openclaw.changelog()

// Safer upgrades and retained history: Doctor preserves session history and repeated-repair state, finishes upgrades with invalid retained histories, and avoids repeatedly stalling or stopping a Gateway that is still starting. (149741, 149956, 148901, 149308, 149823, 149760) Related 149822 Thanks @mas

// OpenClaw 2026.9.4 makes plugins and skills easier to find and lets you turn past conversations into reusable skills through a chat you can steer. Adds GPT Image 2.5 support, more control over cloud sessions, and interactive questions in the terminal. When the CLI cannot start because of Node, OpenClaw can now find a compatible copy or offer to install one just for OpenClaw, without touching your system Node. Release scale: 1,558 pull requests from 294 contributors.

// This is the final June 2026 Extended Stable (LTS) release. Bundled providers and channel adapters now bound untrusted response bodies and reject oversized inputs before expensive work; agent, gateway, retry, and channel paths handle cancellation, retries, partial sends, and stream failures without losing work; and local-model, browser, media, and collaboration plugins recover cleanly from malformed payloads and transient upstream failures.

// Updates now rehearse core and plugin changes in isolated candidate state before activation. Skill Workshop keeps skills in one persistent agent-owned collection across workspaces. You can publish a revocable read-only view of a session's conversation via public link, browse a searchable meeting library with full transcripts, and manage connected provider accounts in Models settings. Warm prompt caches are preserved for snappier sessions. (138839, 135528, 139489)

// Select openai/gpt-6-astra with an OpenAI API-key profile or an eligible ChatGPT/Codex account, with text and image input and Responses tool calls. Chat, dashboards, and session interactions stay responsive while long transcripts and disk usage are processed; active, queued, and delegated replies recover after Gateway restarts; Git backups preserve text containing embedded NUL characters and support Nix-managed config and credential links. (137550, 138071, 136236)

// Mermaid blocks render as diagrams in the Control UI and in the native macOS, iOS, and Android apps. Fresh installs (including npx openclaw@latest) get a quick-start lane that detects existing Claude Code or Codex logins and API keys and verifies them live. Personal skill libraries live beside the workspace set via openclaw skills library, and openclaw update now rolls back the npm candidate when post-update Doctor fails — users on 2026.8.2 without a service manager should run openclaw update --no-restart once. (134913, 134221, 134068, 135462)

// Install the .deb or AppImage on x86-64 Linux, connect to a local or remote Gateway, and open Quick Chat from the system tray; AppImage updates are signature-verified. Open your Home agent in a right or bottom dock with Cmd/Ctrl+Shift+H, create and run background sessions from New Session, and get final answers after settled tool work instead of conversations stopping at tool output. (133676, 128050, 133520)

// Search visible conversation text by exact words or phrases and reopen the surrounding messages. Run work on paired devices or cloud workers and move the session workspace with it. Use widgets in chat, pin them to session dashboards, and export rendered views as images; answer structured agent questions through web and native cards, messaging buttons, or plain text; and follow a durable session progress card across web and native chat. (105057, 123280, 101840, 109922)

// npm plugin updates: accept singleton-array metadata from newer npm clients so tracked official plugins can install and update to correction releases. (108336)

// Codex progress replies: keep app-server turns running after delivered progress messages so GPT/Codex reaches its authoritative terminal response instead of stopping mid-turn. (106961, 108487) Thanks @joshavant.

// Sandboxed browser routes, trusted DNS targets, custom browser origins, and loopback provider endpoints now reject unsafe access paths. Retained session writes, provider fallbacks, and stream progress recover without silently ending active work; pending channel work resumes after recovery and sustained Discord gateway bursts stay bounded. SQLite checkpoints and gateway process signalling no longer turn transient host conditions into failed runs. (97958, 96100, 97041, 99067)

// Package installations can now select, update from, and receive availability notices for the extended-stable channel without silently falling back to another release line. Provider streams, Discord REST responses, browser fetches, and OAuth paths cap hostile response sizes; watchdog semantics distinguish genuine stalls from active long model calls; Discord reconnects no longer silently drop queued messages, and service restarts preserve SecretRef-backed Telegram credentials. (99811, 96989, 102160, 100896)

// A major Control UI rework organizes conversations side by side with live Tasks, clearer usage and cost views, and files, pairing, approvals, and Gateway health kept close to the conversation. Guided setup walks from install to a working first chat and checks connections before saving them. iOS, Android, and macOS apps receive substantial updates; GPT-5.6 compatibility improves across OpenAI and Codex routes; openclaw attach gives Claude Code temporary access to a selected session. 3,063 contributions from 532 contributors.

