Use the server
Your agents and their status
Once you connect an agent, the portal shows it where you connected it: on your own account, or in the organisation you connected it to. Each agent shows which program it is, when it last called Plotwork, and whether its next call would work.
TL;DR
| Question | Answer |
|---|---|
| Where do I see them? | Your own agents on the Keys and agents page and the dashboard. Your agents in an organisation on that organisation's Connect tab. The sidebar shows the agent of the place you are in. |
| Which agents are listed? | Every agent you connected with a browser sign-in, and every API key of yours an agent has opened an MCP session with. A key no agent has used is listed with your keys, not here. |
| How does Plotwork know which program it is? | An agent names itself when it opens an MCP session, such as claude-code 2.1.0. Plotwork keeps that name and version, and when. |
| Can Plotwork tell whether my agent is running? | No. An agent calls Plotwork; Plotwork never calls the agent. The status says what Plotwork has seen: the last call and the last session. |
| What does Check do? | Runs every rule the agent's next call would meet, on our side, and shows each one. |
| Can I press Check again and again? | Three times in two minutes. The third press goes through and the button then rests for five minutes, with the time left on it. |
| Who sees my agents in an organisation? | You, on its Connect tab. The owner and the admins see every member's connections in its Connections list. |
The statuses
| Status | What it means | What to do |
|---|---|---|
| Live | A call in the last 15 minutes | Nothing. |
| Ready | A call in the last 30 days | Nothing. It works the next time you use it. |
| Idle | No call in 30 days | Use it, or disconnect it if you no longer do. |
| Not used yet | Connected, but no call yet | Finish setting it up in the agent. Connecting an agent has each agent's steps. |
| Blocked | Its next call would be refused | The reason is shown: the key expired, the call's network is not allowed, the organisation requires two-step sign-in, you left the organisation, or the organisation's agent policy no longer allows it. |
What a check runs
| Check | It passes when |
|---|---|
| Credential | The connection or key is still live. |
| Expiry | The key is not past the date its organisation's key policy gave it. |
| Network | Its last call was not refused for coming from a network the organisation does not allow. |
| Organisation | You are still a member of the organisation it acts in. |
| Two-step sign-in | The organisation does not require it, or your account has it on. |
| Agent policy | The organisation's "Only approved agents can connect" still allows this agent. |
| Calls left | The allowance it spends has calls left today. |
When every check passes and the agent still does nothing, the problem is on the agent's side: it is not running, or its configuration does not name Plotwork. When a check fails, the reason is on ours, and the page says how to fix it.
What Plotwork keeps
The program name and version the agent sent, cut to 80 and 40 characters with control characters removed, and when it last opened a session. It goes with the connection or key. Nothing the agent asked for or drew is kept. The privacy policy lists it.
See also
| Page | Why |
|---|---|
| Connecting an agent | Connect Claude Code, Cursor, VS Code or Claude Desktop |
| Two-step sign-in | Why an organisation can block an agent until you turn it on |
| Daily limits | The daily allowance an agent's calls spend |