myaiforone 1.0.0

package/docs/user-guide.md

@@ -0,0 +1,2016 @@
+# MyAIforOne — User Guide
+
+> Master reference for every page, button, action, API endpoint, and MCP tool in the MyAIforOne platform.
+> Organized by UI page. Each action includes its API and MCP mapping.
+> Intended audience: hub agent, platform agents, and advanced users.
+
+---
+
+## Table of Contents
+
+1. [Home](#1-home)
+2. [Agents (Org)](#2-agents-org)
+   - [Agent List](#21-agent-list)
+   - [Agent Creation / Edit Modal](#22-agent-creation--edit-modal)
+   - [Agent Dashboard](#23-agent-dashboard)
+3. [Chat](#3-chat)
+4. [Library](#4-library)
+5. [Lab](#5-lab)
+6. [Marketplace](#6-marketplace)
+7. [Admin](#7-admin)
+   - [Channels](#71-channels)
+   - [Activity](#72-activity)
+   - [Settings](#73-settings)
+   - [Docs](#74-docs)
+8. [Monitor](#8-monitor)
+   - [Setup Checklist](#81-setup-checklist)
+   - [Dashboard](#82-dashboard)
+9. [Projects](#9-projects)
+10. [AI Gym](#10-ai-gym)
+   - [Onboarding](#101-onboarding-first-visit)
+   - [Main Gym View](#102-main-gym-view)
+   - [Dimensions](#103-ai-strength-dimensions)
+   - [Programs](#104-programs)
+   - [Activity Digest](#105-activity-digest)
+   - [Trainer Souls](#106-trainer-souls)
+   - [API & MCP Reference](#107-api--mcp-reference)
+
+---
+
+# 1. Home
+
+**URL:** `/` or `/home`
+**Purpose:** The primary interface — a conversational AI hub. Users talk to their personal AI, which routes to the right agent or handles platform operations directly.
+
+## 1.1 Landing View
+
+The initial state before any message is sent.
+
+### Chat Input (Landing)
+- **Large textarea** — type a message to any agent
+- **Send button (→)** — sends the first message and transitions to chat view
+  - If message contains `@alias`, routes to that agent
+  - If no mention, routes to the default/hub agent
+
+### Agent Tiles
+- **Org filter dropdown** — filter visible tiles by organization
+  - Options: "Recent" (default), plus all organization names
+- **Agent tile grid** — clickable tiles showing each agent
+  - Each tile shows: avatar (2-letter initials), name, last message time (in Recent view)
+  - **Click a tile** → inserts `@alias` into the chat input
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List agents for tiles | `GET /api/agents` | `list_agents` |
+| | **Params:** `?org=orgName` (optional) | **Params:** `org` (optional string) |
+
+## 1.2 Chat View
+
+Active after the first message is sent.
+
+### Agent Sidebar (Left Panel)
+- **Org filter dropdown** — filter sidebar agents by organization (top of sidebar)
+- **Recent section** — agents sorted by most recent interaction first
+  - Each entry: avatar, agent name, alias, time since last activity
+  - **Click an agent** → inserts `@alias` into chat input
+- **All Agents section** — collapsible section (click header to expand/collapse)
+  - Shows all agents not in the Recent section, with count badge
+  - Arrow indicator (▶/▼) shows collapsed/expanded state
+- **◀ Collapse button** — collapses the entire sidebar to zero width, giving chat full width
+  - Click again (or the Agents toggle) to re-expand
+
+### Chat Header (Per-Agent)
+When chatting with an agent, the header shows:
+- **Agent avatar and name**
+- **Account override dropdown** — select which Claude account to use for this agent
+  - Options: "default" (agent's configured account), plus all other configured accounts
+  - In-memory only — resets on page reload
+  - The selected account is passed as `accountOverride` on every send/stream call
+- **▨ Canvas button** — toggles the canvas/preview panel (right side)
+  - When active, shows files created or modified by the agent
+  - Canvas has: download button (downloads file), close button, file content display
+- **↺ Reset Session button** — (only for persistent-session agents) compacts and resets the conversation
+  - Shows confirmation dialog: "Reset session? This clears conversation history. Use /opcompact first to save important context."
+  - On confirm, calls reset session API
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Reset session | `POST /api/agents/:agentId/sessions/reset` | `reset_session` |
+| | **Body:** `{ senderId? }` | **Params:** `agentId`, `senderId?` |
+| Account override | Passed as `accountOverride` param on `POST /api/chat/:agentId` and `/stream` | `send_message` / `start_stream` — `accountOverride` param |
+| List accounts (populates dropdown) | `GET /api/config/accounts` | `list_accounts` |
+| | | *(no params)* |
+| Download canvas file | `GET /api/agents/:agentId/download?path=filePath` | `download_agent_file` |
+| | **Query:** `path` (absolute file path) | **Params:** `agentId`, `path` |
+
+### Chat Messages Area
+- **User messages** — right-aligned, light blue background
+- **Agent messages** — left-aligned, purple background, markdown-rendered
+  - Shows `[via:agentId]` tag indicating which agent responded
+  - File references render with **Download** and **▨ Preview** buttons
+- **Thinking indicator** — animated dots while agent is processing
+- **Tool usage cards** — show which tools the agent is actively using
+
+### Chat Input
+- **Textarea** — multi-line input, auto-resizes
+  - Placeholder: "Type @ to mention an agent..."
+  - `Enter` sends, `Shift+Enter` for newline
+- **@mention autocomplete** — appears while typing `@`
+  - Shows matching agents with avatar, alias, name
+  - Arrow keys to navigate, Tab/Enter to select, Escape to dismiss
+- **Send button (→)** — sends message to the mentioned (or default) agent
+- **File attachment button (📎)** — opens file picker to upload files to the agent
+- **Voice record button (🎤)** — start/stop voice recording for transcription
+- **Voice mode button (🔇/🔊)** — toggles reading agent responses aloud via text-to-speech
+
+### Streaming Controls (During Active Response)
+While an agent is streaming a response:
+- **Stop button (■)** — red square button, appears next to Send. Interrupts the agent's response immediately.
+  - Keyboard shortcut: `Escape`
+- **Send button becomes "Queue"** — the send button changes to amber "Queue" mode
+  - Type a follow-up message and click Queue (or press Enter) to queue it
+  - Queued messages auto-send when the current response completes
+- **Message Queue** — appears below the input when messages are queued
+  - Each queued message shows: position number, message text preview
+  - **Edit button (✎)** — edit the queued message text
+  - **Cancel button (✕)** — remove the queued message
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Stop / interrupt streaming | `POST /api/chat/jobs/:jobId/stop` | `stop_chat_job` |
+| | | **Params:** `jobId` |
+| Queue sends via normal chat | `POST /api/chat/:agentId` or `/stream` | `send_message` / `start_stream` |
+| | Queue is client-side; each queued message sends normally when prior response completes | |
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Send message (sync) | `POST /api/chat/:agentId` | `send_message` |
+| | **Body:** `{ text, accountOverride? }` | **Params:** `agentId`, `text`, `accountOverride?` |
+| Start streaming chat | `POST /api/chat/:agentId/stream` | `start_stream` |
+| | **Body:** `{ text, accountOverride? }` | **Params:** `agentId`, `text`, `accountOverride?` |
+| Poll stream output | `GET /api/chat/jobs/:jobId/raw?after=N` | `get_chat_job_raw` |
+| | **Query:** `after` (line offset) | **Params:** `jobId`, `after?` |
+| Stop streaming job | `POST /api/chat/jobs/:jobId/stop` | `stop_chat_job` |
+| | | **Params:** `jobId` |
+| Delegate to agent | `POST /api/delegate` | `delegate_message` |
+| | **Body:** `{ agentId, text }` | **Params:** `agentId`, `text` |
+| Upload file | `POST /api/upload/:agentId/json` | `upload_file` |
+| | **Body:** `{ fileName, base64Content, mode? }` | **Params:** `agentId`, `fileName`, `base64Content`, `mode?` |
+
+### Raw Logs Panel (Bottom Drawer)
+- **Toggle button (⌘)** — shows/hides raw Claude output
+- **Clear button** — clears all log entries
+- **Close button** — hides the panel
+- **Log content area** — live streaming output from the agent's Claude process
+
+---
+
+# 2. Agents (Org)
+
+**URL:** `/org`
+**Purpose:** Manage all agents — create, configure, organize, and monitor.
+
+## 2.1 Agent List
+
+### Sub-Navigation Tabs
+- **Teams** (`/org`) — agents view (default)
+- **Tasks** (`/tasks`) — cross-agent task panel
+- **Automations** (`/automations`) — goals and cron jobs across all agents
+
+### View Controls
+
+| Control | Description |
+|---------|-------------|
+| **Grid view (◊)** | 300px agent cards in a wrapping grid |
+| **Compact view (▬▬)** | Condensed cards with 3-dot menu |
+| **List view (≡)** | Sortable table with columns: Name, Alias, Org, Dept, Title, Class, Status, Last Active, Msgs |
+| **Search input** | Filter agents by name, alias, or description |
+| **Org select dropdown** | Filter by organization |
+| **Class filter dropdown** | Checkboxes: Standard, Builder, Platform |
+| **Hide names button (👁 Hide)** | Blurs all agent names globally for presentations |
+| **Per-org hide button (👁)** | Each organization section has its own hide/show toggle to blur names within that org only |
+| **Select mode button (☑ Select)** | Enables multi-select checkboxes on each agent |
+| **+ New Agent button** | Opens agent creation modal |
+
+### Agent Cards (All Views)
+Each agent shows:
+- **Avatar** — 2-letter initials in colored circle
+- **Heartbeat indicator (❤)** — pulsing animation if heartbeat is active
+- **Automation indicator (⟳)** — spinning icon if goals/crons are active
+- **Name** — agent display name
+- **Subtitle** — org, department, title
+- **Status dot** — green for active
+- **Hover actions:**
+  - **Chat button** → opens `/ui#agentId`
+  - **Dashboard button** → opens `/agent-dashboard?id=agentId`
+  - **Config button** → opens edit modal
+
+### List View (Table)
+- **Sortable columns** — click any header to sort ascending/descending
+- **Row click** → opens config modal for that agent
+
+### Multi-Select Mode
+When activated via Select button:
+- **Checkboxes** appear on each agent
+- **Select All / Deselect All** buttons
+- **Delete Selected** — bulk delete with confirmation
+- **Cancel** — exits multi-select mode
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List all agents | `GET /api/agents` | `list_agents` |
+| | **Query:** `?org=orgName` | **Params:** `org?` (string) |
+| Get agent details | `GET /api/agents/:id` | `get_agent` |
+| | | **Params:** `agentId` |
+| Delete agent | `DELETE /api/agents/:id` | `delete_agent` |
+| | **Query:** `?confirmAlias=alias` | **Params:** `agentId`, `confirmAlias` |
+| Get agent registry | `GET /api/agent-registry` | `get_agent_registry` |
+| | *(no params)* | *(no params)* |
+
+## 2.2 Agent Creation / Edit Modal
+
+Opened by **+ New Agent** button or clicking an agent's Config button. Has 6 standard tabs, plus 4 additional tabs when editing a Gym-class agent (10 total).