// A dependability-focused release fixing misplaced replies, stuck sends, reconnects, model setup failures, and unsafe admin defaults across Telegram, WhatsApp, Matrix, Google Chat, iMessage, Feishu, Mattermost, WebChat, the Control UI, and the terminal UI. Newer Google Chat direct messages reach the correct one-to-one chat, background media results return to the requesting chat, and Telegram answers stay attached to the current question when quoting an earlier bot message. (58993, 89911, 89949, 90475)

// Adds /fast auto so short conversational calls can start quickly, while longer or fallback work returns to normal mode with the effective state still visible. PR 85104, Issue 85087. Thanks @alexph-dev and @vincentkoc. Shows the effective automatic fast-mode state in status instead of reducing it to o

// Richer Telegram delivery: Telegram now sends rich HTML, preserves rich markdown and sticker paths, renders progress drafts and command output more faithfully, normalizes HTML tables safely, and keeps mentions and spooled handlers on the right delivery path. (93286, 93164, 93124, 93364, 93130, 93088,

// Richer channel delivery: Telegram and WhatsApp are less brittle: Telegram renders structured text with tables, lists, expandable blockquotes, preserved intentional line breaks, and CLI-backed replies, while WhatsApp now honors configured ACP bindings. (92679, 93164, 84082, 89421, 92513) Thanks @obvi

// Tighter security boundaries: transcript, sandbox, MCP, browser, channel, and exec-approval paths now fail closed around unsafe access, timed-out approvals, and malformed boundary input. (91529, 91618, 91741, 91750, 89938) Thanks @joshavant, @pgondhi987, @mmaps, @eleqtrizit, @drobison00, @vincentkoc,

// Safer channel output: QQBot strips model reasoning and thinking scaffolding before native delivery, so users see the final answer rather than raw internal markup. (89913, 90132) Thanks @openperf, @dygg2001, @Takhoffman, and @vincentkoc. MCP results no longer poison sessions: resource_link, audio, ma

// Resilient agent and Codex runs: interrupted tool calls, stale session bindings, compaction handoffs, auth-profile failover, reasoning-tag cleanup, yielded subagents, and generated-media delivery all recover without leaving work stranded. (85798, 87484, 88182, 89220) Thanks @RomneyDa, @neeravmakwana,

// Agent and Codex runtime recovery is steadier: subagents keep cwd/workspace separation, hook context stays prompt-local, session locks release on timeout abort while live OpenClaw locks survive cleanup, stale restart continuations are avoided, and Codex app-server/helper failures no longer tear down

// Stronger security and content boundaries: group prompt text is kept out of the system prompt, repeated-dot hostnames are normalized, side-effecting command wrappers and unsafe Node runtime env overrides are blocked, no-auth Tailscale exposure is rejected, and node/device-role approvals now require a

// Faster Gateway and replies: startup avoids repeated plugin, channel, session, usage-cost, warning, scheduled-service, and filesystem scans; visible replies separate user-facing sends from slower follow-up work; Gateway runtime/session caches churn less under load. Transcripts are core: transcript-ba

// Gateway/perf: reuse process-stable channel catalog reads, avoid repeated bundled-channel boundary checks, and rotate gateway watch CPU profiles so benchmark runs do not accumulate unbounded artifacts. Gateway/perf: reuse immutable plugin metadata snapshots across startup, config, model, channel, set

// Exec approvals: remove the old cat SKILL.md && printf ... && <skill-wrapper> allowlist compatibility path so skill files must be loaded with the read tool and only the real skill executable is auto-allowed. Discord: let voice sessions follow configured Discord users into voice channels, with allowed

// Agents: clarify that fixes should default to clean bounded refactors, lean internals, and explicit plugin SDK/API deprecation paths. Dependencies: update @openclaw/proxyline to 0.3.3.

