cowork-os 0.3.88 → 0.3.90

package/README.md

@@ -18,66 +18,57 @@
   <a href="https://opensource.org/licenses/MIT"><img src="https://img.shields.io/badge/License-MIT-yellow.svg" alt="License: MIT"></a>
   <a href="https://www.apple.com/macos/"><img src="https://img.shields.io/badge/platform-macOS-blue.svg" alt="macOS"></a>
   <a href="https://www.electronjs.org/"><img src="https://img.shields.io/badge/electron-40.4.1-47848F.svg" alt="Electron"></a>
-  <a href="https://github.com/sponsors/CoWork-OS"><img src="https://img.shields.io/badge/Sponsor-CoWork%20OS-ff4d4f?logo=github-sponsors&logoColor=white" alt="Sponsor on GitHub"></a>
 </p>
 
 **The operating system for personal AI assistants**
 
-Your AI needs a secure home. CoWork OS provides the runtime, security layers, and I/O channels to run AI agents across WhatsApp, Telegram, Discord, Slack, Microsoft Teams, Google Chat, iMessage, Signal, Mattermost, Matrix, Twitch, LINE, BlueBubbles, and Email — with the control you expect from an operating system.
+CoWork OS is a local-first runtime for secure, multi-channel AI agents. It combines a desktop control plane, gateway channels, plugin/skill extensibility, and strict execution guardrails so assistants can operate safely across real workflows.
 
