openclacky 1.5.16 → 1.5.17

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 9939ab09ac4e473dc7b096dac6a14c67f6d6060fe7678d7bd0b2635c7cc924ed
-  data.tar.gz: 9e0e566a77267f73388a9761d101672c86fd35513dc7248355a3c3df7d8017f6
+  metadata.gz: 23761f95ba9b1fa52c2acc37020fd4e903a92d8f49c0c97668d5cfbb6e572329
+  data.tar.gz: 16838f20ba7d650e15952c738c35b93ca418a8df81baaeb6f9fd0263c5754e65
 SHA512:
-  metadata.gz: 206657c6e08b40482a5029abf25fb8bd8579728ea469e2b2f7fba5b0f95dbbbcbfb8ca75778f73090bbbb34d7145cf58dab64c7d7462325d981a4e5106ced23e
-  data.tar.gz: 0b30e40beb4d544076f2063160d04d9fcc2099b43ce46c78e1694d35ec631859361ffc15b2521a8a53e0c0cd84fb52e157cebf4469557991f273770842808568
+  metadata.gz: d0a29e5ab8a0e6b21e1f2a2f36657bcb8043cfc9d904671acdead8157f082ac032c2d1f7f0a0cee16c0f72a7fdbb939020b3abfe52145a9eea722f4f4856bac2
+  data.tar.gz: '050389221f4f71ab6e54a6038606f0d1f48f1a6571116c5b052493362158a850fe0f1a90a4d520b35fff875315fd1292610bab104366251626d5d9fdcb5b7f54'

data/CHANGELOG.md

@@ -4,6 +4,45 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 
+## [1.5.17] - 2026-09-30
+
+### Added
+- Show a top-up action when a free-trial model rejects the request
+- Add Requesty as a named provider preset (#587 - @Thibaultjaigu)
+- Redesign the skills page around cards and a detail view
+- Redesign the scheduled tasks page and keep a run history per task
+- Redesign the channels page as a compact two-column list
+- Rework the MCP page layout, its row menu and the remove confirmation
+- Add "Add to chat" to the workspace file context menu
+- Add a global switch for IM progress cards (#599)
+- Show the vendor logo on the model chip of a new session
+- Use Chinese names for the built-in skills
+
+### Improved
+- Optimize the Feishu card styling (#597)
+- Improve drag-and-drop for session references and files (#600)
+- Restyle the channel status badge, cron rows, message settings and search boxes
+- Polish the assistant memory page, onboarding page, billing header and trash toolbar
+- Improve the discoverability of overflowing aside tabs (#601)
+- Show skill, MCP and task errors as in-app toasts instead of browser alerts
+- Localize the default extension icon letter
+- Ship meeting as a default-off extension
+
+### Fixed
+- Make an IM message interrupt the running task instead of queueing silently (#586)
+- Hide the sidecar models from the model switchers (#598)
+- Show the "View process" panel in Feishu only when there is process history (#596)
+- Keep a subagent phase running when a message is queued (#594)
+- Fix the skill install click and the extension installed state
+- Strip channel-only prompt prefixes when replaying history (#585)
+- Use the in-app dialog for delete confirmations instead of the browser prompt
+- Toggle a sidebar ··· menu on a repeated click (#595)
+- Hide the auto tag and routing tooltip once the alias is pinned to a model
+
+### More
+- Retire Claude 4.5 / Opus 4.6 and Gemini 3.5 / 3.6 Flash from the catalogs
+- Remove legacy fallback domains and dead failover paths
+
 ## [1.5.16] - 2026-09-24
 
 ### Added

data/README.md

@@ -205,7 +205,7 @@ $ openclacky
 
 Set your **API Key**, **Model**, and **Base URL** (any OpenAI-compatible provider).
 
-Supported out of the box: **Claude (Anthropic) · GPT (OpenAI) · DeepSeek · Kimi (Moonshot) · MiniMax · OpenRouter · OrcaRouter** — or any custom endpoint.
+Supported out of the box: **Claude (Anthropic) · GPT (OpenAI) · DeepSeek · Kimi (Moonshot) · MiniMax · OpenRouter · OrcaRouter · Requesty** — or any custom endpoint.
 
