openclacky 0.9.10 → 0.9.11

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 938c620eb9886a37c1e190c5bf2d8c804a6ed86eb496332a26f5f225e832a5ec
-  data.tar.gz: c165097e40b1bde5339e9819f7ccdd2a3191ecd24f9233a758a4a2406f5d74db
+  metadata.gz: 5db22957073a294578efa60af05e69277891f976bb24d6d9397d5abbefeaf6b8
+  data.tar.gz: 3103ec5cf881eb0ebe16f79a5f136c01ef890255ebeda9757d4aa5f592071470
 SHA512:
-  metadata.gz: 845fef0c86bca42adcb00d2fe34946acd3b72b2d750412a744e78a080cef34b653457ae98318ddad5b623f22325b654ad2a77a61d39abc544edbfd11f44891f3
-  data.tar.gz: ea770a8ac65695d64affd358b4af0788cccc2c4e70eb95d0afaebea61a639ba8829dbcf61af84d0d8229dac9b4af8e969316a5e5c8fb8ec75f9f430647f9b1cc
+  metadata.gz: 55fe30f89017da76ac2a9348fea1a1f4d701006d4c882878e09e915600ac80dcf72131d546b7e786c3588c1c848e47db4628478ce018e619a2d7781bd0cdbfe3
+  data.tar.gz: a6923ec51d5a5d60c2a5bd07b7fa269a76087456ea5adf4d7b33a9dc1c8a43f043570a6d4c6f6309b24dc2ad60850dd4d993cc9444fd22c76445e3749952ca09

data/.clacky/skills/gem-release/SKILL.md

@@ -105,7 +105,7 @@ To use this skill, simply say:
    - Parse the CHANGELOG.md section for `[{version}]`
    - Write it to a temp file (e.g., `/tmp/release_notes_{version}.md`) to avoid shell escaping issues
    - Run `gh release create` with `--notes-file`
-   - Verify the release appears at: `https://github.com/clacky-ai/open-clacky/releases`
+   - Verify the release appears at: `https://github.com/clacky-ai/openclacky/releases`
 
    > **Prerequisite**: `gh` CLI must be installed (`brew install gh`) and authenticated (`gh auth login`)
 
@@ -238,14 +238,14 @@ Present a clear, user-facing release summary after all steps complete:
 
 🔗 Links:
 - RubyGems: https://rubygems.org/gems/openclacky/versions/{version}
-- GitHub Release: https://github.com/clacky-ai/open-clacky/releases/tag/v{version}
+- GitHub Release: https://github.com/clacky-ai/openclacky/releases/tag/v{version}
 
 ⬆️ Upgrade:
 - In the Clacky UI, click "Upgrade" in the bottom-left → detect new version → click upgrade → done
 - Manual upgrade (CLI): `gem update openclacky`
 