+
+### Tab 1: Overview
+
+| Field | Description |
+|-------|-------------|
+| **Agent ID** | Lowercase, hyphens only (e.g., `my-agent`). Auto-generates alias on input. |
+| **Mention Alias** | @-prefixed (e.g., `@myagent`). Auto-generated from ID but editable. |
+| **Name** | Display name shown in UI and chat. |
+| **Description** | Short description of the agent's purpose. |
+| **Instructions (CLAUDE.md)** | System prompt written to the agent's CLAUDE.md file. Multi-line textarea. |
+| **Agent Class** | Dropdown: Standard, Builder, Platform, Gym. |
+| **Executor** | Dropdown: Platform Default, Claude, Ollama models, or cloud providers — OpenAI, Grok (xAI), Gemini, Groq, Together, Mistral (when multi-model is enabled). Format: `provider:model` (e.g., `openai:gpt-4o`, `gemini:gemini-2.5-flash`, `grok:grok-3`). Overrides the service-level default executor for this agent. |
+
+#### Organization Entries
+Each agent can belong to multiple organizations. Per entry:
+- **Organization** — org name
+- **Function/Department** — department within the org
+- **Title** — agent's role title
+- **Reports To** — optional, who this agent reports to
+- **Remove button (×)** — removes this org entry
+- **+ Add Org Entry** — adds another org association
+
+#### Heartbeat Section (Collapsible)
+- **Enabled toggle** — pill switch to enable/disable heartbeat
+- **Instructions** — what the agent should check during heartbeat
+- **Frequency dropdown** — Manual, Daily, Weekdays, Weekly, Every N mins
+- **Day select** — for weekly (Mon-Sun)
+- **Time inputs** — hour (0-23), minute (0-59), AM/PM
+- **Cron preview** — read-only display of the generated cron expression
+
+#### Wiki Learning Section (Collapsible)
+- **Wiki Learning toggle** — enable/disable Wiki Learning (`wiki` field)
+  - When enabled, the agent captures facts and corrections from conversations into `learned.md`
+- **Wiki Sync toggle** — enable/disable scheduled Wiki Sync (`wikiSync.enabled`)
+  - Automatically merges `learned.md` into `context.md` on a schedule
+- **Schedule** — cron expression for sync frequency (`wikiSync.schedule`, default: `0 0 * * *`)
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Create agent | `POST /api/agents` | `create_agent` |
+| | **Body:** `{ agentId, alias, name, description, instructions, agentClass, executor, orgs[], heartbeat{}, wiki, wikiSync{}, ... }` | **Params:** `agentId`, `alias`, `name`, `description`, `instructions`, `agentClass`, `executor`, `orgs`, `heartbeat`, `workspace`, `allowedTools`, `mcps`, `routes`, `persistent`, `streaming`, `advancedMemory`, `autonomousCapable`, `autoCommit`, `timeout`, `claudeAccount`, `wiki`, `wikiSync` |
+| Update agent | `PUT /api/agents/:id` | `update_agent` |
+| | **Body:** same fields as create | **Params:** `agentId`, plus any fields to update |
+| Get agent instructions | `GET /api/agents/:id/instructions` | `get_agent_instructions` |
+| | | **Params:** `agentId` |
+
+### Tab 2: Skills
+
+Controls which tools the agent can use.
+
+- **Default tool pills** (toggle on/off):
+  - Read, Edit, Write, Glob, Grep, Bash, WebFetch, WebSearch
+  - Each is a clickable pill — checked (✓) = enabled
+- **Custom skills** — additional skills from library shown as toggle pills
+- **+ Add New Skill button** — opens skill creation flow
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get agent skills | `GET /api/agents/:agentId/skills` | `get_agent_skills` |
+| | | **Params:** `agentId` |
+| Get org skills | `GET /api/skills/org/:orgName` | `get_org_skills` |
+| | | **Params:** `orgName` |
+| Create skill | `POST /api/skills/create` | `create_skill` |
+| | **Body:** `{ id, name, description, content, scope, orgName?, agentId? }` | **Params:** `id`, `name`, `description`, `content`, `scope`, `orgName?`, `agentId?` |
+| Assign skill to agent | `POST /api/marketplace/assign` | `assign_to_agent` |
+| | **Body:** `{ agentId, itemId, type: "skill" }` | **Params:** `agentId`, `itemId`, `type` |
+
+### Tab 3: MCPs
+
+Manage MCP server connections for the agent.
+
+- **MCP toggle pills** — enable/disable each MCP from the registry
+- **MCP accordion sections** (per MCP, expandable):
+  - **API Key cards** — key name, status badge (OK/Missing), secret input, save/disconnect buttons
+  - **OAuth connections** — account name, status, remove button
+  - **+ Add Account button** — opens MCP auth modal
+- **Unconfigured MCPs section** — MCPs not yet connected
+
+#### MCP Auth Modal
+- **Auth form** — dynamic fields based on MCP requirements (API key input, OAuth link, etc.)
+- **Cancel / Save buttons**
+
+#### Connecting Google OAuth MCPs (Sheets, Gmail, Calendar, Drive, Docs)
+
+Several MCPs use Google OAuth instead of an API key. These include:
+
+| MCP | What it enables |
+|-----|----------------|
+| `googlesheets` | Read, write, and manage Google Sheets spreadsheets |
+| `gmail` | Read, search, send Gmail messages |
+| `googlecalendar` | List, create, and manage calendar events |
+| `googledrive` | Browse, read, and manage Drive files |
+| `googledocs` | Read and edit Google Docs documents |
+
+**To connect a Google OAuth MCP to an agent:**
+1. Go to **Org page** → click the agent → **MCPs tab**
+2. Find the MCP (e.g. Google Sheets) in the accordion and click **+ Add Account**
+3. In the auth modal, click **Authorize** — this opens a Google sign-in flow
+4. Sign in with the Google account you want to connect
+5. The refresh token is stored automatically under a connection name like `googlesheets-agenticledger`
+
+**Multiple accounts:** You can add multiple Google accounts to the same MCP (e.g. `agenticledger`, `bst`, `financeiscooked`) by repeating the flow with a different account. Each becomes a separate named connection.
+
+**Via MCP tool:** Use `create_mcp_connection` with `baseMcp`, `label`, `envVar` (e.g. `GOOGLESHEETS_AGENTICLEDGER`), and `value` (refresh token). The token must be issued by the agenticledger.ai OAuth client — tokens from other Google Cloud projects will fail with `invalid_grant`.
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List registered MCPs | `GET /api/mcps` | `list_mcps` |
+| | | *(no params)* |
+| Get MCP catalog | `GET /api/mcp-catalog` | `get_mcp_catalog` |
+| | | *(no params)* |
+| List MCP keys for agent | `GET /api/agents/:id/mcp-keys` | `list_mcp_keys` |
+| | | **Params:** `agentId` |
+| Save MCP key | `POST /api/agents/:id/mcp-keys` | `save_mcp_key` |
+| | **Body:** `{ mcpName, envVar, value }` | **Params:** `agentId`, `mcpName`, `envVar`, `value` |
+| Delete MCP key | `DELETE /api/agents/:id/mcp-keys/:mcpName` | `delete_mcp_key` |
+| | | **Params:** `agentId`, `mcpName` |
+| List MCP connections | `GET /api/agents/:id/mcp-connections` | `list_mcp_connections` |
+| | | **Params:** `agentId` |
+| Create MCP connection | `POST /api/agents/:id/mcp-connections` | `create_mcp_connection` |
+| | **Body:** `{ mcpName, instanceName, config }` | **Params:** `agentId`, `body` (object with mcpName, instanceName, config) |
+| Delete MCP connection | `DELETE /api/agents/:id/mcp-connections/:instanceName` | `delete_mcp_connection` |
+| | | **Params:** `agentId`, `instanceName` |
+
+### Tab 4: Schedules (Cron Jobs)
+
+Trigger automated messages on a schedule.
+
+- **+ Add Schedule button** — adds a new schedule entry
+- **Per schedule entry:**
+  - **Frequency dropdown** — Manual, Daily, Weekly, Every N minutes
+  - **Day checkboxes** (if weekly) — Mon through Sun
+  - **Time inputs** — hours, minutes
+  - **Message textarea** — the message to send when triggered
+  - **Channel select** — which channel to send to
+  - **Chat ID input** — target chat/conversation
+  - **Cron preview** — human-readable expression
+  - **Trigger Now button** — manually fires the schedule immediately
+  - **Pause/Resume button** — toggles cron on/off
+  - **Remove button (×)** — deletes the schedule
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Create cron | `POST /api/agents/:id/cron` | `create_cron` |
+| | **Body:** `{ schedule, message, channel?, chatId?, enabled? }` | **Params:** `agentId`, `body` (object with schedule, message, channel, chatId, enabled) |
+| Toggle cron on/off | `POST /api/agents/:id/cron/:index/toggle` | `toggle_cron` |
+| | | **Params:** `agentId`, `index` (number) |
+| Trigger cron manually | `POST /api/agents/:id/cron/:index/trigger` | `trigger_cron` |
+| | | **Params:** `agentId`, `index` (number) |
+| Get cron run history | `GET /api/agents/:id/cron/:index/history` | `get_cron_history` |
+| | | **Params:** `agentId`, `index` (number) |
+| Delete cron | `DELETE /api/agents/:id/cron/:index` | `delete_cron` |
+| | | **Params:** `agentId`, `index` (number) |
+
+### Tab 5: Goals
+
+Autonomous task tracking with reporting.
+
+- **+ Add Goal button** — adds a new goal entry
+- **Per goal entry:**
+  - **Goal ID** — unique identifier (auto-generated from title if blank)
+  - **Goal Name/Title** — what is this goal
+  - **Description textarea** — detailed description
+  - **Success Metric textarea** — how to measure completion
+  - **Reporting channels** — where to send reports
+    - Per channel: channel dropdown, chat ID input, remove (×)
+    - **+ Add Channel** — adds another report destination
+  - **Enabled toggle** — pill switch
+  - **Trigger mode dropdown** — Heartbeat (default) or On-demand
+  - **Heartbeat settings** (if Heartbeat mode) — frequency, days, time
+  - **Budget input** — max daily USD spend
+  - **Trigger Now button** — manually executes goal
+  - **Pause/Resume button** — toggles goal on/off
+  - **Remove button (×)** — deletes the goal
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Create goal | `POST /api/agents/:id/goals` | `create_goal` |
+| | **Body:** `{ goalId, name, description, successMetric, reportTargets[], triggerMode, schedule?, budget?, enabled? }` | **Params:** `agentId`, `body` (object) |
+| Toggle goal on/off | `POST /api/agents/:id/goals/:goalId/toggle` | `toggle_goal` |
+| | | **Params:** `agentId`, `goalId` |
+| Trigger goal manually | `POST /api/agents/:id/goals/:goalId/trigger` | `trigger_goal` |
+| | | **Params:** `agentId`, `goalId` |
+| Get goal run history | `GET /api/agents/:id/goals/:goalId/history` | `get_goal_history` |
+| | | **Params:** `agentId`, `goalId` |
+| Delete goal | `DELETE /api/agents/:id/goals/:goalId` | `delete_goal` |
+| | | **Params:** `agentId`, `goalId` |
+| List all automations | `GET /api/automations` | `list_automations` |
+| | *(no params)* | *(no params)* |
+
+### Tab 6: Config
+
+Advanced agent configuration.
+
+#### Behavior Toggles (Pill Switches)
+| Toggle | Description |
+|--------|-------------|
+| **Persistent Session** | Keep conversation context between messages (default: on) |
+| **Streaming** | Stream responses in real-time (default: off) |
+| **Advanced Memory** | Enable long-term semantic recall with daily journals + vector search (default: on) |
+| **Autonomous Capable** | Allow agent to act on goals autonomously (default: on) |
+| **Auto Commit** | Auto-commit code changes (default: off) |
+
+#### Other Fields
+- **Workspace** — working directory for the agent (default: `~`)
+- **Agent Home** — computed path (read-only)
+- **Timeout** — max execution time in seconds (default: 14400)
+- **Claude Account** — which Claude account to use (dropdown of configured accounts)
+
+#### Routes
+Which channels can invoke this agent.
+- Per route: channel dropdown, chat/channel ID, require mention checkbox, remove (×)
+- **+ Add Route** — adds a new channel route
+
+### Gym-Specific Tabs (Agent Class = Gym Only)
+
+When editing an agent with **Agent Class: Gym**, four additional tabs appear after Config. These tabs are hidden for all other agent classes.
+
+#### Trainer Tab
+
+Switch the active coaching personality. Changes how the coach communicates during training sessions without affecting progress or profile data.
+
+| Element | Description |
+|---------|-------------|
+| **Trainer grid** | 5 cards in a responsive grid, each showing avatar (initials), name, coaching style |
+| **Active indicator** | Selected trainer has a colored border, tinted background, and "Active" badge |
+| **Click to switch** | Clicking a card immediately updates the learner profile via `PUT /api/gym/learner-profile` with `{ selectedTrainer }` |
+
+Trainers: Alex (Collaborative, cyan), Jordan (Direct, orange), Morgan (Thoughtful, purple), Riley (Challenging, red), Sam (Patient, green).