-| | |
+| Capability | Current scope |
 |---|---|
-| **20+ AI Providers** | Claude, OpenAI, Gemini, Bedrock, OpenRouter, Ollama (free/local), Groq, xAI, Kimi, Mistral, Cerebras, MiniMax, Qwen, Copilot, and more |
+| **30+ LLM Providers** | 11 built-in providers plus 20+ custom provider types (OpenAI/Anthropic-compatible, Copilot, Cerebras, Mistral, MiniMax, Qwen, and more) |
 | **14 Messaging Channels** | WhatsApp, Telegram, Discord, Slack, Teams, Google Chat, iMessage, Signal, Mattermost, Matrix, Twitch, LINE, BlueBubbles, Email |
 | **9 Enterprise Connectors** | Salesforce, Jira, HubSpot, Zendesk, ServiceNow, Linear, Asana, Okta, Resend |
-| **6 Cloud Storage** | Notion, Box, OneDrive, Google Workspace (Drive/Gmail/Calendar), Dropbox, SharePoint |
-| **Voice Calls** | Outbound phone calls via ElevenLabs Agents |
-| **Agent Teams** | Multi-agent collaboration with shared checklists and coordinated runs |
-| **Workspace Kit** | Workspace `.cowork/` kit (projects, access rules, context injection, per-workspace settings) |
-| **Intent Routing** | Per-message routing across chat/advice/planning/execution/mixed modes |
-| **Relationship Memory** | Layered continuity memory (identity, preferences, context, history, commitments) |
-| **Completion-First** | Answer-first responses, soft-deadline fallback, timeout recovery finalization |
-| **Security-First** | 2800+ unit tests, configurable guardrails, approval workflows, gateway hardening |
-| **Local-First** | Your data stays on your machine. BYOK (Bring Your Own Key) |
-
-### What’s new in 0.3.73
-
-- **Release and install reliability**: fixed publish-blocking type/lint regressions so npm desktop/client releases can be published reliably.
-- **Task config validation**: stricter `personalityId` validation for task `agentConfig` avoids malformed IDs reaching task execution.
-- **Control-plane/LLM error handling alignment**: tests and runtime error codes now match shared validation shapes for predictable client behavior.
-- **Workspace auto-switch confidence**: improved ambiguous temp-task preflight handling when preferred workspace signals are available.
-
-### Current Branch Enhancements (Unreleased)
-
-- **Executor reliability**: tasks can no longer get stuck in `executing` status after follow-ups; varied-failure loop detection nudges the agent to switch strategy after 5 persistent tool failures.
-- **Smarter artifact handling**: read-only analysis steps no longer require artifact output; verification evidence is skipped when the task already produced files; document creation only triggers for explicit DOCX/PDF requests (attachment filenames no longer cause false positives).
-- **Gateway connection resilience**: Slack Socket Mode uses relaxed ping/pong timeouts (15s/60s) to reduce reconnection churn; WhatsApp adds connection-flap detection with enforced backoff when rapid disconnect/reconnect cycles are detected.
-- **XLSX file support**: Excel spreadsheets (.xlsx/.xls) can now be read and extracted as tab-separated text in the file viewer, with support for formulas, rich text, and dates.
-- **Attachment chips in chat**: user message bubbles now show compact attachment chips instead of raw file listings; the collapsible bubble toggle shows both "Show more" and "Show less".
-- **User identity hardening**: the personality prompt now explicitly instructs the LLM to ignore names found in file paths, filenames, and OS metadata when a preferred name is stored, preventing identity leakage.
-- **Hooks auto-token**: enabled hooks server auto-generates a missing authentication token instead of silently disabling.
-- **Reduced log noise**: suppressed zero-count startup messages for plugins, canvas sessions, and legacy settings; demoted MCP stderr to debug level.
-- **Relationship-agent runtime**: tasks are now intent-routed (`chat`, `advice`, `planning`, `execution`, `mixed`) and strategy-bound before execution.
-- **Answer-first behavior**: strategy-marked prompts emit a direct user answer early, then continue with deeper execution when needed.
-- **Soft-deadline execution control**: long-running steps switch to best-effort finalization before hard timeout to reduce unfinished tasks.
-- **Timeout cancellation recovery**: timeout-triggered cancellation paths now attempt best-effort final answers instead of silent termination.
-- **Relationship memory controls**: new layered memory CRUD and commitment APIs (`open`, `done`, `due soon`) are exposed through IPC/preload and renderer memory settings.
+| **6 Cloud Storage Integrations** | Notion, Box, OneDrive, Google Workspace (Drive/Gmail/Calendar), Dropbox, SharePoint |
+| **100+ Built-in Skills** | Developer, productivity, communication, document, and automation skills in `resources/skills/` |
+| **Declarative Plugin System** | Plugins can register skills, agent roles, connectors, and slash commands via manifests |
+| **5 Built-in Plugin Packs** | content-marketing, customer-support, data-analysis, devops, sales-crm |
+| **Agent Teams + Child Control** | Coordinated multi-agent runs with child-task control and telemetry export |
+| **Collaborative Mode** | Ephemeral multi-agent teams with real-time thought sharing and leader synthesis |
+| **Multi-LLM Mode** | Send the same task to multiple providers/models with judge-based synthesis |
+| **Git Worktree Isolation** | Tasks run in isolated git worktrees with auto-commit, merge, and cleanup |
+| **Agent Comparison Mode** | Run the same task across agents/models side by side and compare outputs |
+| **Multimodal Runtime** | Vision/image support across providers, task image attachments, spreadsheet preview/extraction |
+| **Security-First Execution** | Approval workflows, IPC/input validation hardening, redaction, and gateway safety controls |
+| **Local-First + BYOK** | Your data and keys stay with you |
+
+### Platform Capability Overview
+
+- **Agent runtime and orchestration**: task-based execution, dynamic re-planning, intent-aware behavior, agent teams, collaborative mode, multi-LLM mode, agent comparison, and child-task controls.
+- **LLM stack**: broad provider support (built-in + custom provider types), model flexibility, multi-LLM synthesis, and multimodal execution.
+- **Multichannel interaction**: unified operations across 14 chat/email channels with desktop + headless runtime support.
+- **Integrations and productivity**: enterprise connectors, cloud-storage integrations, and native document/spreadsheet/presentation workflows.
+- **Extensibility**: built-in skills, custom skills, declarative plugins/plugin packs, and MCP client/host/registry support.
+- **Memory and continuity**: persistent memory, relationship context, and commitment lifecycle management.
+- **Security and governance**: approval workflows, configurable guardrails, per-context restrictions, validation hardening, and sensitive-data redaction.
+- **Git isolation**: per-task git worktree isolation with auto-commit, merge, conflict detection, and cleanup.
+- **Operator UX**: live task timelines, talk mode, i18n, file previews (including XLSX), task pinning, collaborative thoughts panel, and control-plane telemetry/export surfaces.
+
+For release-by-release history, see `CHANGELOG.md`.
 