 🆕 Fresh install:
-/bin/bash -c "$(curl -sSL https://raw.githubusercontent.com/clacky-ai/open-clacky/main/scripts/install.sh)"
+/bin/bash -c "$(curl -sSL https://raw.githubusercontent.com/clacky-ai/openclacky/main/scripts/install.sh)"
 ```
 
 **Rules for writing the summary:**
@@ -308,7 +308,7 @@ gh release create vX.Y.Z \
 - New version successfully published to RubyGems
 - Git repository updated with version tag
 - CHANGELOG.md updated with release notes
-- GitHub Release created and visible at https://github.com/clacky-ai/open-clacky/releases
+- GitHub Release created and visible at https://github.com/clacky-ai/openclacky/releases
 - No build or deployment errors
 - User-facing release summary presented at the end
 

data/CHANGELOG.md

@@ -7,6 +7,29 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [0.9.11] - 2026-03-25
+
+### Added
+- **Network-aware installer mirrors**: the install script now automatically detects whether you're in China and picks the fastest mirror (RubyGems China mirror, GitHub, etc.) — no manual configuration needed
+- **Shell rc-file loading**: the shell tool now sources your `.zshrc` / `.bashrc` so commands that depend on environment variables or aliases set in your shell profile work correctly
+
+### Improved
+- **Browser tool `evaluate` targets active page**: JavaScript evaluation now automatically targets the currently active browser tab instead of the last opened one, so `evaluate` always runs in the right context
+- **Browser MCP process cleaned up on server shutdown**: the `chrome-devtools-mcp` node process is now stopped when the server shuts down, preventing orphaned processes that held onto port 7070
+- **Server worker process isolation**: workers are now spawned in their own process group, ensuring grandchild processes (e.g. browser MCP) are fully cleaned up during zero-downtime restarts
+- **Channel status via live API**: `channel status` now queries the running server API instead of reading `~/.clacky/channels.yml` directly, so it reflects the actual runtime state
+- **Idle compression timer race fix**: the compression thread is now registered inside a mutex before starting, eliminating a race where `cancel()` could miss an in-flight compression and leave history in an inconsistent state
+- **Compression token display accuracy**: the post-compression token count now uses the rebuilt history estimate instead of the stale pre-compression API count
+- **Shell process group signals**: `SIGTERM`/`SIGKILL` are now sent to the entire process group (`-pgid`) instead of just the child PID, ensuring backgrounded subprocesses are also killed on timeout
+
+### Fixed
+- **Task error session save**: sessions are now correctly saved to disk even when a task ends with an error, preventing session loss on agent failures
+- **History load and model load bugs**: fixed crashes when loading sessions with missing or malformed history/model fields
+- **Default model updated to Claude claude-sonnet-4-6**: bumped the default Gemini model reference from `gemini-2.5-flash` → `gemini-2.7-flash`
+
+### More
+- Renamed gem references from `open-clacky` to `openclacky` across docs, gemspec, and scripts
+
 ## [0.9.10] - 2026-03-24
 
 ### Added

data/README.md

@@ -1,6 +1,6 @@
 # OpenClacky
 
-[![Build](https://img.shields.io/github/actions/workflow/status/clacky-ai/open-clacky/main.yml?label=build&style=flat-square)](https://github.com/clacky-ai/open-clacky/actions)
+[![Build](https://img.shields.io/github/actions/workflow/status/clacky-ai/openclacky/main.yml?label=build&style=flat-square)](https://github.com/clacky-ai/openclacky/actions)
 [![Release](https://img.shields.io/gem/v/openclacky?label=release&style=flat-square&color=blue)](https://rubygems.org/gems/openclacky)
 [![Ruby](https://img.shields.io/badge/ruby-%3E%3D%203.1.0-red?style=flat-square)](https://www.ruby-lang.org)
 [![Downloads](https://img.shields.io/gem/dt/openclacky?label=downloads&style=flat-square&color=brightgreen)](https://rubygems.org/gems/openclacky)
@@ -73,7 +73,7 @@ Built on a production-ready Rails architecture with one-click deployment, dev/pr
 ### Method 1: One-line Install (Recommended)
 
 ```bash
-/bin/bash -c "$(curl -sSL https://raw.githubusercontent.com/clacky-ai/open-clacky/main/scripts/install.sh)"
+/bin/bash -c "$(curl -sSL https://raw.githubusercontent.com/clacky-ai/openclacky/main/scripts/install.sh)"
 ```
 
 ### Method 2: RubyGems
@@ -119,15 +119,15 @@ You'll be prompted to set your **API Key**, **Model**, and **Base URL** (any Ope
 ## Install from Source
 
