Command Centre

Your IntelMCP dashboard, as its own page in Claude.

The Command Centre is IntelMCP's live dashboard as a private page in your own Claude account. It opens in claude.ai or the Claude app, from Artifacts or your sidebar: see what needs attention and work through your alerts without starting a chat.

IntelMCP Command Centre · OverviewIllustrative example
The Overview view of an invented test account, Example SOC Team: 63 alerts to review, 32 high or critical, 63 alerts in 7 days and 11 active rules, each compared with the previous 7 days; a Needs attention list (a paused webhook, 63 alerts waiting for a verdict, a noisy rule, a rule that has never fired and an answered channel request); a chart of alerts per day for the last 7 days, coloured by severity; and the most severe alerts, one per line, those since the last visit marked new, starting with an Arabic post.
Illustrative example from a test account. The channels, posts, rules and indicators are invented.

What you get

The same eight views as the dashboard in chat: Overview, Alerts, Rules, Tuning, Delivery, Sources, Search and Account.

  • The post itself, not a summary. Each alert card shows the Telegram post with the matched word highlighted, the target country and the indicators found in it.
  • Proof is not a victim. Links to check-host reports, pastes and file hosts are listed apart under Proof and services, each labelled with its role (proof, paste, file…), never as victims, and are never offered as something to watch.
  • Click to read it all. Click a card to open the whole post, with where else it was reported and the context around it.
  • Triage with two buttons. Each alert card has Relevant ▾ (the alert's own severity, or medium; the arrow picks critical, high, medium or low) and Not relevant. The ⋯ menu has Open in Telegram, Acknowledge, Open the full post, Hide alerts like this and Copy link.
  • Every action saves at once. A short message says what was saved, with Undo for a verdict. After Not relevant it also offers Fewer like this?, to hide alerts like it. Tick several cards to mark or acknowledge them together.
  • A queue you can finish. The Alerts view shows To review, Relevant, Not relevant and All, with the other filters behind one Filters button. When nothing is left to review it says All caught up. Alerts matched since your last visit are marked new.
  • Rules, tuning and delivery in one place. Check which rules fire too often or never, adjust your watch profile and exclusions, and see whether your webhooks are working.
  • Sources and search. Check whether a channel is collected, request one, and search the collection without leaving the page.
AlertsIllustrative example
The Alerts view: To review, Relevant, Not relevant and All, a Filters button and a ? button for the keys. A message at the top right says Saved: relevant · high, with an Undo button. Each card shows its severity, the channel, new for alerts since the last visit, the matched word and the rule it matched, the post in its original language with the match highlighted (Arabic, English, then Russian), its indicators, and a Relevant button with a severity menu, a Not relevant button and a ⋯ menu. The second card, just marked relevant, is dimmed.
The alert list: each card shows the post with the matched word highlighted, and a verdict saves at once, with Undo.
An alert, openedIllustrative example
One alert opened: an Arabic #OpIsrael post claiming a site is down, with Relevant, Not relevant and the ⋯ menu above it. On the right, the target URL under Indicators; the check-host link listed apart under Proof and services (not victims), labelled proof; the target country, Israel; the other channels that reported it; and links to surrounding posts and to trace forwards.
An opened alert: the check-host link is listed under Proof and services, apart from the target.

Get it in three steps

Today the Command Centre is created from Claude Code, once. After that you open it in claude.ai or the Claude app.

No alerts yet? Build your watch first: answer a few questions and paste the message into Claude, so the page has alerts to show.

  1. Connect IntelMCP in claude.ai

    The page calls IntelMCP through your IntelMCP connection in claude.ai. Open Settings → Connectors → IntelMCP → Connect. Not added yet? Add it first.

  2. Install the Claude Code plugin

    In Claude Code, install the IntelMCP plugin, version 0.1.8 or later:

    /plugin marketplace add intelmcpops-creator/intelmcp-plugin
    /plugin install intelmcp@intelmcp

    Already installed? Update it from Claude Code's /plugin menu.