 ## Coding use case
 

data/README_CN.md

@@ -189,7 +189,7 @@ $ openclacky
 
 设置你的 **API Key**、**模型**和 **Base URL**（任意 OpenAI 兼容提供商）。
 
-开箱即支持：**Claude (Anthropic) · GPT (OpenAI) · DeepSeek · Kimi (Moonshot) · MiniMax · OpenRouter · OrcaRouter**，或任意自定义端点。
+开箱即支持：**Claude (Anthropic) · GPT (OpenAI) · DeepSeek · Kimi (Moonshot) · MiniMax · OpenRouter · OrcaRouter · Requesty**，或任意自定义端点。
 
 ## 代码开发场景
 

data/README_JA.md

@@ -194,7 +194,7 @@ $ openclacky
 
 **API Key**、**Model**、**Base URL**（任意の OpenAI 互換プロバイダー）を設定します。
 
-標準でサポート: **Claude (Anthropic) · GPT (OpenAI) · DeepSeek · Kimi (Moonshot) · MiniMax · OpenRouter · OrcaRouter** — または任意のカスタムエンドポイント。
+標準でサポート: **Claude (Anthropic) · GPT (OpenAI) · DeepSeek · Kimi (Moonshot) · MiniMax · OpenRouter · OrcaRouter · Requesty** — または任意のカスタムエンドポイント。
 
 ## コーディングのユースケース
 

data/lib/clacky/agent/history_navigation.rb

@@ -208,7 +208,7 @@ module Clacky
 
       private def navigation_entry(round)
         msg = round[:user_msg]
-        user = navigation_preview(msg[:display_text] || extract_text_from_content(msg[:content]))
+        user = navigation_preview(msg[:display_text] || strip_channel_prompt_prefix(extract_text_from_content(msg[:content])))
         if user.empty?
           names = Array(msg[:display_files]).map { |file| file[:name] || file["name"] }
           Array(msg[:content]).each do |block|

data/lib/clacky/agent/llm_caller.rb

@@ -211,21 +211,7 @@ module Clacky
             Clacky::Shutdown.sleep(retry_delay)
             retry
           else
-            # All retries exhausted for a timeout. Try the secondary gateway URL if available.
-            if !@config.url_fallback_active?
-              fallback_url = @config.fallback_base_url_for_current_provider
-              if fallback_url
-                @config.activate_url_fallback!(fallback_url)
-                rebuild_client_for_current_model!
-                retries = 0
-                @ui&.show_warning(
-                  I18n.t("llm.warn.switching_to_fallback_url", url: fallback_url)
-                )
-                retry
-              end
-            end
-
-            raise AgentError.new("[LLM] #{I18n.t("llm.error.request_timeout", retries: max_retries)}", raw_message: e.message)
+            raise AgentError.new("#{I18n.t("llm.error.request_timeout", retries: max_retries)}", raw_message: e.message)
           end
 
         rescue Faraday::ConnectionFailed, Faraday::SSLError, Errno::ECONNREFUSED, Errno::ETIMEDOUT, Errno::EPIPE => e
@@ -259,24 +245,9 @@ module Clacky
             Clacky::Shutdown.sleep(retry_delay)
             retry
           else
-            # All retries on the primary endpoint exhausted for a network-level
-            # failure (DNS/TCP/TLS). Try the secondary gateway URL if available.
-            if !@config.url_fallback_active?
-              fallback_url = @config.fallback_base_url_for_current_provider
-              if fallback_url
-                @config.activate_url_fallback!(fallback_url)
-                rebuild_client_for_current_model!
-                retries = 0
-                @ui&.show_warning(
-                  I18n.t("llm.warn.switching_to_fallback_url", url: fallback_url)
-                )
-                retry
-              end
-            end
-
             # Don't show_error here — let the outer rescue block handle it to avoid duplicates.
             # Progress cleanup is the caller's responsibility (via its own ensure block).
-            raise AgentError.new("[LLM] #{I18n.t("llm.error.network_failed", retries: max_retries)}", raw_message: e.message)
+            raise AgentError.new("#{I18n.t("llm.error.network_failed", retries: max_retries)}", raw_message: e.message)
           end
 