+
+#### Learner Tab
+
+View and edit the learner profile built during onboarding. The coach updates this profile as you train.
+
+| Field | Description |
+|-------|-------------|
+| **Identity** | JSON or text — professional background, role, experience level |
+| **Goals** | Newline-separated list of learning goals |
+| **Current Streak** | Read-only counter of consecutive active days |
+| **Longest Streak** | Read-only all-time streak record |
+| **Strengths** | Pill badges of identified strengths (read-only, set by coach) |
+| **Struggles** | Pill badges of identified struggles (read-only, set by coach) |
+| **Save Profile** | Pushes identity + goals changes via `PUT /api/gym/learner-profile` |
+
+#### Dimensions Tab
+
+Current scores across the 5 AI skill dimensions with trend and history.
+
+| Element | Description |
+|---------|-------------|
+| **Dimension bars** | 5 color-coded progress bars (0–10 scale) with score and trend arrow |
+| **Trend indicators** | Up arrow (improving, green), down arrow (declining, red), dot (stable, gray) |
+| **History table** | Last 5 dimension snapshots showing date and scores per dimension |
+
+Data sources: `GET /api/gym/learner-profile` (current scores), `GET /api/gym/dimensions/history` (snapshots).
+
+#### Programs Tab
+
+All training programs with enrollment status and completion progress.
+
+| Element | Description |
+|---------|-------------|
+| **Program cards** | Title, difficulty badge (color-coded), estimated time, target dimensions |
+| **Progress bar** | Shown for started programs — percentage complete with step count |
+| **Current module** | Shows which module the learner is working on |
+| **Description** | Truncated program description (120 chars) |
+
+Data sources: `GET /api/gym/programs` (program list), `GET /api/gym/progress` (completion state).
+
+### Delete Agent
+
+- **Delete Agent button** (red, bottom-left of modal) — only visible when editing
+- **Confirmation overlay:**
+  - Warning: "This action cannot be undone"
+  - Shows agent alias
+  - **Confirmation input** — must type the agent's alias exactly
+  - **Cancel button** — closes without deleting
+  - **Delete Forever button** — disabled until alias matches, then executes delete
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Delete agent | `DELETE /api/agents/:id?confirmAlias=alias` | `delete_agent` |
+| | **Query:** `confirmAlias` (agent alias) | **Params:** `agentId`, `confirmAlias` |
+
+### Save Agent
+
+- **Save button** — "Create Agent" (new) or "Update Agent" (editing)
+  - Collects all fields from all tabs (6 standard + 4 gym-specific if applicable)
+  - Validates required fields (agentId, alias, name)
+  - Creates agent directory, writes CLAUDE.md, updates config.json
+
+## 2.3 Agent Dashboard
+
+**URL:** `/agent-dashboard?id=agentId`
+**Purpose:** At-a-glance view of a single agent — tasks, activity, cost, heartbeat.
+
+### Agent Header
+- **Avatar** — 2-letter initials
+- **Agent name** — large display
+- **Status dot** — green for active
+- **Alias** — @mention format
+- **Description**
+- **Action buttons:**
+  - **← Org** — back to `/org`
+  - **Chat** — opens `/ui#agentId`
+  - **♥ Heartbeat** — trigger manual heartbeat
+    - Shows "Running..." while executing, "Triggered!" on success
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Trigger heartbeat | `POST /api/agents/:id/heartbeat` | `trigger_heartbeat` |
+| | **Body:** `{ triggeredBy? }` | **Params:** `agentId`, `triggeredBy?` |
+
+### Tasks Section
+- **Section title** with count badge
+- **Per task card:**
+  - Status dot (color-coded by status)
+  - Task title
+  - Priority badge (High/Medium/Low with color)
+  - Status label, time ago, assigned by, source agent
+  - **Start button** — moves task to in_progress (if proposed/approved)
+  - **Done button** — marks task complete
+- **Cross-agent tasks** — tasks from other agents shown separately
+- **Empty state:** "No tasks assigned"
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List agent tasks | `GET /api/agents/:id/tasks` | `list_tasks` |
+| | | **Params:** `agentId` |
+| Create task | `POST /api/agents/:id/tasks` | `create_task` |
+| | **Body:** `{ title, description?, priority?, status?, assignedBy?, project? }` | **Params:** `agentId`, `body` (object) |
+| Update task (status change) | `PUT /api/agents/:id/tasks/:taskId` | `update_task` |
+| | **Body:** `{ status?, title?, description?, priority? }` | **Params:** `agentId`, `taskId`, `body` (object) |
+| Delete task | `DELETE /api/agents/:id/tasks/:taskId` | `delete_task` |
+| | | **Params:** `agentId`, `taskId` |
+| Get all tasks (cross-agent) | `GET /api/tasks/all` | `get_all_tasks` |
+| | | *(no params)* |
+| Get task stats | `GET /api/agents/:id/tasks/stats` | `get_task_stats` |
+| | | **Params:** `agentId` |
+| Create project | `POST /api/agents/:id/projects` | `create_project` |
+| | **Body:** `{ name }` | **Params:** `agentId`, `name` |
+
+### Recent Activity Section
+- **Open Chat link** — links to `/ui#agentId`
+- **Activity items** — recent messages with time, direction (IN/OUT), message preview (120 chars)
+- **Empty state:** "No recent activity"
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get activity feed | `GET /api/activity` | `get_activity` |
+| | **Query:** `?limit=N` | **Params:** `limit?` (number) |
+| Get agent logs | `GET /api/agents/:agentId/logs` | `get_agent_logs` |
+| | **Query:** `?limit=N&offset=N&search=text` | **Params:** `agentId`, `limit?`, `offset?`, `search?` |
+
+### Cost Section
+- **3-column stat grid:**
+  - Total Cost ($X.XXXX)
+  - Messages (count)
+  - Average Cost per message ($X.XXXX)
+- **Empty state:** "No cost data"
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get agent cost | `GET /api/agents/:agentId/cost` | `get_agent_cost` |
+| | | **Params:** `agentId` |
+| Get all costs | `GET /api/cost/all` | `get_all_costs` |
+| | | *(no params)* |
+
+### Heartbeat History Section
+- **Count badge** — number of heartbeats
+- **Per heartbeat item** (up to 10 most recent):
+  - Status dot — green (success), red (error), amber (timeout)
+  - Time ago
+  - Duration (ms or s)
+  - Summary (first 100 chars)
+  - Trigger badge — "manual", "schedule", etc.
+- **Empty state:** "No heartbeats yet. Trigger one above."
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get heartbeat history | `GET /api/agents/:id/heartbeat-history` | `get_heartbeat_history` |
+| | **Query:** `?limit=N` | **Params:** `agentId`, `limit?` |
+
+### Wiki Learning
+
+Wiki Learning lets agents automatically capture facts and corrections from conversations into a `learned.md` file. This knowledge accumulates over time and can be merged into the agent's main knowledge base (`context.md`).
+
+**How it works:**
+1. Enable "Wiki Learning" on any agent's config page
+2. The agent will evaluate each conversation and save new facts to `learned.md` in its memory directory
+3. You can manually ask the agent to "update context from learned" to merge verified facts into `context.md`
+
+**Wiki Sync (scheduled):**
+- Optionally enable Wiki Sync to automatically merge learned facts on a schedule
+- Configure via the Schedule tab on the agent config page (same UI as Heartbeat)
+- Default schedule: daily at midnight (`0 0 * * *`)
+- The sync reviews `learned.md`, cross-checks against `context.md`, merges verified facts, and flags contradictions
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Trigger wiki sync | `POST /api/agents/:id/wiki-sync` | `trigger_wiki_sync` |
+| | | **Params:** `agentId` |
+| Get wiki sync history | `GET /api/agents/:id/wiki-sync-history` | `get_wiki_sync_history` |
+| | **Query:** `?limit=N` | **Params:** `agentId`, `limit?` |
+
+---
+
+# 3. Chat
+
+**URL:** `/ui` or `/ui#agentId`
+**Purpose:** Full-screen chat interface with any agent.
+
+The chat interface on `/ui` is functionally identical to the Home chat view (see [Section 1.2](#12-chat-view)) with these differences:
+
+- **URL hash routing** — `/ui#agentId` opens directly to a specific agent
+- **No landing view** — goes straight to chat
+- **Agent drawer** — same sidebar with agent list
+
+All chat actions, APIs, and MCP tools are identical to Home (see [Section 1.2](#12-chat-view)).
+
+### Session Management
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List sessions (low-level) | `GET /api/agents/:agentId/sessions` | `list_sessions` |
+| | | **Params:** `agentId` |
+| Reset session | `POST /api/agents/:agentId/sessions/reset` | `reset_session` |
+| | **Body:** `{ senderId? }` | **Params:** `agentId`, `senderId?` |
+| Delete session | `DELETE /api/agents/:agentId/sessions/:senderId` | `delete_session` |
+| | | **Params:** `agentId`, `senderId` |
+
+### Named Session Tabs (server-side threads)
+
+Each agent supports named, persistent session threads visible in the Web UI tab bar. Sessions survive browser resets and are stored in `agents/<agentId>/memory/session-tabs.json`. Conversation history is replayed from `conversation_log.jsonl` by `tabId`.
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Create/upsert tab | `POST /api/agents/:agentId/session-tabs` | `create_session_tab` |
+| | **Body:** `{ tabId, label, targetAgentId? }` | **Params:** `agentId`, `tabId`, `label`; `targetAgentId` (optional) |
+| List all session tabs | `GET /api/agents/:agentId/session-tabs` | `list_session_tabs` |
+| | Returns tabs sorted newest-first with `lastMessageAt` + `lastPreview` | **Params:** `agentId` |
+| Get conversation history | `GET /api/agents/:agentId/session-tabs/:tabId/history` | `get_session_tab_history` |
+| | Returns `{ messages: [{role, text, time}] }` filtered by tabId | **Params:** `agentId`, `tabId` |
+| Rename tab | `PUT /api/agents/:agentId/session-tabs/:tabId` | `rename_session_tab` |
+| | **Body:** `{ label }` | **Params:** `agentId`, `tabId`, `label` |
+| Delete tab (permanent) | `DELETE /api/agents/:agentId/session-tabs/:tabId` | `delete_session_tab` |
+| | Also clears Claude session file | **Params:** `agentId`, `tabId` |
+
+**Note:** `senderId` in messages sent from the Web UI is the `tabId`. This is what links JSONL log entries to a specific named thread. For agents to maintain separate memory per tab, set `persistent: true` and `perSenderSessions: true` in the agent config.
+
+**`targetAgentId` — cross-agent tab routing:** When a tab is created with `targetAgentId`, all messages sent in that tab are routed to the target agent instead of the host agent. This lets hub (or any agent) create a tab that talks directly to a specialist agent (e.g. `agentcreator`, `skillcreator`) so the sub-agent accumulates its own conversation history and memory. The tab still appears on the host agent's chat page — users don't leave the page, they just switch tabs.
+
+### Model Override (not visible in UI, API/MCP only)
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get current model | `GET /api/agents/:agentId/model` | `get_model` |
+| | | **Params:** `agentId` |
+| Set model override | `PUT /api/agents/:agentId/model` | `set_model` |
+| | **Body:** `{ model }` (opus, sonnet, haiku, or full model ID) | **Params:** `agentId`, `model` |
+| Clear model override | `DELETE /api/agents/:agentId/model` | `clear_model` |
+| | | **Params:** `agentId` |
+
+### Memory (not visible in chat UI, API/MCP only)
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get agent memory | `GET /api/agents/:agentId/memory` | `get_agent_memory` |
+| | **Query:** `?limit=N` | **Params:** `agentId`, `limit?` |
+| Search memory | `POST /api/agents/:agentId/memory/search` | `search_memory` |
+| | **Body:** `{ query }` | **Params:** `agentId`, `query` |
+| Clear memory context | `DELETE /api/agents/:agentId/memory/context` | `clear_memory_context` |
+| | | **Params:** `agentId` |
+
+### Files (not visible in chat UI, API/MCP only)
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List agent files | `GET /api/agents/:agentId/files` | `list_agent_files` |
+| | | **Params:** `agentId` |
+| Download file | `GET /api/agents/:agentId/download?path=filePath` | `download_agent_file` |
+| | **Query:** `path` (absolute file path) | **Params:** `agentId`, `path` |
+| Upload file | `POST /api/upload/:agentId/json` | `upload_file` |
+| | **Body:** `{ fileName, base64Content, mode? }` | **Params:** `agentId`, `fileName`, `base64Content`, `mode?` ("temp" or "permanent") |
+
+### Recover Agent (not visible in UI, API/MCP only)
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Recover from corruption | `POST /api/agents/:agentId/recover` | `recover_agent` |
+| | **Body:** `{ userText?, response? }` | **Params:** `agentId`, `userText?`, `response?` |
+
+---
+
+# 4. Library
+
+**URL:** `/library`
+**Purpose:** Manage your personal library of skills, prompts, apps, MCPs, and agents. Install, assign, create, and publish.