 > **Status**: macOS desktop app + headless/server mode (Linux/VPS). Cross-platform desktop support planned.
 
 ---
 
-## Relationship-Agent Validation (Recommended)
+## Upgrade Validation (Recommended)
 
-Use this quick runbook after upgrades:
+1. Create a task and confirm it transitions to `completed` with a final user-facing answer.
+2. Send a strategic prompt and verify answer-first behavior (direct answer before deeper execution).
+3. Attach an image to a task/follow-up and verify multimodal analysis is available.
+4. In **Settings > Memory**, verify relationship memory and commitments can be edited/forgotten.
+5. In **Settings > LLM**, verify provider/model selection (including Bedrock display names) is functioning.
 
-1. Send a strategic prompt in a channel (for example: "How should I position this product, who should I target, and what should I achieve?").
-2. Confirm you receive a direct answer early (not only research chatter).
-3. Confirm task ends in `completed` with a final response.
-4. If timeout occurs, confirm logs show timeout recovery/finalization instead of silent cancellation.
-5. In **Settings > Memory**, confirm relationship items and commitments can be edited/forgotten and commitment status can be toggled.
-
-For a full acceptance checklist, see `docs/relationship-agent-uat.md`.
+For a detailed acceptance checklist, see `docs/relationship-agent-uat.md`.
 
 ---
 
@@ -253,6 +244,102 @@ For production/persistent setup and Docker/systemd options, use:
 
 ---
 
+## Uninstall
+
+There are two ways to uninstall CoWork OS depending on whether you want to keep local data.
+
+### Option 1: Uninstall app/binaries only (keep database)
+
+This removes installed application files and CLI/package artifacts while keeping workspace, settings, and task data for later restore.
+
+- macOS app (manual drag-installed build):
+
+```bash
+pkill -f '/Applications/CoWork OS.app' || true
+rm -rf "/Applications/CoWork OS.app"
+```
+
+- npm global package install:
+
+```bash
+npm uninstall -g cowork-os
+```
+
+- Local install in a folder:
+
+```bash
+rm -rf ~/cowork-run
+```
+
+- Source/development clone:
+
+```bash
+rm -rf /path/to/CoWork-OS
+```
+
+- VPS/headless Docker install:
+
+```bash
+cd /path/to/docker-compose-dir
+docker compose down
+```
+
+- VPS/headless systemd install:
+
+```bash
+sudo systemctl stop cowork-os cowork-os-node
+sudo systemctl disable cowork-os cowork-os-node
+sudo rm -f /etc/systemd/system/cowork-os.service
+sudo rm -f /etc/systemd/system/cowork-os-node.service
+sudo systemctl daemon-reload
+```
+
+Data locations to keep (choose the one used by your install):
+
+- macOS (Electron): `~/Library/Application Support/cowork-os/`
+- Linux desktop/Electron: `~/.config/cowork-os/`
+- Linux daemon/headless fallback: `~/.cowork/`
+- Node daemon custom path: value passed in `COWORK_USER_DATA_DIR` or `--user-data-dir`
+- Docker/systemd example paths: named volume `cowork_data`, `/var/lib/cowork-os`, and any host bind mount in `/workspace`
+
+### Option 2: Full uninstall + data deletion (database included) — irrecoverable
+
+> ⚠️ **WARNING:** This removes all application data and settings (tasks, tasks timeline, memory, credentials, channel/session state, and the local database). **All data will be deleted and everything will be gone forever.**
+
+Use this only when you are sure you want to destroy local state.
+
+- Delete all user-data locations:
+
+```bash
+rm -rf ~/Library/Application\ Support/cowork-os
+rm -rf ~/.config/cowork-os
+rm -rf ~/.cowork
+```
+
+- Remove with custom user-data path:
+
+```bash
+rm -rf "$COWORK_USER_DATA_DIR"
+```
+
+- Fully remove Docker install data:
+
+```bash
+cd /path/to/docker-compose-dir
+docker compose down -v
+docker compose rm -f
+```
+
+- Fully remove systemd/headless example data:
+
+```bash
+sudo rm -rf /var/lib/cowork-os
+```
+
+After the data wipe, also remove remaining app binaries/shell package entries from Option 1 if you haven't already.
+
+---
+
 ### Security Verified by ZeroLeaks
 