    The plugin also adds an IntelMCP connection to Claude Code. If you added IntelMCP there with claude mcp add, keep one or the other, not both: remove the first with claude mcp remove --scope user intelmcp.

  3. Ask Claude to create it

    Run /intelmcp:command-centre, or ask: "Create my IntelMCP command centre". Claude creates the page and replies with the link. If you already have one, Claude updates it, so the link and pin stay the same.

    This command needs no IntelMCP sign-in in Claude Code: it only publishes the page. The page itself uses IntelMCP as connected in claude.ai (step 1).

In claude.ai and Claude Desktop chat

You can also open the same dashboard inside a conversation: ask "Show my IntelMCP dashboard", or open the + menu in the message box, choose IntelMCP and pick Dashboard. It has an Expand button for full screen where the app offers it. Where the interactive panel does not appear (for example in Cowork), Claude draws a summary page instead.

The first time you open it

  • Allow IntelMCP. The first time, Claude asks whether the page may use IntelMCP. Allow it: the page then loads through your own IntelMCP connection.
  • A short note says what the page is. It appears on your first visit; press Got it and your browser remembers that you have read it. Your browser also keeps the time you last opened the page, so alerts matched since then are marked new; nothing else is kept.
  • Your data, live. This page calls IntelMCP through your own IntelMCP connection, so it shows your alerts, rules and settings, and the collection your plan includes. It changes or deletes something only when you press a button or key; deleting and rotating a secret ask you to confirm.
  • Find it again in Artifacts. It is listed in claude.ai → Artifacts as "IntelMCP Command Centre". Pin it to the sidebar to open it in one click.

Keyboard

The keys work in the Alerts view, when an alert card has focus: click a card, or move to one with Tab. Press ? there to see them.

KeyWhat it does
j / kNext, previous alert
Enter, o or SpaceOpen or close the post
1 to 4Relevant: critical, high, medium, low (1 is critical, 4 is low)
xNot relevant
aAcknowledge
zUndo the last verdict
EscClose the open post
?Show or hide the list of keys

Each key saves at once, and the next alert takes the focus. An acknowledgement cannot be undone from the page. Times are in UTC.

Questions

Why can't I share it?

It is private to you and cannot be shared: Claude does not allow pages that use a connector to be shared by link.

Why does Claude ask permission?

The page calls IntelMCP through your own connection: it reads your alerts and settings, and changes them when you press a button or key. The first time, Claude asks whether the page may use IntelMCP. Allow it, and the page loads your alerts.

Do I need to sign in to IntelMCP in Claude Code?

Not for this. /intelmcp:command-centre needs no IntelMCP sign-in in Claude Code: it only publishes the page. The page uses IntelMCP as connected in claude.ai, so IntelMCP must be connected in claude.ai. The other plugin commands do need the sign-in (/mcp in Claude Code).

How do I get the newest version?

Update the plugin from Claude Code's /plugin menu and run /intelmcp:command-centre again. Claude updates your existing page, so the link and pin stay the same.

I lost the link.

Open claude.ai → Artifacts and look for "IntelMCP Command Centre", or run /intelmcp:command-centre in Claude Code again. Pin it to the sidebar so you don't lose it.

Does it cost extra?

No. Every view is part of the Monitor plan, $25 per month. Loading a view and Refresh call IntelMCP's tools, and so does every save. Each save is one tool call, including a verdict for a page of ticked alerts; Undo is one more. Each tool call counts toward the 5,000 tool calls a day of the plan's fair use, as in chat.

Which data does it show?

This page calls IntelMCP through your own IntelMCP connection, so it shows your alerts, rules and settings, and the collection your plan includes. Posts are third-party content from public channels, and private channels and groups the operator has reviewed: leads to check, not facts. A post from a private channel shows "Private channel" in place of a link.

Open your alerts the way you open your inbox.

Subscribe, connect IntelMCP in claude.ai, then run /intelmcp:command-centre in Claude Code.