// Agents: clarify that fixes should default to clean bounded refactors, lean internals, and explicit plugin SDK/API deprecation paths. Dependencies: update @openclaw/proxyline to 0.3.3.

// Leaner installs: WhatsApp, Slack, Amazon Bedrock, Anthropic Vertex, and related provider/plugin dependency cones moved out of the core runtime so installs only pull what you use. Telegram got much more resilient: isolated polling, durable local spooling, safer group-media handling, and preserved HTM

// - Release/plugin publishing: retry transient ClawHub CLI dependency install failures, keep preview-passing plugins publishable when one preview cell flakes, and verify every expected ClawHub package version after publish so maintenance releases are faster to recover and less likely to hide partial p

// - Doctor/OpenAI Codex: revert the 2026.5.5 doctor --fix repair that rewrote valid openai-codex/ ChatGPT/Codex OAuth routes to openai/, which could break OAuth-only GPT-5.5 setups or accidentally move users onto the OpenAI API-key route. If 2026.5.5 already changed your default model, run openclaw mo

// - Feishu: hydrate missing native topic starter thread IDs before session routing so first turns and follow-ups stay in the same topic session. Fixes 78262. Thanks @joeyzenghuan.

// Google Meet/Voice Call: make Twilio dial-in joins speak through the realtime Gemini voice bridge with paced audio streaming, backpressure-aware buffering, barge-in queue clearing, and no TwiML fallback during realtime speech, giving Meet participants a much snappier OpenClaw voice agent. (77064) Tha

// Core npm hotfix release for v2026.5.3.

// Plugins/file-transfer: add bundled file-transfer plugin with file_fetch, dir_list, dir_fetch, and file_write agent tools for binary file ops on paired nodes; default-deny per-node path policy under plugins.entries.file-transfer.config.nodes with operator approval, symlink traversal refused by defaul

// External plugin installation, update, doctor repair, dependency reporting, and artifact metadata now cover the npm-first cutover, stale configured installs, missing package payloads, and beta-channel plugin fallback. Thanks @vincentkoc. Gateway and agent hot paths are leaner across startup, session

// Memory transformed into a people-aware wiki with provenance views and per-conversation Active Memory filters. Active-run steering enabled by default. Visible-reply enforcement and spawned subagent routing metadata. NVIDIA onboarding/catalogs. Bedrock Opus 4.7 thinking parity. BREAKING: tools.exec and tools.fs no longer implicitly widen restrictive profiles — explicit alsoAllow entries required.

// Major reliability improvements for gateway and packaged-plugin startup diagnostics. Runtime-dependency repair for packaged installs. Stale-session recovery. Channel-specific fixes for Slack thread timestamps, Telegram media transport, Discord gateway metadata, WhatsApp session stability, and Microsoft Teams message formatting.

// Opt-in sandbox.docker.gpus passthrough for local GPU workloads within sandboxed agents. Authenticated node.presence.alive protocol events for iOS background wakes. Node.list last-seen fields. Expanded i18n: Persian, Dutch, Vietnamese, Italian, Arabic, and Thai locales added.

// Full QQBot group chat support with history tracking, @-mention gating, activation modes, per-group config, and FIFO message queuing. Engine refactored into pipeline stages with focused outbound submodules and explicit DI ports via createEngineAdapters(). Generic browser real-time transport for Google Live voice conversations. BREAKING: Engine adapter interface changed.

// SQLite-backed plugin state stores for restart-safe registries. Provider-specific model onboarding stabilization. Opt-in follow-up commitments for heartbeat-delivered reminders. Bounded REM preview diagnostics. Partial recall on memory timeout.

// Safer Codex/OpenAI-compatible replay and streaming behavior. Session age pruning improvements. Gateway startup diagnostics expanded. Plugin state migration tooling for SQLite backends.

// Image generation and editing through Codex OAuth — gpt-image-2 works without OPENAI_API_KEY. OpenRouter image generation support. Subagent forked context inheritance. Per-call timeoutMs for media tools. Pi packages updated to 0.70.0 with GPT-5.5 catalog.