 ```bash
-git clone https://github.com/clacky-ai/open-clacky.git
-cd open-clacky
+git clone https://github.com/clacky-ai/openclacky.git
+cd openclacky
 bundle install
 bin/clacky
 ```
 
 ## Contributing
 
-Bug reports and pull requests are welcome on GitHub at https://github.com/clacky-ai/open-clacky. Contributors are expected to adhere to the [code of conduct](https://github.com/clacky-ai/open-clacky/blob/main/CODE_OF_CONDUCT.md).
+Bug reports and pull requests are welcome on GitHub at https://github.com/clacky-ai/openclacky. Contributors are expected to adhere to the [code of conduct](https://github.com/clacky-ai/openclacky/blob/main/CODE_OF_CONDUCT.md).
 
 ## License
 

data/clacky-legacy/clacky.gemspec

@@ -8,13 +8,13 @@ Gem::Specification.new do |spec|
 
   spec.summary = "Legacy name for openclacky gem"
   spec.description = "This is a transitional gem that depends on openclacky. The clacky project has been renamed to openclacky. Installing this gem will automatically install openclacky."
-  spec.homepage = "https://github.com/clacky-ai/open-clacky"
+  spec.homepage = "https://github.com/clacky-ai/openclacky"
   spec.license = "MIT"
   spec.required_ruby_version = ">= 3.1.0"
 
   spec.metadata["homepage_uri"] = spec.homepage
-  spec.metadata["source_code_uri"] = "https://github.com/clacky-ai/open-clacky"
-  spec.metadata["changelog_uri"] = "https://github.com/clacky-ai/open-clacky/blob/main/CHANGELOG.md"
+  spec.metadata["source_code_uri"] = "https://github.com/clacky-ai/openclacky"
+  spec.metadata["changelog_uri"] = "https://github.com/clacky-ai/openclacky/blob/main/CHANGELOG.md"
 
   spec.files = Dir["lib/**/*", "bin/*", "README.md", "LICENSE.txt"]
   spec.require_paths = ["lib"]

data/clacky-legacy/clarky.gemspec

@@ -8,13 +8,13 @@ Gem::Specification.new do |spec|
 
   spec.summary = "Legacy name for openclacky - AI agent command-line interface"
   spec.description = "This is a placeholder gem. Installing 'clarky' will automatically install 'openclacky'. The clarky command is maintained for backward compatibility."
-  spec.homepage = "https://github.com/clacky-ai/open-clacky"
+  spec.homepage = "https://github.com/clacky-ai/openclacky"
   spec.license = "MIT"
   spec.required_ruby_version = ">= 3.1.0"
 
   spec.metadata["homepage_uri"] = spec.homepage
-  spec.metadata["source_code_uri"] = "https://github.com/clacky-ai/open-clacky"
-  spec.metadata["changelog_uri"] = "https://github.com/clacky-ai/open-clacky/blob/main/CHANGELOG.md"
+  spec.metadata["source_code_uri"] = "https://github.com/clacky-ai/openclacky"
+  spec.metadata["changelog_uri"] = "https://github.com/clacky-ai/openclacky/blob/main/CHANGELOG.md"
 
   spec.files = Dir["lib/**/*", "bin/*", "README.md", "LICENSE.txt"]
   spec.require_paths = ["lib"]

data/docs/HOW-TO-USE-CN.md

@@ -91,6 +91,6 @@ Clacky 可以自动执行复杂任务，内置多种工具：
 
 ## 了解更多
 
-- GitHub：https://github.com/clacky-ai/open-clacky
-- 问题反馈：https://github.com/clacky-ai/open-clacky/issues
+- GitHub：https://github.com/clacky-ai/openclacky
+- 问题反馈：https://github.com/clacky-ai/openclacky/issues
 - 当前版本：0.7.0

data/docs/HOW-TO-USE.md

@@ -89,6 +89,6 @@ Create your own skills in `.clacky/skills/` directory!
 
 ## Learn More
 
-- GitHub: https://github.com/clacky-ai/open-clacky
-- Report Issues: https://github.com/clacky-ai/open-clacky/issues
+- GitHub: https://github.com/clacky-ai/openclacky
+- Report Issues: https://github.com/clacky-ai/openclacky/issues
 - Version: 0.7.0

data/docs/why-developer.md

@@ -285,7 +285,7 @@ models:
 
 ```bash
 # One-line installation (macOS/Linux)
-curl -sSL https://raw.githubusercontent.com/clacky-ai/open-clacky/main/scripts/install.sh | bash
+curl -sSL https://raw.githubusercontent.com/clacky-ai/openclacky/main/scripts/install.sh | bash
 
 # Or via Ruby gem
 gem install openclacky
@@ -362,7 +362,7 @@ clacky tools
 
 ## Get Started
 
-- **GitHub**: https://github.com/clacky-ai/open-clacky
+- **GitHub**: https://github.com/clacky-ai/openclacky
 - **Documentation**: https://docs.clacky.ai
 - **Discord**: https://discord.gg/clacky
 