 <p align="center">
@@ -302,7 +389,7 @@ For production/persistent setup and Docker/systemd options, use:
 
 - Claude Code-style tools: `glob`, `grep`, `edit_file`
 - Browser automation with Playwright
-- 85+ bundled skills for popular services
+- 100+ bundled skills for popular services
 - MCP (Model Context Protocol) support for extensibility
 
 ---
@@ -460,11 +547,18 @@ Configure in **Settings** > **Appearance**.
 
 - **Task-Based Workflow**: Multi-step execution with plan-execute-observe loops
 - **Dynamic Re-Planning**: Agent can revise its plan mid-execution
-- **85+ Built-in Skills**: GitHub, Slack, Notion, Spotify, Apple Notes, and more
+- **100+ Built-in Skills**: GitHub, Slack, Notion, Spotify, Apple Notes, and more
 - **Document Creation**: Excel, Word, PDF, PowerPoint with professional formatting
 - **Persistent Memory**: Cross-session context with privacy-aware observation capture
 - **Workspace Kit**: `.cowork/` project kit + markdown indexing with context injection
 - **Agent Teams**: Multi-agent collaboration with shared checklists, coordinated runs, and team management UI
+- **Collaborative Mode**: Auto-create ephemeral teams where multiple agents work on the same task, sharing thoughts in real-time via the Collaborative Thoughts Panel
+- **Multi-LLM Mode**: Send the same task to multiple LLM providers/models simultaneously, with a judge agent synthesizing the best result
+- **Agent Comparison Mode**: Compare agent or model outputs side by side using the ComparisonService and dedicated comparison UI
+- **Git Worktree Isolation**: Tasks can run in isolated git worktrees with automatic branch creation, auto-commit, merge back to base branch, conflict detection, and cleanup
+- **Task Pinning**: Pin important tasks in the sidebar for quick access regardless of sort order
+- **Wrap-Up Task**: Gracefully wrap up running tasks instead of hard-cancelling, allowing the agent to finish its current thought
+- **Capability Matcher**: Auto-select the best agents for a task based on capability matching
 - **Performance Reviews**: Score and review agent-role outcomes, with autonomy-level recommendations
 - **Voice Calls**: Outbound phone calls via ElevenLabs Agents (list agents, list numbers, initiate calls)
 - **Vision**: Analyze workspace images (screenshots, photos, diagrams) via `analyze_image` tool (OpenAI, Anthropic, or Gemini)
@@ -591,6 +685,10 @@ Coordinate multiple agents working together on complex tasks with shared state.
 | **Queue-Friendly** | Team runs respect global concurrency limits by default |
 | **UI Panel** | Full React UI for creating, managing, and monitoring agent teams |
 | **Data Persistence** | SQLite-backed repositories for teams, members, items, and runs |
+| **Collaborative Mode** | Auto-create ephemeral teams where agents share real-time thoughts during execution |
+| **Multi-LLM Mode** | Dispatch the same task to multiple LLM providers with judge-based synthesis |
+| **Collaborative Thoughts** | Real-time thought panel shows agent reasoning, analysis, and synthesis as it happens |
+| **Wrap-Up Support** | Gracefully wrap up team runs instead of hard-cancelling |
 
 Configure in **Mission Control** > **Teams**.
 