// xAI gains image generation (grok-imagine), TTS, and STT. Voice Call streaming transcription for Deepgram, ElevenLabs, Mistral. TUI embedded mode runs without Gateway. Auto-install missing plugins during setup. OpenAI native web_search tool support.

// Default image generation provider switched to gpt-image-2. Plugin doctor repairs bundled runtime dependencies for packaged installs. Owner identity required for owner-enforced commands. Slack thread alias preservation.

// Restyled security disclaimer in setup wizard. Strengthened default system prompt and GPT-5 overlay. Tiered model pricing from cached catalogs with Moonshot Kimi support. Built-in session entry cap and age pruning enforced by default.

// Default Anthropic model updated to Claude 4.5 Sonnet. Gemini text-to-speech support added. LanceDB cloud storage for durable memory indexes. GitHub Copilot embedding provider for memory search. Control UI Model Auth status card. BREAKING: Anthropic model defaults changed.

// Forward-compat for gpt-5.4-pro with Codex pricing/limits. Telegram forum topics now surface human topic names in agent context. Ollama embedded-run timeout forwarding fix. Image and PDF tool provider/model normalization.

// New optional Active Memory plugin with dedicated memory sub-agent. Experimental local MLX speech provider for macOS Talk Mode. CLI exec-policy command for synchronizing request-level exec approval. Gateway commands.list RPC. Smoother Feishu setup path.

// ChatGPT conversation import ingestion with Imported Insights and Memory Palace diary subtabs. Video generation with URL-only asset delivery and reference audio. Feishu document comment sessions. Microsoft Teams reaction and Graph pagination support.

// Bundled Codex provider with plugin-owned app-server harness for codex/gpt-* models. Active Memory plugin for dedicated memory sub-agent. macOS MLX local speech for Talk Mode. Seedance 2.0 video model support. Teams message pin/unpin/react actions. BREAKING: Codex harness integration changes.

// Grounded REM backfill lane with historical rem-harness and diary commit/reset flows. Control UI structured diary view with timeline navigation. Character-vibes QA evaluation reports. Provider manifest authAliases for shared env vars. iOS CalVer pinning.

// Aligned packaged plugin compatibility metadata with release version. Telegram setup loads through packaged top-level sidecars. Bundled channels share secret contracts. Exec policy alignment with real runtime behavior.

// First-class openclaw infer hub for provider-backed inference workflows. Built-in music_generate tool with Google Lyria and MiniMax providers. Restored memory-wiki stack with sync/query/apply tooling. Bundled webhook ingress plugin for TaskFlows. Persisted compaction checkpoints. BREAKING: media tool auto-fallback behavior changed.

// DeepSeek V4 and V4 Flash support. GPT-5.5 forward-compat handling. Expanded image generation with provider-specific hints. Plugin SDK refactoring for lighter startup. Static model catalogs. BREAKING: Plugin SDK tool-result transform API changes.

// Bundled plugin architecture with manifest-backed model catalogs. Talk/Voice Call realtime voice loops consulting full OpenClaw agent. Browser coordinate clicks for managed automation. Matrix cross-signing identity trust. Gradium TTS provider. BREAKING: Plugin SDK registerEmbeddedExtensionFactory removed.

// Google Meet as bundled participant plugin with personal Google auth, Chrome/Twilio realtime, paired-node Chrome support. DeepSeek V4 Flash as onboarding default. Realtime voice loops for Talk and Voice Call. Lighter plugin startup with static catalogs. BREAKING: Plugin SDK tool-result transforms API changed.

// Built-in video_generate tool with fal Kling and Veo2 providers. MiniMax music generation support. Puppeteer extra stealth plugin for undetectable browser automation. Dreaming wake/sleep cycle improvements and diary formatting.

// New dreaming architecture with configurable wake/sleep cycles and diary persistence. Image understanding model routing respects agent defaults. Plugin health check endpoints. Gateway startup diagnostics. BREAKING: Dreaming config schema changed.

// First-class GPT-5.5 and Claude 4 Sonnet model support. New plugin manifest system for declarative capability contracts. Session compaction improvements with branch/restore. BREAKING: Plugin manifest format introduced.