data/docs/why-openclacky.md

@@ -211,7 +211,7 @@ We believe AI development tools should be accessible to everyone.
 
 ```bash
 # One-line installation (macOS/Linux)
-curl -sSL https://raw.githubusercontent.com/clacky-ai/open-clacky/main/scripts/install.sh | bash
+curl -sSL https://raw.githubusercontent.com/clacky-ai/openclacky/main/scripts/install.sh | bash
 
 # Or if you have Ruby 3.1+
 gem install openclacky
@@ -239,7 +239,7 @@ clacky tools
 
 Clacky is an open-source project. We welcome contributions!
 
-- **GitHub**: https://github.com/clacky-ai/open-clacky
+- **GitHub**: https://github.com/clacky-ai/openclacky
 - **Discord**: https://discord.gg/clacky
 - **Twitter**: https://twitter.com/clacky_ai
 

data/homebrew/openclacky.rb

@@ -2,7 +2,7 @@
 
 class Openclacky < Formula
   desc "Command-line interface for AI models with autonomous agent capabilities"
-  homepage "https://github.com/clacky-ai/open-clacky"
+  homepage "https://github.com/clacky-ai/openclacky"
   url "https://rubygems.org/downloads/openclacky-0.6.1.gem"
   sha256 "" # Will be updated when gem is published
   license "MIT"

data/lib/clacky/agent/cost_tracker.rb

@@ -82,71 +82,7 @@ module Clacky
 
       # Estimate token count for a message content
       # Simple approximation: characters / 4 (English text)
-      # For Chinese/other languages, characters / 2 is more accurate
-      # This is a rough estimate for compression triggering purposes
-      # @param content [String, Array, Object] Message content
-      # @return [Integer] Estimated token count
-      def estimate_tokens(content)
-        return 0 if content.nil?
-
-        text = if content.is_a?(String)
-                 content
-               elsif content.is_a?(Array)
-                 # Handle content arrays (e.g., with images)
-                 # Add safety check to prevent nil.compact error
-                 mapped = content.map { |c| c[:text] if c.is_a?(Hash) }
-                 (mapped || []).compact.join
-               else
-                 content.to_s
-               end
-
-        return 0 if text.empty?
-
-        # Detect language mix - count non-ASCII characters
-        ascii_count = text.bytes.count { |b| b < 128 }
-        total_bytes = text.bytes.length
-
-        # Mix ratio (1.0 = all English, 0.5 = all Chinese)
-        mix_ratio = total_bytes > 0 ? ascii_count.to_f / total_bytes : 1.0
-
-        # English: ~4 chars/token, Chinese: ~2 chars/token
-        base_chars_per_token = mix_ratio * 4 + (1 - mix_ratio) * 2
-
-        (text.length / base_chars_per_token).to_i + 50 # Add overhead for message structure
-      end
-
-      # Calculate total token count for all messages
-      # Returns estimated tokens and breakdown by category
-      # @return [Hash] Token counts by role and total
-      def total_message_tokens
-        system_tokens = 0
-        user_tokens = 0
-        assistant_tokens = 0
-        tool_tokens = 0
-        summary_tokens = 0
-
-        @history.to_a.each do |msg|
-          tokens = estimate_tokens(msg[:content])
-          case msg[:role]
-          when "system"
-            system_tokens += tokens
-          when "user"
-            user_tokens += tokens
-          when "assistant"
-            assistant_tokens += tokens
-          when "tool"
-            tool_tokens += tokens
-          end
-        end
 
-        {
-          total: system_tokens + user_tokens + assistant_tokens + tool_tokens,
-          system: system_tokens,
-          user: user_tokens,
-          assistant: assistant_tokens,
-          tool: tool_tokens
-        }
-      end
 
       private
 

data/lib/clacky/agent/message_compressor_helper.rb

@@ -50,8 +50,8 @@ module Clacky
         # Check if compression is enabled
         return nil unless @config.enable_compression
 
-        # Calculate total tokens and message count
-        total_tokens = total_message_tokens[:total]
+        # Use actual API-reported tokens from last request
+        total_tokens = @previous_total_tokens
         message_count = @history.size
 
         # Force compression (for idle compression) - use lower threshold
@@ -144,11 +144,9 @@ module Clacky
           chunk_path: chunk_path
         }
 
-        final_tokens = total_message_tokens[:total]
-
-        # Show compression info
+        # Show compression info (use estimated tokens from rebuilt history)
         @ui&.show_info(
-          "History compressed (~#{compression_context[:original_token_count]} -> ~#{final_tokens} tokens, " \
+          "History compressed (~#{compression_context[:original_token_count]} -> ~#{@history.estimate_tokens} tokens, " \
           "level #{compression_context[:compression_level]})"
         )
       end

data/lib/clacky/agent/session_serializer.rb

@@ -36,22 +36,17 @@ module Clacky
         @current_task_id = session_data.dig(:time_machine, :current_task_id) || 0
         @active_task_id = session_data.dig(:time_machine, :active_task_id) || 0
 
