows npm env so `openclaw update` no longer dies early on missing `git` or `node-llama-cpp` download setup.
- Sandbox/write: preserve pinned mutation-helper payload stdin so sandboxed `write` no longer reports success while creating empty files. (#43876) Thanks @glitch418x.
- Security/exec approvals: escape invisible Unicode format characters in approval prompts so zero-width command text renders as visible `\u{...}` escapes instead of spoofing the reviewed command. (`GHSA-pcqg-f7rg-xfvv`)(#43687) Thanks @EkiXu and @vincentkoc.
- Hooks/loader: fail closed when workspace hook paths cannot be resolved with `realpath`, so unreadable or broken internal hook paths are skipped instead of falling back to unresolved imports. (#44437) Thanks @vincentkoc.
- Hooks/agent deliveries: dedupe repeated hook requests by optional idempotency key so webhook retries can reuse the first run instead of launching duplicate agent executions. (#44438) Thanks @vincentkoc.
- Security/exec detection: normalize compatibility Unicode and strip invisible formatting code points before obfuscation checks so zero-width and fullwidth command tricks no longer suppress heuristic detection. (`GHSA-9r3v-37xh-2cf6`)(#44091) Thanks @wooluo and @vincentkoc.
- Security/exec allowlist: preserve POSIX case sensitivity and keep `?` within a single path segment so exact-looking allowlist patterns no longer overmatch executables across case or directory boundaries. (`GHSA-f8r2-vg7x-gh8m`)(#43798) Thanks @zpbrent and @vincentkoc.
- Security/commands: require sender ownership for `/config` and `/debug` so authorized non-owner senders can no longer reach owner-only config and runtime debug surfaces. (`GHSA-r7vr-gr74-94p8`)(#44305) Thanks @tdjackey and @vincentkoc.
- Security/gateway auth: clear unbound client-declared scopes on shared-token WebSocket connects so device-less shared-token operators cannot self-declare elevated scopes. (`GHSA-rqpp-rjj8-7wv8`)(#44306) Thanks @LUOYEcode and @vincentkoc.
- Security/browser.request: block persistent browser profile create/delete routes from write-scoped `browser.request` so callers can no longer persist admin-only browser profile changes through the browser control surface. (`GHSA-vmhq-cqm9-6p7q`)(#43800) Thanks @tdjackey and @vincentkoc.
- Security/agent: reject public spawned-run lineage fields and keep workspace inheritance on the internal spawned-session path so external `agent` callers can no longer override the gateway workspace boundary. (`GHSA-2rqg-gjgv-84jm`)(#43801) Thanks @tdjackey and @vincentkoc.
- Security/session_status: enforce sandbox session-tree visibility and shared agent-to-agent access guards before reading or mutating target session state, so sandboxed subagents can no longer inspect parent session metadata or write parent model overrides via `session_status`. (`GHSA-wcxr-59v9-rxr8`)(#43754) Thanks @tdjackey and @vincentkoc.
- Security/agent tools: mark `nodes` as explicitly owner-only and document/test that `canvas` remains a shared trusted-operator surface unless a real boundary bypass exists.
- Security/exec approvals: fail closed for Ruby approval flows that use `-r`, `--require`, or `-I` so approval-backed commands no longer bind only the main script while extra local code-loading flags remain outside the reviewed file snapshot.
- Security/device pairing: cap issued and verified device-token scopes to each paired device's approved scope baseline so stale or overbroad tokens cannot exceed approved access. (`GHSA-2pwv-x786-56f8`)(#43686) Thanks @tdjackey and @vincentkoc.
- Docs/onboarding: align the legacy wizard reference and `openclaw onboard` command docs with the Ollama onboarding flow so all onboarding reference paths now document `--auth-choice ollama`, Cloud + Local mode, and non-interactive usage. (#43473) Thanks @BruceMacD.
- Models/secrets: enforce source-managed SecretRef markers in generated `models.json` so runtime-resolved provider secrets are not persisted when runtime projection is skipped. (#43759) Thanks @joshavant.
- Security/WebSocket preauth: shorten unauthenticated handshake retention and reject oversized pre-auth frames before application-layer parsing to reduce pre-pairing exposure on unsupported public deployments. (`GHSA-jv4g-m82p-2j93`)(#44089) (`GHSA-xwx2-ppv2-wx98`)(#44089) Thanks @ez-lbz and @vincentkoc.
- Security/proxy attachments: restore the shared media-store size cap for persisted browser proxy files so oversized payloads are rejected instead of overriding the intended 5 MB limit. (`GHSA-6rph-mmhp-h7h9`)(#43684) Thanks @tdjackey and @vincentkoc.
- Security/host env: block inherited `GIT_EXEC_PATH` from sanitized host exec environments so Git helper resolution cannot be steered by host environment state. (`GHSA-jf5v-pqgw-gm5m`)(#43685) Thanks @zpbrent and @vincentkoc.
- Security/Feishu webhook: require `encryptKey` alongside `verificationToken` in webhook mode so unsigned forged events are rejected instead of being processed with token-only configuration. (`GHSA-g353-mgv3-8pcj`)(#44087) Thanks @lintsinghua and @vincentkoc.
- Security/Feishu reactions: preserve looked-up group chat typing and fail closed on ambiguous reaction context so group authorization and mention gating cannot be bypassed through synthetic `p2p` reactions. (`GHSA-m69h-jm2f-2pv8`)(#44088) Thanks @zpbrent and @vincentkoc.
