Skip to main content

Dashboard Guide

The AgentHiFive web dashboard is where you manage agents, connections, policies, and approvals. This page describes every section of the UI and what you can do in each.

The top navigation bar contains:

  • AgentHiFive logo link — returns you to the My Agents home page.
  • Requests — opens the Approvals page. A red badge appears when there are pending approval requests.
  • Activity — opens the activity timeline.
  • Advanced dropdown — expands to reveal Agents, Connections, and Policies pages for lower-level management.
  • Documentation icon — opens the docs site in a new tab.
  • User avatar — opens a dropdown with your name, email, and links to Settings, Notifications, Apps, Documentation, and Sign out.

My Agents (Home)​

Path: /dashboard/my-agents

This is your landing page after login. It shows all agents in your workspace with their connected accounts.

What you can do:

  • View agent cards — each agent displays its name, description, status badge, and creation date.
  • See connections per agent — connections are grouped under the agent that has policies for them, showing the service icon, label, provider, status (Healthy / Needs Reconnect / Revoked), and granted scopes.
  • Open connection details — click a connection to view its full details, including credential preview, metadata, and all policies attached to that agent.
  • Add a policy — from the connection detail modal, create a new policy that governs how the agent can use that connection (allowed models, default mode, step-up approval rules).
  • Create a new connection for an agent — open the create-connection modal to link an additional OAuth account to a specific agent.
  • Generate a bootstrap secret — issue a one-time enrolment key so a new agent instance can register itself. Requires confirmation before generating.
  • See available connections — connections in your workspace that do not yet have policies for an agent are listed separately so you can attach them.

Setup Wizard​

Path: /dashboard/setup

The setup wizard runs automatically when you first sign in and have no agents. It creates a default "OpenClaw" agent and walks you through two steps:

  1. Step 1 — AI Provider: Connect an LLM provider (e.g., OpenAI, Anthropic) so your agent has an AI backend. You provide credentials and the wizard creates the connection.
  2. Step 2 — Connect Accounts: Link the accounts your agent will operate on, such as Google (Gmail, Calendar) or Microsoft (Outlook, Teams).

After completing both steps, the wizard displays a one-time enrolment key (bootstrap secret) that you copy and provide to your OpenClaw instance during its own setup. The key expires in 1 hour.

You can skip the wizard at any time using the "Skip setup" button in the top-right corner.

Approvals​

Path: /dashboard/approvals

This page lists all pending permission requests and step-up approval requests from your agents.

What you can do:

  • Review pending requests — each request shows the agent name, connection, action being requested, and when it was submitted.
  • Inspect request details — expand a request to see rich context depending on the action type:
    • Email actions: recipient list (to/cc), sender, subject, and body preview.
    • Calendar events: event details and attendees.
    • Telegram messages: chat ID and message text.
    • Teams messages: channel or chat context and body preview.
    • Slack messages: channel and message text.
    • Attachment access: message subject, sender, attachment name, and file size.
  • View guard triggers — if a safety guard flagged the request (e.g., prompt injection detection or PII bypass attempt), you can see the rule label, pattern type, matched field, and an excerpt of the flagged content.
  • Approve or deny — take action on each request. Approvals grant the agent permission to proceed; denials block the action.
  • Handle OAuth consent — some requests may require you to complete an OAuth popup flow to grant additional scopes before approving.

Activity​

Path: /dashboard/activity

A chronological timeline of every action taken by agents in your workspace.

What you can do:

  • Browse the event timeline — events are listed newest-first with timestamps, showing what happened in plain language.
  • Filter by agent — narrow the view to a specific agent.
  • Filter by connection — show only events for a particular connected account.
  • Filter by date range — select start and end dates to scope the timeline.
  • Read event descriptions — each event is rendered as a human-readable summary (e.g., "OpenClaw obtained access token via google", "OpenClaw executed POST /gmail/v1/users/me/messages/send (200)").
  • See decision badges — every event displays a colored badge indicating the outcome:
    • Success (green) — the action was allowed and completed.
    • Denied (red) — the policy engine or an approval blocked the action.
    • Error (orange) — the action failed due to an unexpected error.
  • Identify provider — a provider icon (G for Google, M for Microsoft, T for Telegram) appears next to each event.