+
+## 4.1 Type Tabs
+
+Five tabs along the top, each showing a different resource type:
+- **Skills** — reusable instruction sets agents can invoke
+- **Prompts** — reusable prompt templates triggered with a character (e.g., `!`)
+- **Apps** — registered web applications
+- **MCPs** — Model Context Protocol servers
+- **Agents** — agent configurations
+
+## 4.2 Search & Filters
+
+- **Search input** — real-time filtering by name, description, or tags
+- **Search clear button (×)** — appears when search has content
+- **Category pills** — dynamic filter buttons:
+  - **"all"** — show everything
+  - **Per-category pills** — auto-generated from item categories
+  - **"★ Platform Defaults" pill** — (skills/prompts/MCPs only) toggle to show only platform defaults
+
+Each tab loads its data from the registry on mount:
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Load Skills/Prompts/Agents/MCPs tab | `GET /api/marketplace/:type` | `browse_registry` |
+| | **URL param:** `type` (skills, prompts, agents, mcps) | **Params:** `type` |
+| Load Apps tab | `GET /api/apps` | `list_apps` |
+| | | *(no params)* |
+| Load MCPs tab (registered servers) | `GET /api/mcps` | `list_mcps` |
+| | | *(no params)* |
+| Load agents (for assign modal) | `GET /api/agents` | `list_agents` |
+| | | *(no params)* |
+
+## 4.3 Personal Items Section ("Built by Me")
+
+### Section Header
+- **Item count** in parentheses
+- **⊕ Add [Type] split button** with dropdown menu:
+
+| Type | Dropdown Options |
+|------|-----------------|
+| Skills | "Scan" (scan folder), "AI Build" (→ Lab), "Marketplace" (→ Marketplace) |
+| Prompts | "Paste" (inline form), "AI Build" (→ Lab), "Marketplace" (→ Marketplace) |
+| Apps | "Register" (inline form), "AI Build" (→ Lab), "Marketplace" (→ Marketplace) |
+| MCPs | "Add" (inline form), "Marketplace" (→ Marketplace) |
+
+### Inline Creation Panels
+
+#### Scan Skills Panel (Skills only)
+- **Directory input** — folder path to scan (default: `~/.claude/commands`)
+- **Scan button** — scans for `.md` skill files
+- **Results** — checkboxes for each found skill with name and description
+- **Import button** — imports selected skills
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Scan for skills | `GET /api/marketplace/scan-skills` | `scan_skills` |
+| | **Query:** `?dir=path` | **Params:** `dir?` (string) |
+| Import scanned skills | `POST /api/marketplace/import-skills` | `import_skills` |
+| | **Body:** `{ agentId, skills[] }` | **Params:** `agentId`, `skills` (string array) |
+| Create skill (programmatic) | `POST /api/skills/create` | `create_skill` |
+| | **Body:** `{ id, name, description, content, scope, orgName?, agentId? }` | **Params:** `id`, `name`, `description`, `content`, `scope` ("global"\|"org"\|"agent"), `orgName?`, `agentId?` |
+
+> Note: The "AI Build" button redirects to Lab (`/lab?build=skill`) where a creator agent builds the skill conversationally. The `create_skill` API/MCP can also create skills directly without the Lab.
+
+#### Create Prompt Panel (Prompts only)
+- **Name input** — prompt display name
+- **ID input** — auto-generated from name
+- **Description input**
+- **Content textarea** — the prompt template text
+- **Save prompt button**
+- **Cancel button**
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Create prompt | `POST /api/marketplace/create-prompt` | `create_prompt` |
+| | **Body:** `{ id, name, content }` | **Params:** `id`, `name`, `content` |
+
+#### Add MCP Panel (MCPs only)
+- **Name input**
+- **ID input** — auto-generated from name
+- **Description input**
+- **Type toggle** — HTTP or stdio buttons
+- **HTTP fields:** URL input
+- **stdio fields:** Command input, Args input (space-separated)
+- **Add MCP button**
+- **Cancel button**
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Add MCP to registry | `POST /api/marketplace/add-mcp` | `add_mcp_to_registry` |
+| | **Body:** `{ name, id, description, type, url?, command?, args? }` | **Params:** `body` (object) |
+
+#### Register App Panel (Apps only)
+- **Name input** (required)
+- **URL input** (required)
+- **Description input**
+- **Category dropdown** — Productivity, Development, Finance, Analytics, Communication, AI/Agents, Other
+- **Status dropdown** — Draft, Live, Maintenance
+- **GitHub Repo input** (optional)
+- **Deploy platform dropdown** — Railway, Vercel, Netlify, Render, Local, Other
+- **Register App button**
+- **Cancel button**
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Create app | `POST /api/apps` | `create_app` |
+| | **Body:** `{ name, url, description?, category?, status?, repo?, deploy? }` | **Params:** `body` (object) |
+
+## 4.4 Item Cards
+
+Each resource is displayed as a card with consistent elements:
+
+### Card Elements
+- **Name and provider**
+- **Badges:**
+  - "✓ verified" — if provider is AgenticLedger/platform
+  - "built by me" — if provider is 'me'
+  - "community" — if external provider
+  - "✓ installed" — if already installed
+  - "★ default" — if set as platform default
+- **Description**
+- **Meta tags** — categories and assigned agents (e.g., "→ agent1, agent2, +1")
+
+### Card Actions
+
+| Button | Description | When Visible |
+|--------|-------------|-------------|
+| **Install** | Install the item | Not yet installed |
+| **Manage** | Open assign modal | Already installed |
+| **☆ Set Default / ★ Default** | Toggle platform default | Skills, Prompts, MCPs |
+| **↑ Publish ▾** | Publish to SaaS (dropdown) | SaaS connected, Skills/Prompts only |
+
+#### Install
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Install item | `POST /api/marketplace/install` | `install_registry_item` |
+| | **Body:** `{ id, type }` | **Params:** `id`, `type` |
+
+After install, the **Assign Modal** opens (see below).
+
+#### Manage / Assign Modal
+- **Title:** "✓ [Item Name] installed" or "Manage — [Item Name]"
+- **Agent checklist** — checkboxes for each configured agent
+- **Missing keys alert** — warns if API keys are needed
+- **Assign selected button** — assigns item to checked agents
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Assign to agent | `POST /api/marketplace/assign` | `assign_to_agent` |
+| | **Body:** `{ agentId, itemId, type }` | **Params:** `agentId`, `itemId`, `type` ("skill", "mcp", "agent") |
+
+#### Set Platform Default
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Set platform default | `POST /api/marketplace/platform-default` | `set_platform_default` |
+| | **Body:** `{ type, id }` | **Params:** `type`, `id` |
+
+#### Publish to SaaS (Dropdown)
+- **"Company Library" button** — publish to SaaS library
+- **"Marketplace" button** — publish to SaaS marketplace
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Publish to SaaS | `POST /api/saas/publish` | `publish_to_saas` |
+| | **Body:** `{ type, id, destination }` | **Params:** `type` ("skill"\|"prompt"\|"agent"\|"app"), `id`, `destination?` ("library"\|"marketplace") |
+
+### App Cards (Special Layout)
+Apps have a richer card:
+- **Initials icon** — colored 2-letter box
+- **App name and URL**
+- **Status badge** — Draft/Live/Maintenance (colored)
+- **Deploy badge** — Railway/Vercel/Netlify/Render/Local/Other
+- **Description**
+- **Category tag and tags**
+- **Agent developer badge** — shows assigned agent
+
+#### App-Specific Actions
+
+| Button | Description |
+|--------|-------------|
+| **↗ Launch** | Opens app URL in new tab |
+| **↑ Publish ▾** | Publish to SaaS (dropdown) |
+| **⌂ GitHub** | Opens GitHub repo (if configured) |
+| **✕ Delete** | Deletes app from registry |
+| **Health chip** | Shows health status; click to check |
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Launch app | Opens `app.url` in new browser tab | N/A — client-side navigation |
+| GitHub link | Opens `app.repo` in new browser tab | N/A — client-side navigation |
+| Delete app | `DELETE /api/apps/:id` | `delete_app` |
+| | | **Params:** `id` |
+| Update app | `PUT /api/apps/:id` | `update_app` |
+| | **Body:** fields to update | **Params:** `id`, `body` (object) |
+| Check app health | `POST /api/apps/:id/check-health` | `check_app_health` |
+| | | **Params:** `id` |
+| Publish app to SaaS | `POST /api/saas/publish` | `publish_to_saas` |
+| | **Body:** `{ type: "app", id, destination }` | **Params:** `type`, `id`, `destination?` |
+
+## 4.5 Prompt Trigger Config (Prompts Tab Only)
+- **Trigger display** — shows current character (e.g., `!`)
+- **✎ Edit button** — prompts for a new 1-character trigger
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get prompt trigger | `GET /api/marketplace/prompt-trigger` | `get_prompt_trigger` |
+| | | *(no params)* |
+| Set prompt trigger | `POST /api/marketplace/prompt-trigger` | `set_prompt_trigger` |
+| | **Body:** `{ trigger }` (single character) | **Params:** `trigger` |
+
+## 4.6 Import from Folder Modal
+- **Step 1:** folder path input + Browse button (opens directory picker) + Scan Folder button
+- **Directory picker:** breadcrumb path, folder listing, up (..) button, Select/Cancel
+- **Step 2:** preview list of found items with type, name, description + Import All button
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Browse directories | `GET /api/browse-dirs` | `browse_dirs` |
+| | **Query:** `?path=dirPath` | **Params:** `path?` |
+
+---
+
+# 5. Lab
+
+**URL:** `/lab`
+**Purpose:** AI-assisted creation of agents, skills, apps, and prompts. Conversational build experience.
+
+## 5.1 Landing View
+
+### Hero Section
+- **Title:** "Lab"
+- **Subtitle:** "What do you want to build?"
+
+### Build Tiles (4 clickable tiles)
+| Tile | Icon | Description |
+|------|------|-------------|
+| **Agent** | Diamond (cyan) | "A purpose-built AI with memory, tools, and channels" |
+| **Skill** | Lightning (green) | "A reusable instruction set any agent can invoke" |
+| **App** | Grid (amber) | "A deployed web application in your platform" |
+| **Prompt** | Star (purple) | "A reusable prompt template triggered with !" |
+
+Clicking a tile opens the **Intake Form** for that type.
+
+### Your Work Section (Artifacts)
+- **Search input** — filter existing artifacts by name
+- **Accordion groups** (click header to expand/collapse, arrow rotates):
+  - **Apps** — each row shows: name, category tag, status badge (Draft/Live/Maintenance)
+    - **Launch button** — opens the app URL in a new tab
+    - **Modify button** — opens the Lab creation view to continue building with the App Creator agent
+    - **Deploy button** — triggers deployment flow
+  - **Agents** — each row shows: name, agent ID, status badge
+  - **Skills** — each row shows: name, category tag, status badge
+  - **Prompts** — each row shows: name, category tag, status badge
+- Each accordion shows a **count badge** with the number of items
+
+## 5.2 Intake Form
+
+Opened by clicking a build tile.