@@ -627,6 +725,9 @@ Claude Code-style tools for efficient code navigation and editing:
 | **glob** | Fast pattern-based file search (e.g., `**/*.ts`, `src/**/*.tsx`) |
 | **grep** | Regex content search across files with context lines |
 | **edit_file** | Surgical file editing with find-and-replace |
+| **git_commit** | Commit changes in the workspace (or worktree) with a message |
+| **git_diff** | View staged/unstaged changes in the workspace |
+| **git_branch** | List, create, or switch branches |
 
 ### Browser Automation
 
@@ -710,7 +811,7 @@ Customize agent behavior via Settings or conversation:
 ┌─────────────────────────────────────────────────────────────────┐
 │                    Execution Layer                               │
 │  File Operations | Document Skills | Browser Automation          │
-│  LLM Providers (20+) | Search Providers (4) | MCP Client          │
+│  LLM Providers (30+) | Search Providers (4) | MCP Client          │
 └─────────────────────────────────────────────────────────────────┘
                               ↕
 ┌─────────────────────────────────────────────────────────────────┐
@@ -2124,7 +2225,7 @@ Configure these in **Settings** > **LLM Provider** by entering API keys/tokens,
 
 ---
 
-## Built-in Skills (85+)
+## Built-in Skills (100+)
 
 | Category | Skills |
 |----------|--------|
@@ -2137,6 +2238,8 @@ Configure these in **Settings** > **LLM Provider** by entering API keys/tokens,
 | **Frontend** | Frontend Design, React Native Best Practices |
 | **Data** | Supabase SDK Patterns |
 | **Search** | Local Web Search (SearXNG), Bird |
+| **Finance** | Crypto Trading, Crypto Execution, Trading Foundation |
+| **Marketing** | Email Marketing Bible |
 | **Use Cases** | Booking Options, Draft Reply, Family Digest, Household Capture, Newsletter Digest, Transaction Scan |
 
 ---
@@ -2260,7 +2363,7 @@ Users must comply with their model provider's terms:
 
 ### Completed
 
-- [x] Multi-provider LLM support (20+ providers including Groq, xAI, Kimi, GitHub Copilot, OpenAI/Anthropic-compatible)
+- [x] Multi-provider LLM support (30+ provider types including Groq, xAI, Kimi, GitHub Copilot, OpenAI/Anthropic-compatible)
 - [x] Multi-channel messaging (14 channels)
 - [x] Configurable guardrails and security
 - [x] Browser automation with Playwright
@@ -2270,7 +2373,7 @@ Users must comply with their model provider's terms:
 - [x] WebSocket Control Plane with API
 - [x] Tailscale and SSH remote access
 - [x] Personality system
-- [x] 85+ bundled skills (code reviewer, PRD, multi-PR review, frontend design, local websearch, developer growth, React Native, Supabase, bird)
+- [x] 100+ bundled skills (code reviewer, PRD, multi-PR review, frontend design, local websearch, developer growth, React Native, Supabase, bird)
 - [x] 2800+ unit tests
 - [x] Docker-based sandboxing (cross-platform)
 - [x] Per-context security policies (DM vs group)
@@ -2314,6 +2417,20 @@ Users must comply with their model provider's terms:
 - [x] Docker and VPS deployment support (Dockerfiles, docker-compose, systemd units)
 - [x] Dedicated workspaces for scheduled cron jobs (auto-created from temp workspace)
 - [x] Hook wake requests for heartbeat processing, including queueing, coalescing, and immediate scheduling
+- [x] Git Worktree Isolation (per-task isolated git worktrees with auto-commit, merge, conflict detection, cleanup)
+- [x] Collaborative Mode (ephemeral multi-agent teams with real-time thought sharing and leader synthesis)
+- [x] Multi-LLM Mode (dispatch same task to multiple LLM providers with judge-based synthesis)
+- [x] Agent Comparison Mode (run same task across agents/models and compare outputs side by side)
+- [x] Task Pinning (pin/unpin tasks in sidebar for quick access)
+- [x] Wrap-Up Task (graceful task completion instead of hard cancel)
+- [x] Git Tools (git_commit, git_diff, git_branch for worktree-aware operations)
+- [x] Capability Matcher (auto-select agents based on task requirements)
+- [x] Executor modular refactoring (event emitter, lifecycle mutex, modular utility files)
+- [x] Task event bridge contract (shared allowlist for control-plane event bridging)
+- [x] Collaborative Thoughts Panel (real-time UI for agent thinking during collaborative mode)
+- [x] Scroll-to-bottom button in task view
+- [x] VPS uninstall/removal documentation
+- [x] New skill packs: Crypto Trading, Crypto Execution, Trading Foundation, Email Marketing Bible
 