-        # Check if the session ended with an error
+        # Check if the session ended with an error.
+        # We record the rollback intent here but do NOT truncate history immediately —
+        # truncating at restore time causes the history replay to return empty results,
+        # leaving the chat panel blank on first open.
+        # Instead, the rollback is deferred: history is trimmed lazily when the user
+        # actually sends the next message (see run() / handle_user_message).
         last_status = session_data.dig(:stats, :last_status)
         last_error = session_data.dig(:stats, :last_error)
 
         if last_status == "error" && last_error
-          # Trim back to just before the last real user message that caused the error
-          last_user_index = @history.last_real_user_index
-          if last_user_index
-            @history.truncate_from(last_user_index)
-
-            @hooks.trigger(:session_rollback, {
-              reason: "Previous session ended with error",
-              error_message: last_error,
-              rolled_back_message_index: last_user_index
-            })
-          end
+          @pending_error_rollback = true
         end
 
         # Rebuild and refresh the system prompt so any newly installed skills

data/lib/clacky/agent.rb

@@ -72,6 +72,7 @@ module Clacky
       @debug_logs = []  # Debug logs for troubleshooting
       @pending_injections = []     # Pending inline skill injections to flush after observe()
       @pending_script_tmpdirs = [] # Decrypted-script tmpdirs to shred when agent.run completes
+      @pending_error_rollback = false  # Deferred rollback flag set by restore_session on error
 
       # Compression tracking
       @compression_level = 0  # Tracks how many times we've compressed (for progressive summarization)
@@ -180,6 +181,22 @@ module Clacky
         cache_hit_requests: 0
       }
 
+      # Deferred error rollback: if the previous session ended with an error,
+      # trim history back to just before that failed user message now — at the
+      # point the user actually sends a new message, not at restore time.
+      # (Trimming at restore time caused replay_history to return empty results.)
+      if @pending_error_rollback
+        @pending_error_rollback = false
+        last_user_index = @history.last_real_user_index
+        if last_user_index
+          @history.truncate_from(last_user_index)
+          @hooks.trigger(:session_rollback, {
+            reason: "Previous session ended with error — rolling back before new message",
+            rolled_back_message_index: last_user_index
+          })
+        end
+      end
+
       # Add system prompt as the first message if this is the first run
       if @history.empty?
         system_prompt = build_system_prompt

data/lib/clacky/cli.rb

@@ -580,8 +580,10 @@ module Clacky
         session_manager&.save(agent.to_session_data(status: :success))
         json_ui.update_sessionbar(tasks: agent.total_tasks, cost: agent.total_cost)
       rescue Clacky::AgentInterrupted
+        session_manager&.save(agent.to_session_data(status: :interrupted))
         json_ui.emit("interrupted")
       rescue => e
+        session_manager&.save(agent.to_session_data(status: :error, error_message: e.message))
         json_ui.emit("error", message: e.message)
       ensure
         json_ui.set_idle_status

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

@@ -1,7 +1,7 @@
 ---
 name: channel-setup
 description: |
-  Configure IM platform channels (Feishu, WeCom, Weixin) for open-clacky.
+  Configure IM platform channels (Feishu, WeCom, Weixin) for openclacky.
   Uses browser automation for navigation; guides the user to paste credentials and perform UI steps.
   Trigger on: "channel setup", "setup feishu", "setup wecom", "setup weixin", "setup wechat", "channel config",
   "channel status", "channel enable", "channel disable", "channel reconfigure", "channel doctor".
@@ -19,7 +19,7 @@ allowed-tools:
 
 # Channel Setup Skill
 
-Configure IM platform channels for open-clacky. Config is stored at `~/.clacky/channels.yml`.
+Configure IM platform channels for openclacky.
 
 ---
 
@@ -38,21 +38,38 @@ Configure IM platform channels for open-clacky. Config is stored at `~/.clacky/c
 
 ## `status`
 
-Read `~/.clacky/channels.yml` and display:
+Call the server API:
+
+```bash
+curl -s http://${CLACKY_SERVER_HOST}:${CLACKY_SERVER_PORT}/api/channels
+```
+
+Response shape (example):
+```json
+{"channels":[
+  {"platform":"feishu","enabled":true,"running":true,"has_config":true,"app_id":"cli_xxx","domain":"https://open.feishu.cn","allowed_users":[]},
+  {"platform":"wecom","enabled":false,"running":false,"has_config":false,"bot_id":""},
+  {"platform":"weixin","enabled":true,"running":true,"has_config":true,"has_token":true,"base_url":"https://ilinkai.weixin.qq.com","allowed_users":[]}
+]}
+```
+
+Display the result:
 