+
+| Field | Description |
+|-------|-------------|
+| **Name** | What you're building (placeholder varies by type) |
+| **Description** | Detailed description of what you want |
+| **Project directory** (Apps only) | Folder path + Browse button with directory picker |
+| **📎 Attach files** | File attachment button + hidden file input |
+| **Create button** | Submits to the appropriate creator agent |
+| **Cancel button** | Returns to landing |
+
+## 5.3 Creator Agents
+
+Each build type is handled by a **hardcoded platform agent** — a real agent with its own ID, system prompt, and tools:
+
+| Type | Agent ID | Name | Alias |
+|------|----------|------|-------|
+| Agent | `agentcreator` | Agent Creator | `@agentcreator` |
+| Skill | `skillcreator` | Skill Creator | `@skillcreator` |
+| App | `appcreator` | App Creator | `@appcreator` |
+| Prompt | `promptcreator` | Prompt Creator | `@promptcreator` |
+
+These are real agents — you can also message them directly from any channel via `@agentcreator`, `@skillcreator`, etc. They are registered as **Platform** class agents.
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List creator agents | `GET /api/platform-agents` | `get_platform_agents` |
+| | Returns list of platform agents with IDs, names, aliases | *(no params)* |
+
+## 5.4 Creation View
+
+The AI-assisted build interface — a split-pane chat + canvas.
+
+### Chat Panel (Left)
+- **Header:** creator agent avatar, name (e.g., "Agent Creator"), alias (e.g., "@agentcreator")
+- **Messages area:** user messages (cyan) and agent messages (purple, markdown-rendered)
+- **Thinking indicator** — animated dots while processing
+- **Tool cards** — show active tool usage
+
+#### Chat Input Controls
+| Control | Description |
+|---------|-------------|
+| **📎 Clip button** | Toggle file drop zone for drag-and-drop |
+| **🎤 Mic button** | Voice input (stub) |
+| **⌘ Raw Logs button** | Toggle raw logs drawer |
+| **Chat textarea** | Enter sends, Shift+Enter newline, auto-resizes |
+| **Send button** | Sends message; becomes "Queue" during streaming |
+| **Stop button** | Appears during streaming; stops active job |
+
+#### Raw Logs Drawer
+- **Mode button** — toggles between "Clean" and "● Raw" modes
+- **Log content area** — live streaming output from the creator agent
+
+### Canvas Panel (Right)
+- **Toggle button (▢ Canvas)** — show/hide canvas
+- **Clear button** — empties canvas
+- **Copy button** — copies all blocks to clipboard
+- **Close button** — hides canvas
+- **Canvas blocks** — code/content artifacts generated by the creator agent
+  - Each block: type label (json, yaml, markdown, typescript, etc.), copy button, content
+- **Empty state:** "Canvas is empty — Artifacts will appear here as they're built"
+- **Resize handle** — draggable divider between chat and canvas
+
+### Deploy Button
+- **Deploy** — disabled until artifacts exist; triggers deployment flow
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get platform agents (creators) | `GET /api/platform-agents` | `get_platform_agents` |
+| | | *(no params)* |
+| Send message to creator | `POST /api/chat/:agentId` | `send_message` |
+| | **Body:** `{ text }` | **Params:** `agentId` (e.g., `agentcreator`), `text` |
+| Start streaming with creator | `POST /api/chat/:agentId/stream` | `start_stream` |
+| | **Body:** `{ text }` | **Params:** `agentId`, `text` |
+| Stop streaming | `POST /api/chat/jobs/:jobId/stop` | `stop_chat_job` |
+| | | **Params:** `jobId` |
+| Poll stream output | `GET /api/chat/jobs/:jobId/raw?after=N` | `get_chat_job_raw` |
+| | **Query:** `after` (line offset) | **Params:** `jobId`, `after?` |
+| Upload file attachment to creator | `POST /api/upload/:agentId` (multipart) | `upload_file` |
+| | Multipart form: `file` + `mode` | **Params:** `agentId`, `fileName`, `base64Content`, `mode?` |
+| Browse dirs (App project picker) | `GET /api/browse-dirs?path=dirPath` | `browse_dirs` |
+| | **Query:** `path` | **Params:** `path?` |
+
+---
+
+# 6. Marketplace
+
+**URL:** `/marketplace`
+**Purpose:** Browse, install, and assign skills, prompts, agents, MCPs, and apps from the registry.
+
+## 6.1 Type Tabs
+
+Same five tabs as Library: **Apps** (default), **Skills**, **Agents**, **MCPs**, **Prompts**
+
+## 6.2 Search & Filters
+
+Identical to Library (see [Section 4.2](#42-search--filters)):
+- Search input with real-time filtering
+- Category pills with "all" and per-category options
+- "★ Platform Defaults" pill (skills/prompts/MCPs only)
+
+## 6.3 Sections
+
+Items are split into two sections:
+- **"Built by Me"** — your personal items (same as Library, with add/create panels)
+- **"Built by Others"** — community and platform items
+
+## 6.4 Item Cards
+
+Same card layout as Library (see [Section 4.4](#44-item-cards) for full card anatomy) with these marketplace-specific behaviors:
+
+### Skills / Prompts / MCPs / Agents Cards
+- **Name, provider, description, badges** — same as Library
+- **Install button** — primary action; downloads item to your local registry
+  - After install, the **Assign Modal** opens automatically to assign to agents
+- **"✓ Installed" badge** — replaces Install button once installed
+- **☆ Set Default / ★ Default** — toggle platform default (skills/prompts/MCPs)
+- **↑ Publish ▾ dropdown** — publish to SaaS (Company Library or Marketplace)
+- **Manage button** — opens assign modal (if already installed)
+
+### App Cards
+- **↗ Launch** — opens app URL in new tab
+- **↑ Publish ▾** — publish to SaaS (Company Library / Marketplace)
+- **⌂ GitHub** — opens repo (if configured)
+- **✕ Delete** — removes app
+- **Health chip** — shows status; click to re-check
+
+All actions, APIs, and MCP tools are identical to Library (see [Section 4.4](#44-item-cards) for full API/MCP tables):
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Browse registry by type | `GET /api/marketplace/:type` | `browse_registry` |
+| | **URL param:** `type` (skills, prompts, agents, mcps, apps) | **Params:** `type` |
+| Install item | `POST /api/marketplace/install` | `install_registry_item` |
+| | **Body:** `{ id, type }` | **Params:** `id`, `type` |
+| Assign to agent | `POST /api/marketplace/assign` | `assign_to_agent` |
+| | **Body:** `{ agentId, itemId, type }` | **Params:** `agentId`, `itemId`, `type` |
+| Set platform default | `POST /api/marketplace/platform-default` | `set_platform_default` |
+| | **Body:** `{ type, id }` | **Params:** `type`, `id` |
+
+---
+
+# 7. Admin
+
+**URL:** `/admin`
+**Purpose:** System-level configuration — channels, activity logs, accounts, service settings, SaaS, and deployment.
+
+## 7.1 Channels
+
+**Tab:** Channels
+
+### Per Channel Card
+One card per messaging channel (Telegram, Slack, iMessage, WhatsApp, Discord). Each shows:
+
+| Element | Description |
+|---------|-------------|
+| **Channel icon** | Emoji for channel type |
+| **Channel name** | e.g., "Telegram" |
+| **Status pill** | "Connected" or "Disabled" |
+| **Sticky mode select** | None, Sticky, Prefix — controls routing behavior |
+| **Prefix input** | Only visible if Prefix mode; the prefix string |
+| **Timeout input** | Minutes before sticky routing expires |
+| **Save Settings button** | Saves channel configuration |
+
+### Monitored Chat IDs (iMessage only)
+- **List of monitored chat IDs** with remove buttons per entry
+- **Add input** — chat ID to add
+- **Add button** — adds the chat ID
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Add monitored chat | `POST /api/channels/:channelName/monitored` | `add_monitored_chat` |
+| | **Body:** `{ chatId }` | **Params:** `channelName`, `chatId` |
+| Remove monitored chat | `DELETE /api/channels/:channelName/monitored` | `remove_monitored_chat` |
+| | **Body:** `{ chatId }` | **Params:** `channelName`, `chatId` |
+
+### Chat Groups (Agent Routing)
+- **Filter dropdown** — filter by chat ID
+- **Agent list** — grouped by chat ID, showing agent name, alias, remove button
+- **Add Agent section:**
+  - Agent select dropdown
+  - Chat/Channel ID input
+  - **Add Agent button**
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List channels | `GET /api/channels` | `list_channels` |
+| | | *(no params)* |
+| Update channel settings | `PUT /api/channels/:channelName` | `update_channel` |
+| | **Body:** `{ stickyMode?, prefix?, timeout?, enabled? }` | **Params:** `channelName`, `body` (object) |
+| Add agent route | `POST /api/channels/:channelName/agents` | `add_agent_route` |
+| | **Body:** `{ agentId, chatId?, requireMention? }` | **Params:** `channelName`, `body` (object) |
+| Remove agent route | `DELETE /api/channels/:channelName/agents/:agentId` | `remove_agent_route` |
+| | | **Params:** `channelName`, `agentId` |
+| Get sticky routing | `GET /api/sticky-routing` | `get_sticky_routing` |
+| | | *(no params)* |
+
+## 7.2 Activity
+
+**Tab:** Activity
+
+### Filters
+- **Agent filter dropdown** — dynamically populated with all agents
+- **Channel filter dropdown** — All Channels, Telegram, Slack, iMessage, Discord, WhatsApp
+- **Search input** — searches message content (debounced)
+- **Result count** — "X entries"
+
+### Activity Entries
+Each entry shows:
+- Agent avatar and name (clickable link to chat)
+- Channel tag
+- Session indicator
+- Timestamp
+- "Chat →" button to view in chat
+- Query text (first 200 chars)
+- Response preview (first 800 chars, expandable on click)
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get activity feed | `GET /api/activity` | `get_activity` |
+| | **Query:** `?limit=N` | **Params:** `limit?` |
+| Get agent logs | `GET /api/agents/:agentId/logs` | `get_agent_logs` |
+| | **Query:** `?limit=N&offset=N&search=text` | **Params:** `agentId`, `limit?`, `offset?`, `search?` |
+
+## 7.3 Settings
+
+**Tab:** Settings (default)
+
+### Profile Section
+- **Section label:** "Profile" (cyan highlight)
+- **Description:** "Your identity — agents can use this to personalize responses"
+- **Fields:**
+
+| Field | Description |
+|-------|-------------|
+| **Name** | Your name |
+| **Role / Title** | Your job role (e.g., product manager, developer, student) |
+| **Industry** | Your industry or domain (e.g., fintech, healthcare, education) |
+| **AI Experience** | Dropdown: Beginner, Intermediate, Advanced |
+| **Interests** | Comma-separated list of focus areas (e.g., automation, coding, writing) |
+
+- **Save button** — persists profile to `profile.json`
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get profile | `GET /api/profile` | `get_profile` |
+| | | *(no params)* |
+| Update profile | `PUT /api/profile` | `update_profile` |
+| | **Body:** `{ name?, role?, industry?, aiExperience?, interests? }` | **Params:** `name?`, `role?`, `industry?`, `aiExperience?`, `interests?` |
+
+### Claude Accounts Section
+
+#### Account List
+Each account row shows:
+- **Status dot** — green (authenticated), red (error), grey (unknown)
+- **Account name**
+- **Email address** (fetched from account status)
+- **Account path** (e.g., `~/.claude-main`)
+- **↻ Re-login button** — restart OAuth for this account
+- **✕ Delete button** — remove account
+
+#### Add Account
+- **Account name input**
+- **Login & Add button (→)** — starts OAuth flow
+- **Login panel** (appears after starting):
+  - Step 1: Login URL display (copyable, clickable)
+  - Step 2: Code input field + **Submit Code button**
+  - **Verify & Save button**
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List accounts | `GET /api/config/accounts` | `list_accounts` |
+| | | *(no params)* |
+| Add account | `POST /api/config/accounts` | `add_account` |
+| | **Body:** `{ name, path }` | **Params:** `name`, `path` |
+| Delete account | `DELETE /api/config/accounts/:name` | `delete_account` |
+| | | **Params:** `name` |
+| Check account status | `GET /api/config/accounts/:name/status` | `check_account_status` |
+| | | **Params:** `name` |
+| Start login | `POST /api/config/accounts/login` | `start_account_login` |
+| | **Body:** `{ name, path }` | **Params:** `name`, `path` |
+| Submit login code | `POST /api/config/accounts/login/code` | `submit_login_code` |
+| | **Body:** `{ accountName, code }` | **Params:** `accountName`, `code` |
+| Check auth status | `GET /api/whoami/:agentId` | `whoami` |
+| | | **Params:** `agentId` |
+
+### Service Section
+- **Note:** "Restart required for most changes"
+- **Settings grid:**
+
+| Field | Description |
+|-------|-------------|
+| **Personal Agents Directory** | Where agent folders are stored |
+| **Personal Registry Directory** | Where skills/prompts registry lives |
+| **Web UI Port** | Port for the web interface (default: 4888) |
+| **Log Level** | Logging verbosity |
+
+- **Save button** — saves service configuration
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get service config | `GET /api/config/service` | `get_service_config` |
+| | | *(no params)* |
+| Update service config | `PUT /api/config/service` | `update_service_config` |
+| | **Body:** `{ personalAgentsDir?, personalRegistryDir?, port?, logLevel?, multiModelEnabled?, platformDefaultExecutor?, ollamaBaseUrl? }` | **Params:** `body` (object) |
+
+### AI Models Section
+- **Section label:** "AI Models" (amber highlight)
+- **Toggle:** Enable Multi-Model — enables/disables alternative model routing (Ollama + cloud providers)
+- **Setup guide** (shown when enabled): Platform-specific install instructions for Ollama (macOS, Windows, Linux), model pull commands, and a "Test Connection" button that checks Ollama availability and lists downloaded models
+- **Settings grid:**
+
+| Field | Description |
+|-------|-------------|
+| **Platform Default Executor** | Default executor for all agents: `claude` (default), `ollama:<model>`, or any supported provider:model combo |
+| **Ollama Base URL** | Ollama API endpoint (default: `http://localhost:11434`) |
+
+- **Save button** — saves AI model settings
+
+#### Supported Providers
+
+| Provider | Executor format | API key field | Example |
+|----------|----------------|---------------|---------|
+| **Claude** | `claude` (default) | — (uses local CLI auth) | `claude` |
+| **Ollama** | `ollama:<model>` | — (local, no key) | `ollama:gemma2`, `ollama:llama3.1` |
+| **OpenAI** | `openai:<model>` | `openai` | `openai:gpt-4o`, `openai:gpt-4o-mini` |
+| **Grok (xAI)** | `grok:<model>` | `xai` | `grok:grok-3`, `grok:grok-3-mini` |
+| **Gemini** | `gemini:<model>` | `google` | `gemini:gemini-2.5-flash`, `gemini:gemini-2.5-pro` |
+| **Groq** | `groq:<model>` | `groq` | `groq:llama-3.3-70b-versatile` |
+| **Together** | `together:<model>` | `together` | `together:meta-llama/Llama-3.3-70B-Instruct` |
+| **Mistral** | `mistral:<model>` | `mistral` | `mistral:mistral-large-latest` |
+
+API keys are stored in `config.json` under `service.providerKeys` (e.g., `{ "openai": "sk-...", "xai": "xai-...", "google": "AIza..." }`). Set them in Admin → Settings → Provider Keys.