 ### Planned
 

package/dist/electron/electron/acp/agent-registry.js

@@ -33,15 +33,15 @@ class ACPAgentRegistry {
             id: `local:${role.name}`,
             name: role.displayName,
             description: role.description || `${role.displayName} agent`,
-            version: '1.0.0',
-            provider: 'CoWork OS',
+            version: "1.0.0",
+            provider: "CoWork OS",
             icon: role.icon,
             capabilities,
-            origin: 'local',
+            origin: "local",
             localRoleId: role.id,
             registeredAt: Date.now(),
             lastActiveAt: Date.now(),
-            status: role.isActive ? 'available' : 'offline',
+            status: role.isActive ? "available" : "offline",
         };
     }
     /**
@@ -71,7 +71,7 @@ class ACPAgentRegistry {
         if (remote)
             return remote;
         // Check local agents
-        if (agentId.startsWith('local:')) {
+        if (agentId.startsWith("local:")) {
             const roleName = agentId.slice(6); // Remove 'local:' prefix
             const role = roles.find((r) => r.name === roleName && r.isActive);
             if (role)
@@ -107,12 +107,12 @@ class ACPAgentRegistry {
      * Register a remote agent
      */
     registerRemoteAgent(params) {
-        const id = `remote:${(0, crypto_1.randomUUID)().slice(0, 8)}-${params.name.toLowerCase().replace(/\s+/g, '-')}`;
+        const id = `remote:${(0, crypto_1.randomUUID)().slice(0, 8)}-${params.name.toLowerCase().replace(/\s+/g, "-")}`;
         const card = {
             id,
             name: params.name,
             description: params.description,
-            version: params.version || '1.0.0',
+            version: params.version || "1.0.0",
             provider: params.provider,
             icon: params.icon,
             capabilities: (params.capabilities || []).map((c) => ({
@@ -125,10 +125,10 @@ class ACPAgentRegistry {
             outputContentTypes: params.outputContentTypes,
             supportsStreaming: params.supportsStreaming,
             endpoint: params.endpoint,
-            origin: 'remote',
+            origin: "remote",
             registeredAt: Date.now(),
             lastActiveAt: Date.now(),
-            status: 'available',
+            status: "available",
             metadata: params.metadata,
         };
         this.remoteAgents.set(id, card);