         rescue RetryableError => e
@@ -318,28 +289,9 @@ module Clacky
           Clacky::Shutdown.sleep(retry_delay)
           retry
         else
-          # All retries on the current endpoint exhausted.
-          # If a secondary gateway URL is configured and not yet tried, switch
-          # to it now, rebuild the client, and reset the retry counter so the
-          # same request gets a fresh MAX_RETRIES_ON_FALLBACK budget on the
-          # new endpoint.  This is a one-shot switch — url_fallback_active?
-          # prevents a second switch if the fallback URL also fails.
-          if !@config.url_fallback_active?
-            fallback_url = @config.fallback_base_url_for_current_provider
-            if fallback_url
-              @config.activate_url_fallback!(fallback_url)
-              rebuild_client_for_current_model!
-              retries = 0
-              @ui&.show_warning(
-                I18n.t("llm.warn.switching_to_fallback_url", url: fallback_url)
-              )
-              retry
-            end
-          end
-
           # Don't show_error here — let the outer rescue block handle it to avoid duplicates.
           # Progress cleanup is the caller's responsibility (via its own ensure block).
-          raise AgentError.new("[LLM] #{I18n.t("llm.error.service_unavailable", retries: current_max)}", raw_message: e.message)
+          raise AgentError.new("#{I18n.t("llm.error.service_unavailable", retries: current_max)}", raw_message: e.message)
         end
 
         rescue Clacky::BadRequestError => e

data/lib/clacky/agent/session_serializer.rb

@@ -332,10 +332,20 @@ module Clacky
         { has_more: has_more }
       end
 
+      # ChannelManager prefixes channel prompts with agent-only context for
+      # the LLM; the clean original rides in display_text. Replay paths that
+      # fall back to raw content (pre-display_text sessions, chunk MD archives)
+      # must strip that block for UI parity with live rendering. Gated on
+      # channel_info so a normal user literally typing these lines keeps them.
+      private def strip_channel_prompt_prefix(text)
+        return text unless @channel_info
+        Clacky::Utils::ChannelPrompt.strip(text)
+      end
+
       private def replay_rounds(ui, page)
         page.each do |round|
           msg = round[:user_msg]
-          raw_text    = msg[:display_text] || extract_text_from_content(msg[:content])
+          raw_text    = msg[:display_text] || strip_channel_prompt_prefix(extract_text_from_content(msg[:content]))
           # Images: recovered from inline image_url blocks in content (carry data_url for <img> rendering)
           image_files = extract_image_files_from_content(msg[:content])
           # Disk files (PDF, doc, etc.): stored in display_files on the user message at send time
@@ -507,7 +517,7 @@ module Clacky
             synthetic_ts = base_time - ((index ? index[:section_count] : sections.size) - round_offset - round_index) * 1.0
             user_msg = {
               role: "user",
-              content: text,
+              content: strip_channel_prompt_prefix(text),
               created_at: synthetic_ts,
               ext_events: sec_ext_events,
               _from_chunk: true

data/lib/clacky/agent.rb

@@ -238,7 +238,7 @@ module Clacky
       entry = @config.current_model
       @client = Clacky::Client.new(
         @config.api_key,
-        base_url: @config.effective_base_url,
+        base_url: @config.base_url,
         model: @config.model_name,
         anthropic_format: @config.anthropic_format?,
         api_format: @config.api_format,

data/lib/clacky/agent_config.rb

@@ -523,13 +523,7 @@ module Clacky
       @current_model_id = id
       @current_model_index = index
 
-      if previous_id != id
-        @session_model_overlay = nil
-        # Reset URL fallback when switching models — the new model has its own
-        # base_url and the old fallback endpoint is irrelevant (or even wrong).
-        @url_fallback_active = false
-        @url_fallback_base_url = nil
-      end
+      @session_model_overlay = nil if previous_id != id
 
       true
     end
@@ -545,16 +539,9 @@ module Clacky
       index = @models.find_index { |m| m["model"].to_s.downcase == name_str }
       return false if index.nil?
 