 ```
 Channel Status
 ─────────────────────────────────────────────────────
-Platform   Enabled   Details
-feishu     ✅ yes    app_id: cli_xxx...  domain: feishu.cn
-wecom      ❌ no     (not configured)
-weixin     ✅ yes    2 account(s) logged in
+Platform   Enabled   Running   Details
+feishu     ✅ yes    ✅ yes    app_id: cli_xxx...
+wecom      ❌ no     ❌ no     (not configured)
+weixin     ✅ yes    ✅ yes    has_token: true
 ─────────────────────────────────────────────────────
 ```
 
-For Weixin, show `has_token: true/false` from the channels.yml entry (token is never displayed).
+- Feishu: show `app_id` (truncated to 12 chars)
+- WeCom: show `bot_id` if present
+- Weixin: show `has_token: true/false` (token value is never displayed)
 
-If the file doesn't exist: "No channels configured yet. Run `/channel-setup setup` to get started."
+If the API is unreachable or returns an empty list: "No channels configured yet. Run `/channel-setup setup` to get started."
 
 ---
 
@@ -101,7 +118,7 @@ Only reach here if the automated script failed.
 
 ##### Phase 2 — Create a new app
 
-4. **Always create a new app** — do NOT reuse existing apps. Guide the user: "Click 'Create Enterprise Self-Built App', fill in name (e.g. Open Clacky) and description (e.g. AI assistant powered by open-clacky), then submit. Reply done." Wait for "done".
+4. **Always create a new app** — do NOT reuse existing apps. Guide the user: "Click 'Create Enterprise Self-Built App', fill in name (e.g. Open Clacky) and description (e.g. AI assistant powered by openclacky), then submit. Reply done." Wait for "done".
 
 ##### Phase 3 — Enable Bot capability
 
@@ -140,6 +157,7 @@ Only reach here if the automated script failed.
       -H "Content-Type: application/json" \
       -d '{"app_id":"<APP_ID>","app_secret":"<APP_SECRET>","domain":"https://open.feishu.cn"}'
     ```
+    **CRITICAL: This curl call is the ONLY way to save credentials. NEVER write `~/.clacky/channels.yml` or any file under `~/.clacky/channels/` directly. The server API handles persistence and hot-reload.**
 11. **Wait for connection** — Poll until log shows `[feishu-ws] WebSocket connected ✅`:
     ```bash
     for i in $(seq 1 20); do
@@ -238,29 +256,39 @@ Tell the user while waiting:
 
 ## `enable`
 
-Read `~/.clacky/channels.yml`, set `channels.<platform>.enabled: true`, write back.
+Call the server API to re-enable the platform (this reads from disk, sets enabled, saves, and hot-reloads):
 
-If the platform has no credentials, redirect to `setup`.
+```bash
+curl -s -X POST http://${CLACKY_SERVER_HOST}:${CLACKY_SERVER_PORT}/api/channels/<platform> \
+  -H "Content-Type: application/json" \
+  -d '{"enabled": true}'
+```
+
+If the platform has no credentials (404 or error), redirect to `setup`.
 
-Say: "✅ `<platform>` channel enabled. Restart `clacky server` to activate."
+Say: "✅ `<platform>` channel enabled."
 
 ---
 
 ## `disable`
 
-Read `~/.clacky/channels.yml`, set `channels.<platform>.enabled: false`, write back.
+Call the server API to disable the platform:
+
+```bash
+curl -s -X DELETE http://${CLACKY_SERVER_HOST}:${CLACKY_SERVER_PORT}/api/channels/<platform>
+```
 
-Say: "❌ `<platform>` channel disabled. Restart `clacky server` to deactivate."
+Say: "❌ `<platform>` channel disabled."
 
 ---
 
 ## `reconfigure`
 
-1. Show current config (mask secrets).
+1. Show current config via `GET /api/channels` (mask secrets — show last 4 chars only).
 2. Ask: update credentials / change allowed users / add a new platform / enable or disable a platform.
-3. For credential updates, re-run the relevant setup flow.
-4. Write atomically: write to `~/.clacky/channels.yml.tmp` then rename to `~/.clacky/channels.yml`.
-5. Say: "Restart `clacky server` to apply changes."
+3. For credential updates, re-run the relevant setup flow (which calls `POST /api/channels/<platform>`).
+4. **NEVER write `~/.clacky/channels.yml` directly** — always use the server API.
+5. Say: "Channel reconfigured."
 
 ---
 

data/lib/clacky/default_skills/channel-setup/feishu_setup.rb

@@ -121,33 +121,24 @@ end
 # ---------------------------------------------------------------------------
 
 class BrowserSession