package/dist/electron/electron/acp/handler.js

@@ -29,12 +29,15 @@ function getACPRegistry() {
 // ===== Validation helpers =====
 function requireAuth(client) {
     if (!client.isAuthenticated) {
-        throw { code: protocol_1.ErrorCodes.UNAUTHORIZED, message: 'Authentication required' };
+        throw { code: protocol_1.ErrorCodes.UNAUTHORIZED, message: "Authentication required" };
     }
 }
 function requireString(value, field) {
-    if (typeof value !== 'string' || !value.trim()) {
-        throw { code: protocol_1.ErrorCodes.INVALID_PARAMS, message: `${field} is required and must be a non-empty string` };
+    if (typeof value !== "string" || !value.trim()) {
+        throw {
+            code: protocol_1.ErrorCodes.INVALID_PARAMS,
+            message: `${field} is required and must be a non-empty string`,
+        };
     }
     return value.trim();
 }
@@ -57,7 +60,7 @@ function registerACPMethods(server, deps) {
     server.registerMethod(types_1.ACPMethods.AGENT_GET, async (client, params) => {
         requireAuth(client);
         const p = params;
-        const agentId = requireString(p?.agentId, 'agentId');
+        const agentId = requireString(p?.agentId, "agentId");
         const roles = deps.getActiveRoles();
         const agent = reg.getAgent(agentId, roles);
         if (!agent) {
@@ -69,8 +72,8 @@ function registerACPMethods(server, deps) {
     server.registerMethod(types_1.ACPMethods.AGENT_REGISTER, async (client, params) => {
         requireAuth(client);
         const p = (params || {});
-        requireString(p.name, 'name');
-        requireString(p.description, 'description');
+        requireString(p.name, "name");
+        requireString(p.description, "description");
         const card = reg.registerRemoteAgent(p);
         // Broadcast registration event
         server.broadcast(types_1.ACPEvents.AGENT_REGISTERED, { agent: card });
@@ -80,9 +83,9 @@ function registerACPMethods(server, deps) {
     server.registerMethod(types_1.ACPMethods.AGENT_UNREGISTER, async (client, params) => {
         requireAuth(client);
         const p = params;
-        const agentId = requireString(p?.agentId, 'agentId');
-        if (!agentId.startsWith('remote:')) {
-            throw { code: protocol_1.ErrorCodes.INVALID_PARAMS, message: 'Only remote agents can be unregistered' };
+        const agentId = requireString(p?.agentId, "agentId");
+        if (!agentId.startsWith("remote:")) {
+            throw { code: protocol_1.ErrorCodes.INVALID_PARAMS, message: "Only remote agents can be unregistered" };
         }
         const removed = reg.unregisterRemoteAgent(agentId);
         if (!removed) {
@@ -96,8 +99,8 @@ function registerACPMethods(server, deps) {
     server.registerMethod(types_1.ACPMethods.MESSAGE_SEND, async (client, params) => {
         requireAuth(client);
         const p = (params || {});
-        const to = requireString(p.to, 'to');
-        const body = requireString(p.body, 'body');
+        const to = requireString(p.to, "to");
+        const body = requireString(p.body, "body");
         // Validate target agent exists
         const roles = deps.getActiveRoles();
         const targetAgent = reg.getAgent(to, roles);
@@ -108,12 +111,12 @@ function registerACPMethods(server, deps) {
             id: (0, crypto_1.randomUUID)(),
             from: p.from || `client:${client.id}`,
             to,
-            contentType: p.contentType || 'text/plain',
+            contentType: p.contentType || "text/plain",
             body,
             data: p.data,
             correlationId: p.correlationId,
             replyTo: p.replyTo,
-            priority: p.priority || 'normal',
+            priority: p.priority || "normal",
             timestamp: Date.now(),
             ttlMs: p.ttlMs,
         };
@@ -123,15 +126,15 @@ function registerACPMethods(server, deps) {
         server.broadcast(types_1.ACPEvents.MESSAGE_RECEIVED, { message });
         // If the target is a local agent and we have task creation capability,
         // auto-create a task from high-priority messages
-        if (targetAgent.origin === 'local' &&
+        if (targetAgent.origin === "local" &&
             targetAgent.localRoleId &&
-            p.priority === 'high' &&
+            p.priority === "high" &&
             deps.createTask) {
             try {
                 const result = await deps.createTask({
                     title: `ACP message from ${message.from}`,
                     prompt: body,
-                    workspaceId: '', // Will use default workspace
+                    workspaceId: "", // Will use default workspace
                     assignedAgentRoleId: targetAgent.localRoleId,
                 });
                 message.data = { ...(message.data || {}), autoTaskId: result.taskId };
@@ -146,7 +149,7 @@ function registerACPMethods(server, deps) {
     server.registerMethod(types_1.ACPMethods.MESSAGE_LIST, async (client, params) => {
         requireAuth(client);
         const p = (params || {});
-        const agentId = requireString(p.agentId, 'agentId');
+        const agentId = requireString(p.agentId, "agentId");
         const drain = p.drain === true;
         const messages = reg.getMessages(agentId, drain);
         return { messages };
@@ -155,9 +158,9 @@ function registerACPMethods(server, deps) {