+
+- **Limitations:** Non-Claude agents (Ollama + cloud providers) get text-in/text-out only — no tool use (Read, Write, Bash), no MCP tools, no persistent sessions. Good for Q&A, content generation, and advisory roles.
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Enable multi-model | `PUT /api/config/service` | `update_service_config` |
+| | **Body:** `{ multiModelEnabled: true }` | **Params:** `{ multiModelEnabled: true }` |
+| Set platform default | `PUT /api/config/service` | `update_service_config` |
+| | **Body:** `{ platformDefaultExecutor: "openai:gpt-4o" }` | **Params:** `{ platformDefaultExecutor: "openai:gpt-4o" }` |
+| Set Ollama URL | `PUT /api/config/service` | `update_service_config` |
+| | **Body:** `{ ollamaBaseUrl: "http://host:11434" }` | **Params:** `{ ollamaBaseUrl: "http://host:11434" }` |
+| Set provider API keys | `PUT /api/config/service` | `update_service_config` |
+| | **Body:** `{ providerKeys: { openai: "sk-...", xai: "xai-..." } }` | **Params:** `{ providerKeys: { ... } }` |
+| Test provider API key | `POST /api/config/provider-test` | `test_provider` |
+| | **Body:** `{ provider: "openai" }` — tests the configured key for that provider | **Params:** `provider` (string: openai, grok, gemini, groq, together, mistral) |
+| Proxy Ollama tag list | `GET /api/ollama-proxy?url=<ollamaUrl>/api/tags` | — (internal) |
+| | **Query:** `url` — only `/api/tags` proxy is allowed | N/A — used by the Settings UI to list local Ollama models |
+
+### Deployment Section
+- **Section label:** "Deployment" (green highlight)
+- **Note:** "Restart required for most changes"
+- **Settings grid:**
+
+| Field | Description |
+|-------|-------------|
+| **Provider** | Deployment platform (e.g., "railway") |
+| **Deploy Token** | Platform deploy token (password field with reveal toggle) |
+| **GitHub Org/User** | GitHub organization or username for deployments |
+| **GitHub Token** | GitHub personal access token (password field with reveal toggle) |
+
+- **Save button** — saves deployment configuration
+- **Reveal toggle buttons (👁)** — show/hide password fields for Deploy Token and GitHub Token
+
+> Note: Deployment settings are stored in `config.json` and do not currently have dedicated API/MCP endpoints. They are managed through the `update_service_config` API.
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get deployment config | `GET /api/config/service` | `get_service_config` |
+| | Returns service config including deployment settings | *(no params)* |
+| Update deployment config | `PUT /api/config/service` | `update_service_config` |
+| | **Body:** `{ deployment: { provider?, deployToken?, githubOrg?, githubToken? } }` | **Params:** `body` (object) |
+
+### SaaS Publishing Section
+- **Description:** "Publish skills, prompts, agents, and apps from your Library to a shared SaaS workspace"
+- **Status dot** — green when connected, hidden otherwise
+- **Settings grid:**
+
+| Field | Description |
+|-------|-------------|
+| **SaaS Base URL** | The SaaS platform URL |
+| **API Key** | SaaS API key (password field with reveal toggle) |
+
+- **Save button** — saves SaaS config
+- **Test Connection button** — tests credentials
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get SaaS config | `GET /api/saas/config` | `get_saas_config` |
+| | | *(no params)* |
+| Update SaaS config | `PUT /api/saas/config` | `update_saas_config` |
+| | **Body:** `{ baseUrl?, apiKey? }` | **Params:** `baseUrl?`, `apiKey?` |
+| Test SaaS connection | `POST /api/saas/test` | `test_saas_connection` |
+| | **Body:** `{ baseUrl?, apiKey? }` | **Params:** `baseUrl?`, `apiKey?` |
+
+### Status Indicator Section
+- **Description:** "Show a live status dot in your menu bar (Mac) or system tray (Windows)"
+- **Install xbar Plugin button** (macOS only) — installs the status bar indicator
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Install xbar plugin | `POST /api/install-xbar` | `install_xbar` |
+| | | *(no params)* |
+
+### Authorized Senders (Pairing)
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List paired senders | `GET /api/pairing` | `list_paired_senders` |
+| | | *(no params)* |
+| Pair sender | `POST /api/pairing` | `pair_sender` |
+| | **Body:** `{ senderKey }` | **Params:** `senderKey` |
+| Unpair sender | `DELETE /api/pairing/:senderKey` | `unpair_sender` |
+| | | **Params:** `senderKey` |
+
+## 7.4 Docs
+
+**Tab:** Docs
+
+Documentation cards:
+
+| Card | Description | Action |
+|------|-------------|--------|
+| **API Docs** | Full REST API reference — endpoints, schemas, auth | External link (↗) |
+| **MCP Tools** | MCP tool definitions — available actions via MCP | External link (↗) |
+| **Changelog** | Release history — features, fixes, breaking changes | In-app link (→) |
+| **User Guide** | This document — comprehensive platform walkthrough | In-app link (→) — **to be added** |
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get changelog | `GET /api/changelog` | `get_changelog` |
+| | | *(no params)* |
+
+---
+
+# 8. Monitor
+
+**URL:** `/monitor`
+**Purpose:** At-a-glance platform health and first-time setup checklist. This is the first page new users see after running `/setup`.
+
+## 8.1 Setup Checklist
+
+The **Getting Started** card appears at the top of the page. It tracks three onboarding milestones and auto-updates as you complete them.
+
+| Element | Description |
+|---------|-------------|
+| **Connect a messaging channel** | Turns green when at least one channel (Telegram, Slack, WhatsApp, iMessage) is connected |
+| **Create your first agent** | Turns green when at least one personal (non-platform) agent exists |
+| **Send a test message** | Turns green when any personal agent has received at least one message |
+| **Start Setup button** | Navigates to `/?setup=true`, which opens the hub chat and auto-triggers the `/onboarding` skill to walk through channel + agent setup |
+| **"Setup complete!" badge** | Appears and the checklist collapses once all three items are done |
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get dashboard data (drives checklist) | `GET /api/dashboard` | `get_dashboard` |
+
+## 8.2 Dashboard
+
+Below the setup checklist, a 2-column dashboard grid shows live platform metrics. Data auto-refreshes every 30 seconds.
+
+### Platform Status Card
+
+| Stat | Description |
+|------|-------------|
+| **Uptime** | How long the service has been running (hours + minutes) |
+| **Total Agents** | Count of all registered agents (platform + personal) |
+| **Personal Agents** | Count of non-platform agents (excludes hub, creators) |
+| **Organizations** | Distinct organizations across all agents |
+| **Channels** | Pills showing each connected channel name |
+| **MCPs** | Number of MCP servers in the registry |
+| **Active Automations** | Sum of active goals + active cron jobs across personal agents |
+| **Open Tasks** | Sum of approved + in-progress + review tasks across personal agents |
+
+### Agent Activity Card
+
+Shows personal agents sorted by most recent activity.
+
+| Element | Description |
+|---------|-------------|
+| **Search input** | Filter agents by name or alias |
+| **Filter pills** | Toggle filters: MCPs, Skills, Automations, Open Tasks. Filters combine with AND logic. |
+| **Agent rows** | Each row shows: avatar (initials), name, @alias, badge counts (MCPs, skills, automations, tasks), message count, and time since last message |
+| **Click an agent row** | Navigates to `/ui?agent={agentId}` to chat with that agent |
+
+### Recent Activity Card (full width)
+
+A chronological feed of the last 15 messages across all agents.
+
+| Element | Description |
+|---------|-------------|
+| **Time column** | Timestamp (HH:MM) |
+| **Agent column** | Agent that handled the message |
+| **Message column** | First 80 characters of the message. User messages shown in accent blue, agent responses in purple. |
+
+### Costs Card (full width, conditional)
+
+Only appears when cost data is available from `GET /api/costs`.
+
+| Element | Description |
+|---------|-------------|
+| **Period rows** | Shows cost per time period (today, this week, this month) with dollar amounts |
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get dashboard data | `GET /api/dashboard` | `get_dashboard` |
+| Get recent activity | `GET /api/activity?limit=15` | `get_activity` |
+| Get cost data | `GET /api/costs` | `get_all_costs` |
+
+---
+
+# 9. Projects
+
+Cross-agent initiatives that bundle tasks, agents, orgs, apps, and artifacts under a single trackable entity. Unlike tasks (finite, single-agent), projects are long-running containers that group related work across your entire platform.
+
+**Route:** `/projects`
+
+## 9.1 Projects List
+
+A responsive card grid showing all projects. Header displays active/paused/completed counts with a **+ New Project** button.
+
+### Filter Bar
+
+| Element | Description |
+|---------|-------------|
+| **Search** | Text search filtering by project name or description |
+| **Status pills** | Filter by All, Active, Paused, Completed, Archived |
+
+### Project Cards
+
+| Element | Description |
+|---------|-------------|
+| **Name** | Project name (display font) |
+| **Status badge** | Color-coded: active (green), paused (amber), completed (accent), archived (muted) |
+| **Owner badge** | Agent ID that owns the project |
+| **Team count** | Number of team member agents |
+| **Description** | Truncated to 2 lines |
+| **Progress bar** | Based on linked task completion percentage |
+| **Task stats** | "X/Y tasks done" |
+| **Linked counts** | Agents, apps, artifacts counts |
+| **Created date** | When the project was created |
+| **Click** | Opens the project detail panel |
+
+## 9.2 Project Detail Panel
+
+Slide-in panel showing full project details with editable fields.