- Security/LINE webhook: require signatures for empty-event POST probes too so unsigned requests no longer confirm webhook reachability with a `200` response. (`GHSA-mhxh-9pjm-w7q5`)(#44090) Thanks @TerminalsandCoffee and @vincentkoc.
- Security/Zalo webhook: rate limit invalid secret guesses before auth so weak webhook secrets cannot be brute-forced through unauthenticated churned requests without pre-auth `429` responses. (`GHSA-5m9r-p9g7-679c`)(#44173) Thanks @zpbrent and @vincentkoc.
- Security/Zalouser groups: require stable group IDs for allowlist auth by default and gate mutable group-name matching behind `channels.zalouser.dangerouslyAllowNameMatching`. Thanks @zpbrent.
- Security/Slack and Teams routing: require stable channel and team IDs for allowlist routing by default, with mutable name matching only via each channel's `dangerouslyAllowNameMatching` break-glass flag.
- Security/exec approvals: fail closed for ambiguous inline loader and shell-payload script execution, bind the real script after POSIX shell value-taking flags, and unwrap `pnpm`/`npm exec`/`npx` script runners before approval binding. (`GHSA-57jw-9722-6rf2`)(`GHSA-jvqh-rfmh-jh27`)(`GHSA-x7pp-23xv-mmr4`)(`GHSA-jc5j-vg4r-j5jx`)(#44247) Thanks @tdjackey and @vincentkoc.
- Doctor/gateway service audit: canonicalize service entrypoint paths before comparing them so symlink-vs-realpath installs no longer trigger false "entrypoint does not match the current install" repair prompts. (#43882) Thanks @ngutman.
- Doctor/gateway service audit: earlier groundwork for this fix landed in the superseded #28338 branch. Thanks @realriphub.
- Gateway/session stores: regenerate the Swift push-test protocol models and align Windows native session-store realpath handling so protocol checks and sync session discovery stop drifting on Windows. (#44266) thanks @jalehman.
- Context engine/session routing: forward optional `sessionKey` through context-engine lifecycle calls so plugins can see structured routing metadata during bootstrap, assembly, post-turn ingestion, and compaction. (#44157) thanks @jalehman.
- Agents/failover: classify z.ai `network_error` stop reasons as retryable timeouts so provider connectivity failures trigger fallback instead of surfacing raw unhandled-stop-reason errors. (#43884) Thanks @hougangdev.
- Memory/session sync: add mode-aware post-compaction session reindexing with `agents.defaults.compaction.postIndexSync` plus `agents.defaults.memorySearch.sync.sessions.postCompactionForce`, so compacted session memory can refresh immediately without forcing every deployment into synchronous reindexing. (#25561) thanks @rodrigouroz.
- Telegram/model picker: make inline model button selections persist the chosen session model correctly, clear overrides when selecting the configured default, and include effective fallback models in `/models` button validation. (#40105) Thanks @avirweb.
- Telegram/native command sync: suppress expected `BOT_COMMANDS_TOO_MUCH` retry error noise, add a final fallback summary log, and document the difference between command-menu overflow and real Telegram network failures.
- Mattermost/reply media delivery: pass agent-scoped `mediaLocalRoots` through shared reply delivery so allowed local files upload correctly from button, slash-command, and model-picker replies. (#44021) Thanks @LyleLiu666.
- Plugins/env-scoped roots: fix plugin discovery/load caches and provenance tracking so same-process `HOME`/`OPENCLAW_HOME` changes no longer reuse stale plugin state or misreport `~/...` plugins as untracked. (#44046) thanks @gumadeiras.
- Gateway/session discovery: discover disk-only and retired ACP session stores under custom templated `session.store` roots so ACP reconciliation, session-id/session-label targeting, and run-id fallback keep working after restart. (#44176) thanks @gumadeiras.
- Models/OpenRouter native ids: canonicalize native OpenRouter model keys across config writes, runtime lookups, fallback management, and `models list --plain`, and migrate legacy duplicated `openrouter/openrouter/...` config entries forward on write.
- Gateway/hooks: bucket hook auth failures by forwarded client IP behind trusted proxies and warn when `hooks.allowedAgentIds` leaves hook routing unrestricted.
- Agents/compaction: skip the post-compaction `cache-ttl` marker write when a compaction completed in the same attempt, preventing the next turn from immediately triggering a second tiny compaction. (#28548) thanks @MoerAI.
- Native chat/macOS: add `/new`, `/reset`, and `/clear` reset triggers, keep shared main-session aliases aligned, and ignore stale model-selection completions so native chat state stays in sync across reset and fast model changes. (#10898) Thanks @Nachx639.
- Agents/compaction safeguard: route missing-model and missing-API-key cancellation warnings through the shared subsystem logger so they land in structured and file logs. (#9974) Thanks @dinakars777.
- Cron/doctor: stop flagging canonical `agentTurn` and `systemEvent` payload kinds as legacy cron storage, while still normalizing whitespace-padded and non-canonical variants. (#44012) Thanks @shuicici.
- ACP/client final-message delivery: preserve terminal assistant text snapshots before resolving `end_turn`, so ACP clients no longer drop the last visible reply when the gateway sends the final message body on the terminal chat event. (#17615) Thanks @pjeby.
- Telegram/Discord status reactions: show a temporary compacting reaction during auto-compaction pauses and restore thinking afterward so the bot no longer appears frozen while context is being compacted. (#35474) thanks @Cypherm.