-      previous_id = @current_model_id
       @current_model_id = @models[index]["id"]
       @current_model_index = index
 
-      if previous_id != @current_model_id
-        # Reset URL fallback when switching models — same rationale as switch_model_by_id.
-        @url_fallback_active = false
-        @url_fallback_base_url = nil
-      end
-
       true
     end
 
@@ -601,9 +588,27 @@ module Clacky
       true
     end
 
-    # List all model names
+    # Model entries the user can actually chat with.
+    #
+    # `type` doubles as a record discriminator: `default`/`lite` tag a chat
+    # model's role, while the media/ocr kinds mark sidecar capability configs
+    # that only the dedicated media UI should surface. Every model switcher
+    # (CLI, IM, Web UI) must go through here so the rule lives in one place.
+    def chat_models
+      @models.select { |m| self.class.chat_model?(m) }
+    end
+
+    # @param entry [Hash, nil]
+    def self.chat_model?(entry)
+      return false unless entry
+      return false if entry["auto_injected"]
+
+      !Clacky::Providers::SIDECAR_KINDS.include?(entry["type"].to_s)
+    end
+
+    # List all chat model names
     def model_names
-      @models.map { |m| m["model"] }
+      chat_models.map { |m| m["model"] }
     end
 
     # Get API key for current model
@@ -1138,63 +1143,6 @@ module Clacky
       end
     end
 
-    # ── URL-level fallback ─────────────────────────────────────────────────
-    # Independent from the model-name fallback above.  When all max_retries
-    # on the primary endpoint are exhausted, the caller may switch to a
-    # secondary gateway URL (same model, different host) via these methods.
-    # The URL fallback is intentionally one-shot per session — we do not
-    # probe or reset it automatically, keeping the logic simple.
-
-    # Look up the fallback base URL for the current model's provider.
-    # Returns nil if:
-    #   - provider cannot be determined
-    #   - provider is not openclacky (URL failover is only supported for the
-    #     OpenClacky gateway; other providers have no secondary endpoint)
-    #   - current base_url is already the fallback URL (user picked the
-    #     Secondary/China node directly — switching to the same URL does nothing)
-    # @return [String, nil]
-    def fallback_base_url_for_current_provider
-      m = current_model
-      return nil unless m
-
-      # Use strict URL matching to identify the provider — no inference.
-      # resolve_provider would incorrectly return "openclacky" for localhost
-      # addresses and clacky-* api keys, triggering unwanted failover.
-      # find_by_base_url only matches registered preset URLs, so localhost
-      # and other non-openclacky endpoints correctly return nil.
-      provider_id = Clacky::Providers.find_by_base_url(m["base_url"])
-      return nil unless provider_id == Clacky::Providers::OPENCLACKY_ID
-
-      fallback_url = Clacky::Providers.fallback_base_url(provider_id)
-      return nil unless fallback_url
-
-      # Don't offer a fallback when the current base_url is already the
-      # fallback endpoint — switching to the same URL accomplishes nothing.
-      current_base = m["base_url"].to_s.chomp("/")
-      return nil if current_base == fallback_url.to_s.chomp("/")
-
-      fallback_url
-    end
-
-    # Activate the URL fallback: override base_url with the secondary gateway.
-    # Idempotent — safe to call multiple times.
-    # @param fallback_url [String] the secondary gateway URL
-    def activate_url_fallback!(fallback_url)
-      @url_fallback_active = true
-      @url_fallback_base_url = fallback_url
-    end
-
-    # Returns true when the secondary gateway URL is in use.
-    def url_fallback_active?
-      @url_fallback_active == true
-    end
-
-    # The effective base URL for API calls.
-    # Returns the URL-fallback override when active, otherwise the model's configured base_url.
-    def effective_base_url
-      @url_fallback_active ? @url_fallback_base_url : base_url
-    end
-
     # Get current model configuration.
     #
     # Resolution order:

data/lib/clacky/client.rb

@@ -817,7 +817,7 @@ module Clacky
 
       if error_code == "insufficient_credit" || response.status == 402
         raise InsufficientCreditError.new(
-          "[LLM] #{I18n.t("llm.error.insufficient_credit")}",
+          "#{I18n.t("llm.error.insufficient_credit")}",
           error_code: "insufficient_credit",
           provider_id: @provider_id,
           raw_message: error_message
@@ -827,30 +827,37 @@ module Clacky
       case response.status
       when 400
         if error_message.match?(/ThrottlingException|unavailable|quota/i)
-          raise RetryableError.new("[LLM] #{I18n.t("llm.error.rate_limit_400")}", routed_tier: routed_tier)
+          raise RetryableError.new("#{I18n.t("llm.error.rate_limit_400")}", routed_tier: routed_tier)
         end
 
         raise BadRequestError.new(
           "[LLM] Client request error: #{error_message}",
-          display_message: "[LLM] #{I18n.t("llm.error.bad_request")}",
+          display_message: "#{I18n.t("llm.error.bad_request")}",
           raw_message: error_message
         )
       when 401
-        raise AgentError.new("[LLM] #{I18n.t("llm.error.invalid_api_key")}", raw_message: error_message)
+        raise AgentError.new("#{I18n.t("llm.error.invalid_api_key")}", raw_message: error_message)
       when 403
         i18n_key = "llm.error.403.#{error_code}"
         translated = I18n.t(i18n_key)
         translated = I18n.t("llm.error.403.default") if translated == i18n_key
-        raise AgentError.new("[LLM] #{translated}", raw_message: error_message)
+        raise AgentError.new(translated, raw_message: error_message) unless error_code == "model_not_allowed"
+
+        raise ModelNotAllowedError.new(
+          translated,
+          error_code: error_code,
+          provider_id: @provider_id,
+          raw_message: error_message
+        )
       when 404
-        raise AgentError.new("[LLM] #{I18n.t("llm.error.endpoint_not_found")}", raw_message: error_message)
+        raise AgentError.new("#{I18n.t("llm.error.endpoint_not_found")}", raw_message: error_message)
       when 429
         if error_code == "quota_exceeded"
-          raise AgentError.new("[LLM] #{I18n.t("llm.error.quota_exhausted")}", raw_message: error_message)
+          raise AgentError.new("#{I18n.t("llm.error.quota_exhausted")}", raw_message: error_message)
         end
-        raise RetryableError.new("[LLM] #{I18n.t("llm.error.rate_limit_429")}", routed_tier: routed_tier)
-      when 500..599 then raise RetryableError.new("[LLM] #{I18n.t("llm.error.server_error", status: response.status)}", routed_tier: routed_tier)
-      else raise AgentError.new("[LLM] #{I18n.t("llm.error.unexpected", status: response.status)}", raw_message: error_message)
+        raise RetryableError.new("#{I18n.t("llm.error.rate_limit_429")}", routed_tier: routed_tier)
+      when 500..599 then raise RetryableError.new("#{I18n.t("llm.error.server_error", status: response.status)}", routed_tier: routed_tier)
+      else raise AgentError.new("#{I18n.t("llm.error.unexpected", status: response.status)}", raw_message: error_message)
       end
     end
 
@@ -858,7 +865,7 @@ module Clacky
     def check_html_response(response)
       body = response.body.to_s.lstrip
       if body.start_with?("<!DOCTYPE", "<!doctype", "<html", "<HTML")
-        raise RetryableError, "[LLM] #{I18n.t("llm.error.html_response")}"
+        raise RetryableError, "#{I18n.t("llm.error.html_response")}"
       end
     end
 

data/lib/clacky/default_extensions/advisor/panels/advisor/view.js

@@ -366,7 +366,7 @@
 
   function settingsTab() {
     const btn = document.createElement("button");
-    btn.className = "settings-tab";
+    btn.className = "settings-tab page-tab";
     btn.dataset.tab = "advisor";
     btn.textContent = t("advisor.settingsTab", "工作建议");
     return btn;

data/lib/clacky/default_extensions/computer-use/skills/computer-use/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: computer-use
+name_zh: 电脑操控
+description_zh: 操控桌面上的原生应用、弹窗、菜单栏和文件选择器，处理终端和浏览器够不着的操作，支持 macOS 和 WSL 下的 Windows。
 description: Drive the desktop — macOS, or the Windows desktop through WSL — when a task needs a native app, dialog, menu bar, file picker or anything the terminal and browser tools cannot reach. Screenshot first, then click and type by coordinates read from that image.
 ---
 