+
+| Element | Description |
+|---------|-------------|
+| **Name** | Editable project name |
+| **Description** | Editable description |
+| **Status** | Dropdown: active, paused, completed, archived |
+| **Owner** | Dropdown of all agents |
+| **Team Members** | Multi-select checkboxes of agents |
+| **Plan** | Editable markdown textarea |
+| **Notes** | Editable freeform textarea |
+| **Linked Tasks** | List with status dots, link/unlink controls |
+| **Linked Agents** | List with unlink controls |
+| **Linked Orgs** | List with unlink controls |
+| **Linked Apps** | List with unlink controls |
+| **Linked Artifacts** | List with name + type, unlink controls |
+| **Save button** | Saves all changes |
+| **Delete button** | Deletes the project |
+| **Link entity** | Type selector + value input to add new links |
+
+## 9.3 Create Project Modal
+
+Form for creating a new project.
+
+| Field | Description |
+|-------|-------------|
+| **Name** | Required — project name |
+| **Description** | What this project is about |
+| **Owner** | Agent dropdown (defaults to hub) |
+| **Team Members** | Checkboxes for team agent selection |
+| **Plan** | Markdown textarea for the project plan |
+| **Notes** | Additional notes |
+
+## 9.4 API & MCP Reference
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List all projects | `GET /api/projects` | `list_projects` |
+| Get project detail | `GET /api/projects/:id` | `get_project` |
+| Create project | `POST /api/projects` | `create_initiative` |
+| Update project | `PUT /api/projects/:id` | `update_project` |
+| Delete project | `DELETE /api/projects/:id` | `delete_project` |
+| Link entity to project | `POST /api/projects/:id/link` | `link_to_project` |
+| Unlink entity from project | `POST /api/projects/:id/unlink` | `unlink_from_project` |
+| Get project status report | `GET /api/projects/:id/status` | `get_project_status` |
+| Start autonomous execution | `POST /api/projects/:id/execute` | `execute_project` |
+| Pause autonomous execution | `POST /api/projects/:id/pause` | `pause_project` |
+
+### Autonomous Execution
+
+Projects can execute autonomously in the background. When you call `execute_project`, it creates a scheduled goal on the owner agent that:
+1. Reads the project status and finds the next undone task
+2. Marks it in-progress, executes it, marks it done
+3. Continues to the next task
+4. Reports to the owner agent's Slack channel when complete or blocked
+
+Options: `schedule` (cron, default every 15 min), `reportTo` (channel:chatId), `budget` (maxDailyUsd, default $5).
+
+Use `pause_project` to stop execution. The goal is disabled but preserved.
+
+---
+
+# 10. AI Gym
+
+Personal AI training area with a coach agent, structured programs, dimension tracking, and personalized learning plans. Gated behind `gymEnabled: true` in settings.
+
+**Route:** `/gym`
+
+## 10.1 Onboarding (First Visit)
+
+Full-screen 3-step flow shown when `onboardingComplete: false`. Must be completed before accessing the main gym.
+
+### Step 1 — Choose Your Trainer
+
+| Element | Description |
+|---------|-------------|
+| **Trainer cards** | Grid of 5 trainer profiles, each with avatar, name, style, and intro quote |
+| **Alex** | Collaborative · Steady — works alongside you |
+| **Jordan** | Direct · Steady — holds you accountable |
+| **Morgan** | Thoughtful · Steady — explains the "why" |
+| **Riley** | Challenging · Immersive — pushes you hard |
+| **Sam** | Patient · Steady — meets you where you are |
+| **Next button** | Enabled after selecting a trainer |
+
+### Step 2 — Get to Know You
+
+| Element | Description |
+|---------|-------------|
+| **Coach header** | Selected trainer's avatar and name |
+| **Coach message** | One-question prompt asking about professional background, AI experience, and goals |
+| **Text input** | Free-form textarea for response |
+| **Send button** | Sends response to gym agent via `POST /api/chat/gym/stream` |
+
+### Step 3 — Get Your Plan
+
+| Element | Description |
+|---------|-------------|
+| **Loading spinner** | Shown while coach generates the plan |
+| **Plan summary** | Two-bucket display: On-the-Job Training + Platform-Driven Learning |
+| **Enter the Gym** | Marks onboarding complete and enters main gym |
+
+## 10.2 Mode Picker (Landing)
+
+After onboarding, returning users see a **mode picker** instead of jumping straight into chat. Three tiles answer "what brings you to the gym?"
+
+| Tile | Title | Description |
+|------|-------|-------------|
+| **1** | **I have work to do** | Bring a real task. Coach helps you execute while teaching along the way. Chat opens with "Describe what you need to get done..." placeholder. |
+| **2 (Recommended)** | **You tell me** | Coach picks what to work on based on activity, dimension gaps, and unused features. Visually emphasized with accent border. Auto-sends a coaching request — coach arrives with one specific recommendation. |
+| **3** | **I want to get smart** | Self-directed learning. Pick a topic or program. If in-progress program exists, coach offers to continue it. Otherwise prompts for topic. |
+
+Additional elements:
+- **Time-aware greeting** — "Good morning/afternoon/evening. What brings you to the gym?"
+- **Streak counter** — shown below tiles if streak > 0
+
+Clicking a tile sets the **session mode** (task/coach/learn), transitions to the main gym view, and primes the coach chat with mode-appropriate context.
+
+## 10.3 Main Gym View
+
+Three-panel layout: left sidebar, center panel (tabbed), bottom feed strip.
+
+### Top Bar
+
+| Element | Description |
+|---------|-------------|
+| **Configure button** | Opens the agent editor modal (same as Org page) in an iframe overlay, pre-loaded with the gym agent. Shows all 6 standard tabs plus 4 gym-specific tabs (Trainer, Learner, Dimensions, Programs). See section 2.2 for tab details. |
+| **Gym nav link** | Returns to gym home |
+| **Theme toggle** | Light/dark mode switch |
+
+### Left Sidebar
+
+| Element | Description |
+|---------|-------------|
+| **Streak counter** | Fire icon with current day streak count |
+| **Programs list** | All programs with status indicators: ✓ (done), N/M (in progress), ○ (not started) |
+| **Gym cards** | 2–3 cards with title, description, CTA button, and dismiss (×) control |
+| **"New" badge** | Shown on cards generated by the latest digest |
+
+### Center Panel — Coach Tab
+
+| Element | Description |
+|---------|-------------|
+| **Coach header** | Trainer name + avatar + style subtitle |
+| **Chat messages** | Scrollable conversation area (user right-aligned, coach left-aligned) |
+| **Chat input** | Textarea with Enter-to-send + send button. Routes to `POST /api/chat/gym/stream` |
+| **Typing indicator** | Three bouncing dots shown while coach is responding |
+| **SSE streaming** | Responses stream in real-time via `GET /api/chat/jobs/:jobId/stream` |
+
+### Center Panel — Progress Tab
+
+| Element | Description |
+|---------|-------------|
+| **Radar chart** | SVG spider chart showing 5 dimensions at current scores (0–5) |
+| **Dimension breakdown** | Row per dimension: label, progress bar, score/5, trend arrow (↑ ↓ →) |
+| **Next focus** | Callout highlighting the weakest dimension with guidance text |
+| **Milestones** | Chronological list of achievements (program completions, level-ups) |
+
+### Bottom Feed Strip
+
+Placeholder for P1. Displays "Feed coming in P1".
+
+## 10.4 AI Strength Dimensions
+
+Learners are scored across 5 dimensions on a **1–5 scale** (0 = not yet assessed):
+
+| Dimension | What It Measures |
+|-----------|-----------------|
+| **Application** | How deeply AI is integrated into actual work — usage frequency, breadth, habit |
+| **Communication** | Prompt effectiveness — context, specificity, iteration efficiency |
+| **Knowledge** | Conceptual understanding of AI, agents, tools, MCPs, memory |
+| **Orchestration** | Multi-agent workflows — goals, cron, delegation, automation |
+| **Craft** | Building AI systems — agent creation, system prompts, tool/MCP config |
+
+Score labels: 1 = Beginner, 2 = Developing, 3 = Proficient, 4 = Advanced, 5 = Expert.
+
+## 10.5 Programs
+
+Structured training curricula. Each program has modules, and each module has steps with verification.
+
+### Getting Started (Bundled)
+
+Pre-installed beginner program with 3 modules and 7 steps:
+- **Meet Your AI** — What is an agent? + Have your first conversation
+- **Give It Context** — Prompt quality + Try a detailed prompt + Upload a file
+- **Make It Yours** — Create a specialized agent + Write a system prompt
+
+### Step Verification Types
+
+| Type | How It Works |
+|------|-------------|
+| **knowledge** | Coach asks comprehension questions, evaluates understanding |
+| **platform-check** | Coach calls MCP tools to verify platform state changed (e.g., new agent created) |
+| **self-report** | User describes what they did and learned |
+
+## 10.6 Activity Digest
+
+Scheduled daily at 6am (when `gymEnabled: true`). Also triggerable manually via `POST /api/gym/digest/run`.
+
+What it does:
+1. Scans all agent activity summaries
+2. Scores 5 dimensions using observed patterns
+3. Snapshots dimension history weekly
+4. Updates learner profile (activity, features, streak, dimensions)
+5. Writes daily journal to `agents/platform/gym/memory/daily/<date>.md`
+6. Generates 2–3 gym cards (weakest dimension tips, dormant agent nudges, unused feature discovery)
+
+## 10.7 Trainer Souls
+
+Each trainer is a `soul.md` file prepended to the gym agent's system prompt at spawn time. Switching trainers changes the coaching voice without affecting progress, plan, or profile. Switch anytime via the **Trainer tab** in the gym agent's Configure modal (top bar → Configure → Trainer tab).
+
+Avatar SVGs at `public/trainers/{alex,jordan,morgan,riley,sam}.svg`.
+
+## 10.8 Coach-Created Guides
+
+After any substantive coaching session, the gym agent auto-generates a reusable guide and asks the user to review it.
+
+| Element | Description |
+|---------|-------------|
+| **Guide generation** | Coach distills session into clean step-by-step instructions via `create_gym_guide` MCP tool |
+| **Review prompt** | Coach asks "I wrote up a guide from what we just did — want to review it?" |
+| **Approval** | User can edit, approve, or discard before saving |
+| **Where guides appear** | Gym sidebar → Guides tab → "Coach Created" filter |
+| **Skill publishing** | Guides can also be published as agent-executable skills via `create_skill` |
+
+Guides have two forms: human-readable (browsable in gym) and agent-executable (assignable as skills).