     server.registerMethod(types_1.ACPMethods.TASK_CREATE, async (client, params) => {
         requireAuth(client);
         const p = (params || {});
-        const assigneeId = requireString(p.assigneeId, 'assigneeId');
-        const title = requireString(p.title, 'title');
-        const prompt = requireString(p.prompt, 'prompt');
+        const assigneeId = requireString(p.assigneeId, "assigneeId");
+        const title = requireString(p.title, "title");
+        const prompt = requireString(p.prompt, "prompt");
         // Validate assignee exists
         const roles = deps.getActiveRoles();
         const assignee = reg.getAgent(assigneeId, roles);
@@ -170,26 +173,26 @@ function registerACPMethods(server, deps) {
             assigneeId,
             title,
             prompt,
-            status: 'pending',
+            status: "pending",
             workspaceId: p.workspaceId,
             createdAt: Date.now(),
             updatedAt: Date.now(),
         };
         // If assignee is a local agent, delegate to the CoWork task system
-        if (assignee.origin === 'local' && assignee.localRoleId && deps.createTask) {
+        if (assignee.origin === "local" && assignee.localRoleId && deps.createTask) {
             try {
                 const result = await deps.createTask({
                     title,
                     prompt,
-                    workspaceId: p.workspaceId || '',
+                    workspaceId: p.workspaceId || "",
                     assignedAgentRoleId: assignee.localRoleId,
                 });
                 acpTask.coworkTaskId = result.taskId;
-                acpTask.status = 'running';
+                acpTask.status = "running";
             }
             catch (err) {
-                acpTask.status = 'failed';
-                acpTask.error = err?.message || 'Failed to create task';
+                acpTask.status = "failed";
+                acpTask.error = err?.message || "Failed to create task";
             }
         }
         acpTasks.set(acpTask.id, acpTask);
@@ -201,7 +204,7 @@ function registerACPMethods(server, deps) {
     server.registerMethod(types_1.ACPMethods.TASK_GET, async (client, params) => {
         requireAuth(client);
         const p = params;
-        const taskId = requireString(p?.taskId, 'taskId');
+        const taskId = requireString(p?.taskId, "taskId");
         const acpTask = acpTasks.get(taskId);
         if (!acpTask) {
             throw { code: protocol_1.ErrorCodes.INVALID_PARAMS, message: `ACP task not found: ${taskId}` };
@@ -211,17 +214,17 @@ function registerACPMethods(server, deps) {
             const coworkTask = deps.getTask(acpTask.coworkTaskId);
             if (coworkTask) {
                 const statusMap = {
-                    pending: 'pending',
-                    running: 'running',
-                    completed: 'completed',
-                    failed: 'failed',
-                    cancelled: 'cancelled',
+                    pending: "pending",
+                    running: "running",
+                    completed: "completed",
+                    failed: "failed",
+                    cancelled: "cancelled",
                 };
                 const newStatus = statusMap[coworkTask.status] || acpTask.status;
                 if (newStatus !== acpTask.status) {
                     acpTask.status = newStatus;
                     acpTask.updatedAt = Date.now();
-                    if (newStatus === 'completed' || newStatus === 'failed' || newStatus === 'cancelled') {
+                    if (newStatus === "completed" || newStatus === "failed" || newStatus === "cancelled") {
                         acpTask.completedAt = Date.now();
                     }
                     if (coworkTask.error) {
@@ -254,12 +257,12 @@ function registerACPMethods(server, deps) {
     server.registerMethod(types_1.ACPMethods.TASK_CANCEL, async (client, params) => {
         requireAuth(client);
         const p = params;
-        const taskId = requireString(p?.taskId, 'taskId');
+        const taskId = requireString(p?.taskId, "taskId");
         const acpTask = acpTasks.get(taskId);
         if (!acpTask) {
             throw { code: protocol_1.ErrorCodes.INVALID_PARAMS, message: `ACP task not found: ${taskId}` };
         }
-        if (acpTask.status === 'completed' || acpTask.status === 'cancelled') {
+        if (acpTask.status === "completed" || acpTask.status === "cancelled") {
             throw { code: protocol_1.ErrorCodes.INVALID_PARAMS, message: `Task is already ${acpTask.status}` };
         }
         // Cancel the underlying CoWork task if it exists
@@ -271,14 +274,14 @@ function registerACPMethods(server, deps) {
                 // Best-effort cancellation
             }
         }
-        acpTask.status = 'cancelled';
+        acpTask.status = "cancelled";
         acpTask.updatedAt = Date.now();
         acpTask.completedAt = Date.now();
         // Broadcast cancellation event
         server.broadcast(types_1.ACPEvents.TASK_UPDATED, { task: acpTask });
         return { task: acpTask };
     });
-    console.log('[ACP] Registered 10 ACP method handlers on Control Plane');
+    console.log("[ACP] Registered 10 ACP method handlers on Control Plane");
 }
 /**
  * Cleanup ACP state (call on shutdown)