Event types include: token_vended, token_vend_denied, execution_requested, execution_completed, execution_denied, execution_error, rate_limit_exceeded, connection_revoked, connection_needs_reauth, and policy_created.

Connections (Advanced)​

Path: /dashboard/connections

A workspace-wide view of all OAuth connections, independent of which agent uses them.

What you can do:

  • Browse available services — connections are organized by service category (e.g., Email, Calendar, Messaging, AI Providers) using the service catalog.
  • Create a new connection — select a service, then complete the OAuth popup flow to authorize AgentHiFive to access that account. Singleton services allow only one connection per service.
  • View connection status — each connection shows one of three statuses:
    • Healthy (green) — credentials are valid and working.
    • Needs Reconnect (yellow) — the token has expired or been invalidated; click to re-authorize.
    • Revoked (red) — the connection was explicitly revoked.
  • See connection metadata — details vary by provider. For example, Telegram shows bot username; Microsoft shows email, display name, and tenant ID.
  • View attached policies — see which agents have policies that use this connection and what permissions they grant.
  • Revoke a connection — permanently disable a connection. Some providers support instant revocation; others are marked as revoked locally.
  • Re-authorize — for connections in "Needs Reconnect" state, launch the OAuth flow again to refresh credentials.

Agents (Advanced)​

Path: /dashboard/agents

Direct management of agent registrations in your workspace.

What you can do:

  • List all agents — see every registered agent with its name, description, icon, status, and creation date.
  • Create a new agent — fill in a name, optional description, and optional icon URL. After creation, a bootstrap secret is displayed for the agent to use during enrolment.
  • View agent status lifecycle:
    • Created — the agent record exists but has not yet enrolled using its bootstrap secret.
    • Active — the agent has enrolled and is operational.
    • Disabled — the agent has been manually disabled and cannot make requests.
  • Generate a new bootstrap secret — if the original secret expired or was lost, generate a replacement. This invalidates any previous unused secret.
  • Disable or enable an agent — toggle an agent between active and disabled states.
  • Delete an agent — permanently remove an agent and all its associated policies. Requires confirmation.

Settings​

Workspace​

Path: /dashboard/settings

General workspace configuration.

What you can do:

  • View workspace ID — your unique workspace identifier, useful for API integrations.
  • Rename your workspace — update the display name.
  • View backend version — see the current build number and date (or "Development" for local builds).
  • Manage API tokens — programmatic access tokens for the AgentHiFive API:
    • Create a token — provide a name and select an expiry period (e.g., 30 days, 90 days, 1 year). The token value is shown once after creation; copy it immediately.
    • View existing tokens — see token name, creation date, expiry date, and last-used date.
    • Revoke a token — permanently invalidate a token so it can no longer authenticate API requests.

Notifications​

Path: /dashboard/settings/notifications

Configure where AgentHiFive sends real-time notifications (e.g., when an approval is needed).

What you can do:

  • Set up Telegram notifications — select a Telegram bot connection, then choose or enter a chat ID. The system can auto-detect recent chats from your bot.
  • Set up Slack notifications — select a Slack bot connection and specify a channel ID for delivery.
  • Test a channel — send a test notification to verify the configuration works.
  • Enable or disable channels — toggle individual notification channels on or off without deleting them.
  • View verification status — see whether each channel has been successfully verified.

Apps​

Path: /dashboard/settings/apps

Register your own OAuth applications instead of using the shared AgentHiFive defaults. This is useful for organizations that want full control over their OAuth credentials.

What you can do:

  • Register a Google OAuth app — provide your Client ID and Client Secret from the Google Cloud Console. A label helps you identify the app.
  • Register a Microsoft OAuth app — provide your Client ID, Client Secret, and Tenant ID from the Azure portal.
  • View the callback URL — the redirect URI you need to configure in your OAuth provider's console.
  • See registered apps — list of all custom OAuth apps with their provider, client ID, label, and registration date.
  • Delete a custom app — remove a custom OAuth app registration to revert to the shared defaults.

Notification Bell​

The Requests link in the top navigation doubles as a notification indicator. When there are pending approval requests (permission requests or step-up approvals), a red badge appears on the link showing the total count. This count refreshes automatically every 30 seconds and updates immediately when you approve or deny a request.

Clicking "Requests" takes you to the Approvals page where you can review and act on all pending items.