data/lib/clacky/default_extensions/ext-studio/panels/studio/view.js

@@ -1643,7 +1643,11 @@
   }
 
   function _skeletonHtml() {
-    return Array.from({ length: 3 }).map(() => `
+    const bar = `
+      <div class="studio-seg-bar" style="pointer-events:none">
+        <span class="skel" style="display:inline-block;height:2rem;width:12rem;border-radius:8px"></span>
+      </div>`;
+    return bar + Array.from({ length: 3 }).map(() => `
       <div class="studio-skill-card" style="pointer-events:none">
         <div class="studio-skill-head">
           <span class="skel" style="display:inline-block;height:1rem;width:40%"></span>
@@ -2078,16 +2082,10 @@
         headerActions.appendChild(newBtn);
       }
 
-      if (locked) {
-        const lockHint = el("p", { class: "studio-empty", text: t("skills.locked") });
-        lockHint.style.marginBottom = "12px";
-        container.appendChild(lockHint);
-      }
-
       const sections = [{
         id: "local",
         label: t("skills.section.local"),
-        hint: t("skills.section.localHint"),
+        hint: locked ? t("skills.locked") : t("skills.section.localHint"),
         items: local,
         cardFn: localCard,
         emptyKey: "skills.local.empty",
@@ -2122,12 +2120,12 @@
     root.appendChild(page);
 
     const header = el("div", { class: "studio-page-head" });
-    header.appendChild(el("h2", { class: "studio-page-title", text: t("ws.title") }));
+    header.appendChild(el("h2", { class: "studio-page-title page-title", text: t("ws.title") }));
     const headerActions = el("div", { class: "studio-page-head-actions" });
     header.appendChild(headerActions);
     page.appendChild(header);
 
-    const tabsBar = el("div", { class: "studio-tabs" });
+    const tabsBar = el("div", { class: "studio-tabs page-tabs" });
     const body = el("div", { class: "studio-tab-body" });
     page.appendChild(tabsBar);
     page.appendChild(body);
@@ -2140,7 +2138,7 @@
 
     function select(id) {
       active = id;
-      Array.from(tabsBar.children).forEach((b) => b.classList.toggle("studio-tab-active", b.dataset.tab === id));
+      Array.from(tabsBar.children).forEach((b) => b.classList.toggle("active", b.dataset.tab === id));
       body.innerHTML = "";
       headerActions.innerHTML = "";
       const tab = tabs.find((t2) => t2.id === id);
@@ -2148,7 +2146,7 @@
     }
 
     tabs.forEach((tab) => {
-      const btn = el("button", { class: "studio-tab", text: tab.label });
+      const btn = el("button", { class: "studio-tab page-tab", text: tab.label });
       btn.dataset.tab = tab.id;
       btn.addEventListener("click", () => select(tab.id));
       tabsBar.appendChild(btn);
@@ -2282,16 +2280,12 @@
     .studio-published-row { display: flex; align-items: center; justify-content: space-between; gap: 8px; padding: 8px 0; font-size: 12px; border-bottom: 1px solid var(--color-border-primary); }
     .studio-published-name { display: flex; align-items: center; gap: 6px; color: var(--color-text-primary); }
     .studio-page { width: 100%; }
-    .studio-page-head { display: flex; align-items: center; justify-content: space-between; margin-bottom: 20px; }
+    .studio-page-head { display: flex; align-items: center; justify-content: space-between; min-height: 36px; margin-bottom: 20px; }
     .studio-page-head-actions { display: flex; align-items: center; gap: 8px; flex-shrink: 0; }
-    .studio-page-title { margin: 0; font-size: 22px; font-weight: 600; color: var(--color-text-primary); }
     .studio-btn-sm { padding: 5px 12px; font-size: 12px; }
-    .studio-seg-bar { display: flex; align-items: center; gap: 12px; margin-bottom: 14px; }
+    .studio-seg-bar { display: flex; align-items: center; gap: 12px; min-height: 36px; margin-bottom: 14px; }
     .studio-seg-hint { font-size: 11px; color: var(--color-text-muted); }
-    .studio-tabs { display: flex; gap: 4px; border-bottom: 1px solid var(--color-border-primary); margin-bottom: 24px; }
-    .studio-tab { padding: 8px 16px; border: none; background: transparent; color: var(--color-text-tertiary); cursor: pointer; font-size: 14px; font-weight: 500; border-bottom: 2px solid transparent; margin-bottom: -1px; }
-    .studio-tab:hover { color: var(--color-text-primary); }
-    .studio-tab-active { color: var(--color-text-primary); border-bottom-color: var(--color-accent-primary); }
+    .studio-tabs { margin-bottom: 24px; }
     .studio-ext-block { margin-bottom: 20px; }
     .studio-block-title { font-size: 13px; font-weight: 600; color: var(--color-text-secondary); padding: 8px 0 0; }
     .studio-skill-card { border: 1px solid var(--color-border-primary); border-radius: var(--radius-md, 8px); padding: 12px; margin: 0 0 10px; background: var(--color-bg-secondary); }