+
+| Action | API | MCP |
+|--------|-----|-----|
+| List coach guides | `GET /api/gym/guides` | `list_gym_guides` |
+| Create coach guide | `POST /api/gym/guides` | `create_gym_guide` |
+| | **Body:** `{ title, description?, difficulty?, dimensions?, estimatedTime?, modules?, content? }` | **Params:** `title`, `description?`, `difficulty?`, `dimensions?`, `estimatedTime?`, `modules?`, `content?` |
+
+## 10.9 API & MCP Reference
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Get learner profile | `GET /api/gym/learner-profile` | `get_learner_profile` |
+| Update learner profile | `PUT /api/gym/learner-profile` | `update_learner_profile` |
+| | **Body:** any profile fields (merged) | **Params:** `data` (object) |
+| Get training plan | `GET /api/gym/plan` | `get_plan` |
+| Update training plan | `PUT /api/gym/plan` | `update_plan` |
+| | **Body:** full plan object | **Params:** `data` (object) |
+| Get program progress | `GET /api/gym/progress` | `get_gym_progress` |
+| Update program progress | `PUT /api/gym/progress` | `update_gym_progress` |
+| | **Body:** progress data | **Params:** `data` (object) |
+| List gym cards | `GET /api/gym/cards` | `list_gym_cards` |
+| Create gym card | `POST /api/gym/cards` | `create_gym_card` |
+| | **Body:** `{ title, description, type, cta?, ctaAction? }` | **Params:** `title`, `description`, `type` |
+| Dismiss gym card | `DELETE /api/gym/cards/:id` | `dismiss_gym_card` |
+| | | **Params:** `id` |
+| Snapshot dimensions | `POST /api/gym/dimensions/snapshot` | `snapshot_dimensions` |
+| | **Body:** `{ date?, dimensions }` | **Params:** `date?`, `dimensions` |
+| Get dimension history | `GET /api/gym/dimensions/history` | `get_dimension_history` |
+| List programs | `GET /api/gym/programs` | `list_gym_programs` |
+| Get program detail | `GET /api/gym/programs/:slug` | `get_gym_program` |
+| | | **Params:** `slug` |
+| Create program | `POST /api/gym/programs` | -- |
+| | **Body:** `{ title, slug?, description, difficulty, modules[] }` | |
+| Update program | `PATCH /api/gym/programs/:slug` | `update_gym_program` |
+| | **Body:** fields to update | **Params:** `slug`, `body` |
+| Delete program | `DELETE /api/gym/programs/:slug` | `delete_gym_program` |
+| | | **Params:** `slug` |
+| Import program from markdown | `POST /api/gym/programs/import-markdown` | `import_program` |
+| | **Body:** `{ markdown }` (H1=program, H2=module, H3=step) | **Params:** `markdown` |
+| Get agent activity summary | `GET /api/agents/:id/activity-summary` | `get_agent_activity_summary` |
+| | | **Params:** `agentId` |
+| Get agent logs (paginated) | `GET /api/agents/:id/logs?limit=50&offset=0` | `get_agent_logs` |
+| | | **Params:** `agentId`, `limit?`, `offset?` |
+| Search agent logs | `GET /api/agents/logs/search?q=keyword&agentIds=a,b` | `search_agent_logs` |
+| | | **Params:** `q`, `agentIds?` |
+| Run activity digest | `POST /api/gym/digest/run` | `run_gym_digest` |
+| Get gym feed | `GET /api/gym/feed` | `get_gym_feed` |
+| | Returns tips, platformUpdates, briefing | *(no params)* |
+| Get gym config flags | `GET /api/gym/config` | `get_gym_config` |
+| | Returns gymEnabled, gymOnlyMode, aibriefingEnabled | *(no params)* |
+| List coach-created guides | `GET /api/gym/guides` | `list_gym_guides` |
+| | Returns array of programs with source=coach | *(no params)* |
+| Create coach guide | `POST /api/gym/guides` | `create_gym_guide` |
+| | **Body:** `{ title, description?, difficulty?, dimensions?, estimatedTime?, modules?, content? }` | **Params:** `title`, `description?`, `difficulty?`, `dimensions?`, `estimatedTime?`, `modules?`, `content?` |
+
+---
+
+# Appendix A: Global Navigation
+
+Present on every page:
+
+| Element | Description |
+|---------|-------------|
+| **MyAIforOne logo** | Home link |
+| **Agents tab** | → `/org` |
+| **Chat tab** | → `/ui` |
+| **Library tab** | → `/library` |
+| **Projects tab** | → `/projects` |
+| **Lab tab** | → `/lab` |
+| **Gym tab** | → `/gym` (only visible when `gymEnabled: true`) |
+| **Marketplace link** | → `/marketplace` |
+| **Admin button (⚙)** | → `/admin` |
+| **User Guide button** | → `/user-guide` — this document |
+| **Monitor button** | → `/monitor` — setup checklist + live dashboard |
+| **Mini Bar button** | Opens a compact floating popup window (440×460) at `/mini`. A lightweight chat interface you can keep open while working — has agent selection, @mention, send, and basic chat. Useful as a quick-access sidebar. |
+| **Theme toggle** | Switches between light and dark mode. Persists across sessions via localStorage. |
+
+---
+
+# Appendix B: Dashboard & Health
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Full dashboard | `GET /api/dashboard` | `get_dashboard` |
+| | Returns all agents, channels, accounts, uptime | *(no params)* |
+| Health check | `GET /health` | `health_check` |
+| | Returns `{ status: "ok" }` | *(no params)* |
+
+---
+
+# Appendix C: Webhook
+
+External systems can trigger agent messages via webhook:
+
+| Action | API | MCP |
+|--------|-----|-----|
+| Send webhook | `POST /webhook/:agentId` | `send_webhook` |
+| | **Headers:** `x-webhook-secret: secret` (optional) | **Params:** `agentId`, `text`, `secret?`, `channel?`, `chatId?` |
+| | **Body:** `{ text, channel?, chatId? }` | |
+
+---
+
+# Appendix D: Complete MCP Tool Index
+
+Quick reference — all MCP tools alphabetically:
+
+| # | Tool | Category |
+|---|------|----------|
+| 1 | `add_account` | Accounts |
+| 2 | `add_agent_route` | Channels |
+| 3 | `add_monitored_chat` | Channels |
+| 4 | `add_mcp_to_registry` | Marketplace |
+| 5 | `assign_to_agent` | Marketplace |
+| 6 | `browse_dirs` | Utilities |
+| 7 | `browse_registry` | Marketplace |
+| 8 | `check_account_status` | Accounts |
+| 9 | `check_app_health` | Apps |
+| 10 | `clear_memory_context` | Memory |
+| 11 | `clear_model` | Model |
+| 12 | `create_agent` | Agents |
+| -- | `create_gym_card` | AI Gym |
+| -- | `create_gym_guide` | AI Gym |
+| 13 | `create_app` | Apps |
+| 14 | `create_cron` | Cron |
+| 15 | `create_goal` | Goals |
+| 16 | `create_mcp_connection` | MCPs |
+| 17 | `create_initiative` | Projects |
+| 18 | `create_project` | Tasks |
+| 19 | `create_prompt` | Marketplace |
+| 19 | `create_skill` | Skills |
+| 20 | `create_task` | Tasks |
+| 21 | `delegate_message` | Chat |
+| 22 | `delete_account` | Accounts |
+| 23 | `delete_agent` | Agents |
+| -- | `delete_gym_program` | AI Gym |
+| 24 | `delete_app` | Apps |
+| 25 | `delete_cron` | Cron |
+| 26 | `delete_goal` | Goals |
+| 27 | `delete_mcp_connection` | MCPs |
+| -- | `delete_project` | Projects |
+| 28 | `delete_mcp_key` | MCPs |
+| 29 | `delete_session` | Sessions |
+| -- | `create_session_tab` | Session Tabs |
+| -- | `delete_session_tab` | Session Tabs |
+| 30 | `delete_task` | Tasks |
+| -- | `execute_project` | Projects |
+| -- | `dismiss_gym_card` | AI Gym |
+| 31 | `download_agent_file` | Files |
+| 32 | `get_activity` | Activity |
+| 33 | `get_agent` | Agents |
+| -- | `get_agent_activity_summary` | AI Gym |
+| 34 | `get_agent_cost` | Cost |
+| 35 | `get_agent_instructions` | Agents |
+| 36 | `get_agent_logs` | Activity |
+| 37 | `get_agent_memory` | Memory |
+| -- | `get_session_tab_history` | Session Tabs |
+| 38 | `get_agent_registry` | Agents |
+| 39 | `get_agent_skills` | Skills |
+| 40 | `get_all_costs` | Cost |
+| 41 | `get_all_tasks` | Tasks |
+| 42 | `get_changelog` | Utilities |
+| 43 | `get_chat_job_raw` | Chat |
+| 44 | `get_cron_history` | Cron |
+| 45 | `get_dashboard` | Dashboard |
+| -- | `get_dimension_history` | AI Gym |
+| 46 | `get_goal_history` | Goals |
+| -- | `get_heartbeat_history` | Heartbeat |
+| -- | `get_gym_config` | AI Gym |
+| -- | `get_gym_feed` | AI Gym |
+| -- | `get_gym_program` | AI Gym |
+| -- | `get_gym_progress` | AI Gym |
+| -- | `get_learner_profile` | AI Gym |
+| -- | `get_mcp_catalog` | MCPs |
+| 49 | `get_model` | Model |
+| -- | `get_org_skills` | Skills |
+| -- | `get_plan` | AI Gym |
+| -- | `get_platform_agents` | Lab |
+| -- | `get_project` | Projects |
+| -- | `get_project_status` | Projects |
+| -- | `get_profile` | Profile |
+| -- | `get_prompt_trigger` | Marketplace |
+| 53 | `get_saas_config` | SaaS |
+| 54 | `get_service_config` | Config |
+| 55 | `get_sticky_routing` | Channels |
+| 56 | `get_task_stats` | Tasks |
+| 57 | `get_wiki_sync_history` | Wiki |
+| 58 | `health_check` | Dashboard |
+| -- | `import_program` | AI Gym |
+| 59 | `import_skills` | Marketplace |
+| 60 | `install_registry_item` | Marketplace |
+| 61 | `install_xbar` | Utilities |
+| -- | `link_to_project` | Projects |
+| 62 | `list_agents` | Agents |
+| -- | `list_gym_cards` | AI Gym |
+| -- | `list_gym_guides` | AI Gym |
+| -- | `list_gym_programs` | AI Gym |
+| 63 | `list_agent_files` | Files |
+| 64 | `list_apps` | Apps |
+| 65 | `list_automations` | Automations |
+| 66 | `list_channels` | Channels |
+| 67 | `list_mcp_connections` | MCPs |
+| 68 | `list_mcp_keys` | MCPs |
+| 69 | `list_mcps` | MCPs |
+| 70 | `list_paired_senders` | Pairing |
+| -- | `list_projects` | Projects |
+| 71 | `list_sessions` | Sessions |
+| -- | `list_session_tabs` | Session Tabs |
+| 72 | `list_tasks` | Tasks |
+| 73 | `list_accounts` | Accounts |
+| 74 | `pair_sender` | Pairing |
+| -- | `pause_project` | Projects |
+| 75 | `publish_to_saas` | SaaS |
+| 76 | `recover_agent` | Agents |
+| -- | `run_gym_digest` | AI Gym |
+| -- | `rename_session_tab` | Session Tabs |
+| 77 | `remove_agent_route` | Channels |
+| 78 | `remove_monitored_chat` | Channels |
+| 79 | `reset_session` | Sessions |
+| 80 | `save_mcp_key` | MCPs |
+| 81 | `scan_skills` | Marketplace |
+| -- | `search_agent_logs` | AI Gym |
+| 82 | `search_memory` | Memory |
+| 83 | `send_message` | Chat |
+| -- | `snapshot_dimensions` | AI Gym |
+| 84 | `send_webhook` | Webhook |
+| 85 | `set_model` | Model |
+| 86 | `set_platform_default` | Marketplace |
+| 87 | `set_prompt_trigger` | Marketplace |
+| 88 | `start_account_login` | Accounts |
+| 89 | `start_stream` | Chat |
+| 90 | `stop_chat_job` | Chat |
+| 91 | `submit_login_code` | Accounts |
+| 92 | `test_provider` | Service Config / Multi-Provider |
+| 93 | `test_saas_connection` | SaaS |
+| 94 | `toggle_cron` | Cron |
+| 94 | `toggle_goal` | Goals |
+| 95 | `trigger_cron` | Cron |
+| 96 | `trigger_goal` | Goals |
+| 97 | `trigger_heartbeat` | Heartbeat |
+| 98 | `trigger_wiki_sync` | Wiki |
+| 99 | `unpair_sender` | Pairing |
+| -- | `unlink_from_project` | Projects |
+| 100 | `update_agent` | Agents |
+| -- | `update_gym_program` | AI Gym |
+| -- | `update_gym_progress` | AI Gym |
+| -- | `update_learner_profile` | AI Gym |
+| 101 | `update_app` | Apps |
+| 102 | `update_channel` | Channels |
+| -- | `update_plan` | AI Gym |
+| -- | `update_profile` | Profile |
+| -- | `update_project` | Projects |
+| 103 | `update_saas_config` | SaaS |
+| 104 | `update_service_config` | Config |
+| 105 | `update_task` | Tasks |
+| 106 | `upload_file` | Files |
+| 107 | `whoami` | Accounts |