// ModelStudio/Qwen: add standard (pay-as-you-go) DashScope endpoints for China and global Qwen API keys alongside the existing Coding Plan endpoints, and relabel the provider group to Qwen (Alibaba Cloud Model Studio). (43878) UI/clarity: consolidate button primitives (btn--icon, btn--ghost, btn--xs),

// Plugin SDK v2 with typed contracts, lifecycle hooks, and manifest-first architecture. Ollama inference performance improvements. macOS native Talk Mode integration. Feishu channel support improvements. BREAKING: Plugin SDK v2 migration required for custom plugins.

// This recovery release uses v2026.3.13-1 because GitHub immutable releases do not allow reusing v2026.3.13 after publication.

// Real Chrome browser session control for AI agents. Redesigned Android chat settings UI. Docker OPENCLAW_TZ timezone support. Compaction fixes, Telegram media transport policy improvements, and Discord gateway metadata reliability.

// Complete Control UI refresh with modular views and command palette. Configurable Fast Mode for OpenAI/Anthropic speed toggles. Starter Kubernetes manifests. 15+ critical CVEs patched. Implicit workspace plugin autoloads disabled (BREAKING).

// Memory search upgraded with multimodal image/audio indexing via gemini-embedding-2-preview. iOS home screen redesign with live agent overview. First-class Ollama browser sign-in. macOS model picker in chat. Cron/notification migration required (BREAKING).

// CLI/backup: add openclaw backup create and openclaw backup verify for local state archives, including --only-config, --no-include-workspace, manifest/payload validation, and backup guidance in destructive flows. (40163) thanks @shichangs. macOS/onboarding: add a remote gateway token field for remote

// New pluggable ContextEngine with 7 lifecycle hooks for context management. GPT-5.4 and Gemini Flash 3.1 added as default model aliases. ClawJacked vulnerability patched. Dual auth mode (token + password) requires explicit gateway.auth.mode (BREAKING).

// Critical sandbox security fixes (TOCTOU, symlink escapes, media staging). New first-class PDF analysis tool with Anthropic/Google provider support. Telegram streaming now defaults to 'partial' for all new setups.

// OpenAI Responses API is now WebSocket-first by default. Docker/Kubernetes health check endpoints added. Android gains camera.list, device.permissions, and notification actions. Node exec approval payloads now require systemRunPlan (BREAKING).

// Full openclaw secrets workflow (audit, configure, apply, reload). ACP agents are now first-class thread session runtimes. New openclaw agents bindings/bind/unbind CLI. Dozens of security fixes for gateway auth, pairing, and channel ingress.

// High-severity security patch (CVE-level): hardened gateway auth, node exec approvals, symlink sandbox escapes, and channel reaction authorization. Heartbeat DM delivery default changed back to 'allow' (BREAKING from 2026.2.24).

// Shipped native four-step Android onboarding flow with a new five-tab shell. Security hardening for sandbox Docker network namespace join (blocked by default). Heartbeat now blocks DM delivery by default (BREAKING).

// First-class Kilocode provider support (default model: kilocode/anthropic/claude-opus-4.6). Browser SSRF policy now defaults to trusted-network mode (BREAKING). New sessions cleanup tooling with disk-budget controls.

// Models/Google: add Gemini 3.1 support (google/gemini-3.1-pro-preview). Providers/Onboarding: add Volcano Engine (Doubao) and BytePlus providers/models (including coding variants), wire onboarding auth choices for interactive + non-interactive flows, and align docs to volcengine-api-key. (7967) Thank

// Native Synology Chat channel. Optional built-in auto-updater (off by default). Removed Google Antigravity provider (BREAKING). Gateway device-auth v1 removed — v2 required (BREAKING). Unified channel streaming config.