data/lib/clacky/default_extensions/ext-studio/skills/ext-develop/SKILL.md

@@ -3,6 +3,8 @@
 # After context compression, previously inspected host files may be
 # mistaken for extension edit targets.
 name: ext-develop
+name_zh: 扩展开发
+description_zh: 开发、调试和发布扩展：从想法搭出新扩展，修复加载失败或不显示的问题，或发布到扩展市场。
 description: Build, debug, or publish an OpenClacky extension — scaffold a new one from an idea, fix a broken/invisible panel/api/skill/agent, or ship it to the marketplace. Trigger on create/start extension, plugin, panel, ext verify error, "won't load", "not showing up", publish/ship/unpublish an extension.
 agent: ext-developer
 ---

data/lib/clacky/default_extensions/meeting/ext.yml

@@ -8,6 +8,7 @@ version: "0.1.0"
 author: OpenClacky
 homepage: https://www.openclacky.com
 license: MIT
+enabled_by_default: false
 contributes:
   api: api/handler.rb
   panels:

data/lib/clacky/default_extensions/meeting/skills/meeting-summarizer/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: meeting-summarizer
+name_zh: 会议纪要
+description_zh: 根据会议转写内容生成结构化纪要，包含关键决策、待办事项和讨论要点，会议结束后自动触发。
 description: Summarize a completed meeting from its transcript. Produces a structured summary with key decisions, action items, and discussion highlights. Triggered automatically when a meeting ends.
 user-invocable: false
 auto_summarize: false

data/lib/clacky/default_skills/browser-setup/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: browser-setup
+name_zh: 浏览器配置
+description_zh: 配置 Chrome 或 Edge 浏览器工具并检测连接，支持 macOS、Linux 和 WSL。
 description: |
   Configure the browser tool for Clacky. Guides the user through Chrome or Edge setup,
   verifies the connection, and writes ~/.clacky/browser.yml.

data/lib/clacky/default_skills/channel-manager/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: channel-manager
+name_zh: 渠道管理
+description_zh: 接入飞书、企业微信、微信、钉钉、Discord、Telegram 等 IM 渠道，支持配置、启停、诊断和发消息。
 description: |
   Configure IM platform channels (Feishu, WeCom, Weixin, Discord, Telegram, DingTalk) for openclacky.
   Uses browser automation for navigation; guides the user to paste credentials and perform UI steps.

data/lib/clacky/default_skills/channel-manager/install_feishu_skills.rb

@@ -23,8 +23,7 @@ require_relative '../skill-add/scripts/install_from_zip'
 
 class FeishuSkillsInstaller
   PRIMARY_HOST     = ENV.fetch('CLACKY_LICENSE_SERVER', 'https://www.openclacky.com')
-  FALLBACK_HOST    = 'https://openclacky.up.railway.app'
-  API_HOSTS        = ENV['CLACKY_LICENSE_SERVER'] ? [PRIMARY_HOST] : [PRIMARY_HOST, FALLBACK_HOST]
+  API_HOSTS        = [PRIMARY_HOST]
   API_PATH         = '/api/v1/skills/feishu'
   API_OPEN_TIMEOUT = 5
   API_READ_TIMEOUT = 10