-  attr_reader :target_id
-
   def initialize(client)
-    @client    = client
-    @target_id = nil
+    @client = client
   end
 
   def navigate(url)
-    if @target_id
-      @client.call("navigate", target_id: @target_id, url: url)
-    else
-      result = @client.call("open", url: url)
-      @target_id = result["targetId"]
-    end
+    @client.call("open", url: url)
     sleep 2
     snapshot
   end
 
   def snapshot(interactive: true, compact: true)
-    result = @client.call("snapshot",
-      target_id: @target_id, interactive: interactive, compact: compact)
+    result = @client.call("snapshot", interactive: interactive, compact: compact)
     result["output"].to_s
   end
 
   # Run JavaScript in the page context and return the raw output string.
   def evaluate(js)
-    result = @client.call("act", target_id: @target_id, kind: "evaluate", js: js)
+    result = @client.call("act", kind: "evaluate", js: js)
     result["output"].to_s
   end
 
@@ -155,7 +146,6 @@ class BrowserSession
   # The browser evaluate output may be wrapped in MCP markdown — parse it out.
   def cookies
     result = @client.call("act",
-      target_id: @target_id,
       kind: "evaluate",
       js: "document.cookie")
     raw = result["output"].to_s
@@ -173,7 +163,6 @@ class BrowserSession
 
     # Try lark_oapi_csrf_token specifically via JS
     result = @client.call("act",
-      target_id: @target_id,
       kind: "evaluate",
       js: "document.cookie.split(';').map(c=>c.trim()).find(c=>c.startsWith('lark_oapi_csrf_token='))?.split('=')[1] || document.cookie.split(';').map(c=>c.trim()).find(c=>c.startsWith('lgw_csrf_token='))?.split('=')[1] || document.cookie.split(';').map(c=>c.trim()).find(c=>c.startsWith('swp_csrf_token='))?.split('=')[1] || ''")
     token = extract_string_value(result["output"].to_s)
@@ -181,7 +170,6 @@ class BrowserSession
 
     # Try from window object
     result = @client.call("act",
-      target_id: @target_id,
       kind: "evaluate",
       js: "window.csrfToken || ''")
     extract_string_value(result["output"].to_s)

data/lib/clacky/idle_compression_timer.rb

@@ -41,27 +41,42 @@ module Clacky
         Thread.current.name = "idle-compression-timer"
         sleep IDLE_DELAY
 
-        # Kick off compression in a separate thread so it can be interrupted
-        compress_thread = Thread.new do
-          Thread.current.name = "idle-compression-work"
-          run_compression
+        # Register @compress_thread inside the mutex BEFORE the thread starts running,
+        # so cancel() can always find and interrupt it even if it fires immediately.
+        compress_thread = nil
+        @mutex.synchronize do
+          compress_thread = Thread.new do
+            Thread.current.name = "idle-compression-work"
+            run_compression
+          end
+          @compress_thread = compress_thread
         end
 
-        @mutex.synchronize { @compress_thread = compress_thread }
         compress_thread.join
         @mutex.synchronize { @compress_thread = nil; @timer_thread = nil }
       end
     end
 
     # Cancel the timer and any in-progress compression.
-    # Safe to call even if the timer is not running.
+    # Raises AgentInterrupted on the compress thread and waits for it to fully exit,
+    # ensuring history rollback completes before the caller starts a new agent.run.
     def cancel
+      compress_thread_to_join = nil
+
       @mutex.synchronize do
         @timer_thread&.kill
-        @compress_thread&.raise(Clacky::AgentInterrupted, "Idle timer cancelled")
+        if @compress_thread&.alive?
+          @compress_thread.raise(Clacky::AgentInterrupted, "Idle timer cancelled")
+          compress_thread_to_join = @compress_thread
+        end
         @timer_thread    = nil
         @compress_thread = nil
       end
+
+      # Join outside the mutex to avoid deadlock.
+      # This blocks until the compress thread has finished rolling back history,
+      # so the subsequent agent.run sees a clean, consistent history.
+      compress_thread_to_join&.join(5)
     end
 
     # True if the timer or compression is currently active.