// iOS/Watch: add an Apple Watch companion MVP with watch inbox UI, watch notification relay handling, and gateway command surfaces for watch status/send flows. (20054) Thanks @mbelinky. iOS/Gateway: wake disconnected iOS nodes via APNs before nodes.invoke and auto-reconnect gateway sessions on silent

// Agents/Anthropic: add opt-in 1M context beta header support for Opus/Sonnet via model params.context1m: true (maps to anthropic-beta: context-1m-2025-08-07). Agents/Models: support Anthropic Sonnet 4.6 (anthropic/claude-sonnet-4-6) across aliases/defaults with forward-compat fallback when upstream c

// Discord: unlock rich interactive agent prompts with Components v2 (buttons, selects, modals, and attachment-backed file blocks) so for native interaction through Discord. Thanks @thewilloftheshadow. Discord: components v2 UI + embeds passthrough + exec approval UX refinements (CV2 containers, button

// Telegram: add poll sending via openclaw message poll (duration seconds, silent delivery, anonymity controls). (16209) Thanks @robbyczgw-cla. Slack/Discord: add dmPolicy + allowFrom config aliases for DM access control; legacy dm.policy + dm.allowFrom keys remain supported and openclaw doctor --fix c

// Discord: send voice messages with waveform previews from local audio files (including silent delivery). (7253) Thanks @nyanjou. Discord: add configurable presence status/activity/type/url (custom status defaults to activity text). (10855) Thanks @h0tp-ftw.

// CLI: add openclaw logs --local-time to display log timestamps in local timezone. (13818) Thanks @xialonglee. Telegram: render blockquotes as native <blockquote> tags instead of stripping them. (14608)

// iOS: alpha node app + setup-code onboarding. (11756) Thanks @mbelinky.

// Models: support Anthropic Opus 4.6 and OpenAI Codex gpt-5.3-codex (forward-compat fallbacks). (9853, 10720, 9995) Thanks @TinyTb, @calvin-hpnet, @tyler6204. Providers: add xAI (Grok) support. (9885) Thanks @grp06.

// Telegram: remove last @ts-nocheck from bot-handlers.ts, use Grammy types directly, deduplicate StickerMetadata. Zero @ts-nocheck remaining in src/telegram/. (9206) Telegram: remove @ts-nocheck from bot-message.ts, type deps via Omit<BuildTelegramMessageContextParams>, widen allMedia to TelegramMedia

// Feishu: add Feishu/Lark plugin support + docs. (7313) Thanks @jiulingyun (openclaw-cn). Web UI: add Agents dashboard for managing agent files, tools, skills, models, channels, and cron jobs.

// Docs: onboarding/install/i18n/exec-approvals/Control UI/exe.dev/cacheRetention updates + misc nav/typos. (3050, 3461, 4064, 4675, 4729, 4763, 5003, 5402, 5446, 5474, 5663, 5689, 5694, 5967, 6270, 6300, 6311, 6416, 6487, 6550, 6789) Telegram: use shared pairing store. (6127) Thanks @obviyus.

// CLI: add completion command (Zsh/Bash/PowerShell/Fish) and auto-setup during postinstall/onboarding. CLI: add per-agent models status (--agent filter). (4780) Thanks @jlowin.

// Rebrand: rename the npm package/CLI to openclaw, add a openclaw compatibility shim, and move extensions to the @openclaw/ scope. Onboarding: strengthen security warning copy for beta + access control expectations.

// Providers: Ollama discovery + docs; Venice guide upgrades + cross-links. (1606) Thanks @abhaymundhara. https://docs.clawd.bot/providers/ollama https://docs.clawd.bot/providers/venice Channels: LINE plugin (Messaging API) with rich replies + quick replies. (1630) Thanks @plum-dawg.

// TTS: move Telegram TTS into core + enable model-driven TTS tags by default for expressive audio replies. (1559) Thanks @Glucksberg. https://docs.clawd.bot/tts Gateway: add /tools/invoke HTTP endpoint for direct tool calls (auth + tool policy enforced). (1575) Thanks @vignesh07. https://docs.clawd.bo

// Highlight: Compaction safeguard now uses adaptive chunking, progressive fallback, and UI status + retries. (1466) Thanks @dlauer. Providers: add Antigravity usage tracking to status output. (1490) Thanks @patelhiren.