data/lib/clacky/default_skills/code-explorer/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: code-explorer
+name_zh: 代码探索
+description_zh: 分析项目结构，梳理代码逻辑，回答「这个功能是怎么实现的」。
 description: Use this skill when exploring, analyzing, or understanding project/code structure. Required for tasks like "analyze project", "explore codebase", "understand how X works".
 agent: coding
 fork_agent: true

data/lib/clacky/default_skills/cron-task-creator/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: cron-task-creator
+name_zh: 定时任务
+description_zh: 创建和管理定时自动任务，支持按天、周、小时执行，可随时编辑、启停或删除。
 description: 'Create, manage, and run scheduled automated tasks (cron jobs) in Clacky. Use this skill whenever the user wants to create a new automated task or cron job, set up recurring automation, schedule something to run daily/weekly/hourly, view all scheduled tasks, edit an existing task prompt or cron schedule, enable or disable a task, delete a task, check task run history or logs, or run a task immediately via the WebUI. Trigger on phrases like cron, scheduled task, run every day, automate this; 定时任务, 每天自动, 定时执行.'
 disable-model-invocation: false
 user-invocable: true

data/lib/clacky/default_skills/deploy/SKILL.md

@@ -1,6 +1,8 @@
 ---
 name: deploy
-description: Deploy Rails applications to Railway. Handles first-time setup and re-deploys idempotently using Railway CLI. Trigger on: "deploy", "deploy to railway", "railway deploy", "发布", "部署", "上线".
+name_zh: 一键部署
+description_zh: 把 Rails 应用部署到 Railway，首次配置和后续更新都能自动完成。
+description: 'Deploy Rails applications to Railway. Handles first-time setup and re-deploys idempotently using Railway CLI. Trigger on: "deploy", "deploy to railway", "railway deploy", "发布", "部署", "上线".'
 agent: coding
 disable-model-invocation: false
 ---

data/lib/clacky/default_skills/mcp-manager/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: mcp-manager
+name_zh: MCP 管理
+description_zh: 添加、查看、检测、删除和重新配置 MCP 服务，无需手动编辑 JSON。
 description: |
   Manage MCP (Model Context Protocol) servers for openclacky: add, list, probe, remove,
   reconfigure. Edits ~/.clacky/mcp.json so the user never writes JSON by hand.

data/lib/clacky/default_skills/media-gen/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: media-gen
+name_zh: 媒体生成
+description_zh: 生成或编辑图片、视频，以及文字转语音，适用于封面、海报、插画、配音等。
 description: 'Generate or edit images, videos, or audio (text-to-speech) in the current task. Use whenever the user asks to create/generate/produce or edit/modify a picture / image / illustration / cover / poster / icon / artwork, a video / clip / animation, or speech / voiceover / narration / TTS — e.g. generate image, draw, design a cover, edit this image, change the background, text-to-video, generate speech; 画一张, 配图, 编辑图片, 改图, 换背景, 做个视频, 配音, 文字转语音. Also use when a document (slides, poster, README hero) needs an inline image.'
 disable-model-invocation: false
 user-invocable: true

data/lib/clacky/default_skills/new/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: new
+name_zh: 新建项目
+description_zh: 快速创建一个新项目，马上开始开发。
 description: Create a new project to start development quickly
 agent: coding
 disable-model-invocation: false

data/lib/clacky/default_skills/onboard/SKILL.md

@@ -1,5 +1,7 @@
 ---
 name: onboard
+name_zh: 初始设置
+description_zh: 首次使用引导：设置助手名字、性格和你的个人档案，也可单独调整其中一项。
 description: Onboard a new user OR curate a single piece of the assistant's inner state. Without arguments, runs the full first-run ceremony (AI name, personality, user profile, SOUL.md + USER.md, optional browser + personal website). With `scope:soul` or `scope:user`, runs a quick chat to update just that one profile file. With `path:<abs>`, runs a quick chat to update / keep / delete one memory file under ~/.clacky/memories/.
 disable-model-invocation: true
 user-invocable: true