// Lobster optional plugin tool for typed workflows + approval gates. docs Custom assistant identity + avatars in the Control UI. docs docs

// Control UI: add copy-as-markdown with error feedback. (1345) https://docs.clawd.bot/web/control-ui Control UI: drop the legacy list view. (1345) https://docs.clawd.bot/web/control-ui

// Hooks: add hooks system with bundled hooks, CLI tooling, and docs. (1028) — thanks @ThomsenDrake. https://docs.clawd.bot/hooks Media: add inbound media understanding (image/audio/video) with provider + CLI fallbacks. https://docs.clawd.bot/nodes/media-understanding

// Plugins: add provider auth registry + clawdbot models auth login for plugin-driven OAuth/API key flows. Browser: improve remote CDP/Browserless support (auth passthrough, wss upgrade, timeouts, clearer errors).

// Web search: web_search/web_fetch tools (Brave API) + first-time setup in onboarding/configure. Browser control: Chrome extension relay takeover mode + remote browser control via clawdbot browser serve.

// Packaging: include dist/memory/ in the npm tarball (fixes ERR_MODULE_NOT_FOUND for dist/memory/index.js).

// BREAKING: rename chat “providers” (Slack/Telegram/WhatsApp/…) to channels across CLI/RPC/config; legacy config keys auto-migrate on load (and are written back as channels.). Memory: add vector search for agent memories (Markdown-only) with SQLite index, chunking, lazy sync + file watch, and per-agen

// CLI: avoid top-level await warnings in the entrypoint on fresh installs.

// Installer: ensure the CLI entrypoint is executable after npm installs.

// Installer: include patches/ in the npm package so postinstall patching works for npm/bun installs.

// Plugins are now first-class: loader + CLI management, plus the new Voice Call plugin. Config: modular $include support for split config files. (731) — thanks @pasogott.

// CLI: clawdbot status now table-based + shows OS/update/gateway/daemon/agents/sessions; status --all adds a full read-only debug report (tables, log tails, Tailscale summary, and scan progress via OSC-9 + spinner). CLI Backends: add Codex CLI fallback with resume support (text output) and JSONL parsi

// Microsoft Teams provider: polling, attachments, outbound CLI send, per-channel policy. Models/Auth expansion: OpenCode Zen + MiniMax API onboarding; token auth profiles + auth order; OAuth health in doctor/status.

// Security: DMs locked down by default across providers; pairing-first + allowlist guidance. Sandbox: per-agent scope defaults + workspace access controls; tool/session isolation tuned.

// NPM package: include missing runtime dist folders (slack/signal/imessage/tui/wizard/control-ui/daemon) to avoid ERR_MODULE_NOT_FOUND in Node 25 npx installs.

// NPM package: include dist/sessions so clawdbot agent resolves session helpers in npx installs.

// Models: add image-specific model config (agent.imageModel + fallbacks) and scan support. Agent tools: new image tool routed to the image model (when configured).

// Pluggable agents (Claude, Pi, Codex, Opencode): New inbound.reply.agent block chooses the CLI and parser per command reply; per-agent argv builders inject the right flags/identity/prompt handling and parse NDJSON streams, enabling Pi/Codex swaps without changing templates. Safety stop words for agen

// Manual heartbeat sends: warelay heartbeat accepts --message/--body with --provider web|twilio to push real outbound messages; --dry-run previews payloads without sending.

// warelay 1.2.1\n\n- Heartbeat: warelay heartbeat now supports --message/--body for manual sends on web or twilio, with --dry-run to preview payloads.\n- Media handling: MIME-first detection (magic bytes > headers > extension) so images/audio/video/doc send correctly even with wrong suffixes, and imag

// Heartbeat UX: default 10 m; prompt HEARTBEAT ultrathink; HEARTBEAT_OK suppresses sends; skipped heartbeats don’t refresh sessions; fallback heartbeats no longer start new sessions; supports heartbeatIdleMinutes.

// Web auto-replies now resize/recompress media, honor inbound.reply.mediaMaxMb, detect media kind, and enforce provider caps.

// Added cwd option to command reply config so external reply commands (e.g., Claude Code) run in the intended working directory for correct context.

// Fix commander help config for TypeScript build (subcommandTerm).

// Added executable shim so npx warelay@0.1.x runs the CLI directly.