# Manage every WordPress site at once with Paperclip

> Managing every WordPress site at once with Paperclip means assigning one of its agents a ticket or routine that calls xSpeed Hub across the whole workspace, then reading the per-site result on the task.

Page: https://xspeedcache.com/agent/paperclip/fleet/
Last updated: October 2026

Paperclip suits an agency that already thinks in tickets. You run a team of agents, one of them a Claude Code agent with the Hub added on its host, and fleet maintenance becomes work you assign. A ticket says to clear and warm every client site after Thursday's plugin update. The agent calls purge_cache with site set to "all", then start_preloader, and posts the per-site result on the task: which sites cleared, which did not, and the error for each failure. You read it in the Paperclip dashboard beside everything else the team did that day.

Routines give you the recurring version. A routine fires on a cron schedule, creates a task and wakes the agent you assign, and timer heartbeats stay off until you turn them on. The Hub has no scheduler for agent prompts, so the cron is Paperclip's. A weekly routine that runs run_benchmark across every site is a sound choice, because that tool is a read and changes nothing. The other six fan-out tools (purge_cache, toggle_cache, update_settings, start_preloader, stop_preloader and optimize_site) write, and a routine runs with nobody watching.

Which gate applies depends on how the agent reaches the Hub. Added with claude mcp add on the host, the server does not pass through Paperclip's action permissions or review queue, and the Claude Code adapter runs headless with dangerouslySkipPermissions on by default, so a fleet purge runs as soon as the agent decides to. Set it to false and a tool that needs approval does not run. Attached on the Connectors page instead, each Hub tool is set to Allowed, Ask first or Off. xSpeed Hub has no confirmation step of its own; its instruction to confirm once before a fleet-wide purge, toggle, settings or preloader call reaches the agent as text, and it does not name optimize_site.

## Set up Paperclip once

### 1. Put your sites in xSpeed Hub

Sign in at app.xspeedcache.com with Google or email; the Hub is free and has no site cap. Then connect each WordPress site from its own dashboard: click Connect Hub in the xSpeed Cache top bar, then Connect via xSpeed Hub. Each site needs the free xSpeed Cache plugin.

### 2. Add xSpeed Hub to the agent Paperclip runs

For a Claude Code agent, run this on the machine that hosts it, as the same OS user the Paperclip heartbeat runs as. Paperclip's docs say MCP wiring lives at the adapter and runtime layer, and user scope makes the Hub available to every Claude Code agent that user runs.

```bash
claude mcp add --transport http --scope user xspeedhub https://app.xspeedcache.com/xspeed/mcp
```

### 3. Sign in once, or use a token

Start claude interactively as that same user, run /mcp, choose xspeedhub and approve access on the xSpeed Hub page. Claude Code keeps the credentials, so later headless runs need no browser; claude mcp get xspeedhub shows the status. On a host with no browser, add the server with the connection token from Connect AI in the Hub instead, ideally with Read-only everywhere on.

```bash
# Interactive sign-in
claude
/mcp

# Or a token, no browser
claude mcp add --transport http --scope user xspeedhub https://app.xspeedcache.com/xspeed/mcp \
  --header "Authorization: Bearer <your connection token>"
```

Full setup: https://xspeedcache.com/agent/paperclip/

### Before you send a prompt that changes something

Reads change nothing on your sites, though contact_support emails xSpeed support. Writes run once your connection allows them, so your client's approval prompt and a read-only connection (the token's Read-only everywhere switch, or a Viewer sign-in) are the gates that matter. Paperclip's Claude Code adapter runs headless with dangerouslySkipPermissions on by default, which lets the agent call MCP tools without asking. Set it to false and a tool that needs approval does not run. A server added on the adapter does not pass through Paperclip's action permissions or review queue. A Paperclip connector does: each Hub tool is set to Allowed, Ask first or Off, and because the Hub publishes no annotations Paperclip classifies them by name, which can under-classify, so start every write Off. xSpeed Hub has no confirmation step of its own, so a write that gets past those settings runs as soon as your connection allows writes. For unattended routines, send a connection token with Read-only everywhere on, which makes the Hub refuse every write tool.

## Prompts

- Weekly routine: use xSpeed Hub to benchmark every site and post one line per site on this task. Report only.
- Ticket: with xSpeed Hub, purge every client site after the plugin update and list the failures here.
- Use xSpeed Hub to list the workspaces this agent can reach and the site count in each, without calling any write tool.

## Example session

```
You: Ticket: the plugin update is out on every client site. Use xSpeed Hub to clear and re-warm them all. Wait for my reply before any change.
  → list_sites [read-only]: 14 sites in the workspace
Agent: This will touch all 14 sites: a full purge, then the preloader on each. Reply to approve.
You: Approved.
  → purge_cache(site: all, type: all) [changes site]: 13 of 14 purged; client-q returned an error
  → start_preloader(site: all) [changes site]: 13 started; client-q unreachable
Agent: Thirteen sites are clear and warming. client-q failed both calls, so I left the ticket open for it and noted the error.
```

## How it works

1. **See what you can reach**: list_sites returns each site's handle, URL, status and scopes. When you have more than one workspace, list_workspaces returns each one's handle, id, your role, its site count, which one is the default, and whether this connection can reach it.
2. **Choose the scope of the call**: The agent passes site: "all" for every site in one workspace, or an array of handles for exactly those sites. For a workspace other than the default it also passes the workspace argument. Omit workspace and the call acts on the default workspace.
3. **Confirm once**: Before a purge, toggle, settings change or preloader call that touches every site, the Hub tells the agent to say "this will touch all N sites" and wait for you, then make one call. That instruction does not name optimize_site, so for it your client's prompt is the gate. It is guidance to the agent, not a lock. It is one tool call, so one approval in your client covers every site in it.
4. **Run it as one call**: The Hub runs the sites one after another with a one-second pause between live calls, so a fleet purge is not a burst of traffic at your servers. Revoked sites are skipped from "all". Each site's action lands in the activity log as its own row.
5. **Read the result per site**: The reply is a summary of targets, succeeded and failed, plus one row per site with ok and either its data or its error. Some sites failing is a partial success. Only a run where every site failed comes back as an error.
6. **Follow up on the failures**: The agent reports which sites failed and why, then retries just those handles with an array. Tools that read one site, such as get_cache_status and get_health, do not fan out, so a fleet-wide check means one call per site.

## Reference

| | |
| --- | --- |
| Fan-out tools | purge_cache, toggle_cache, update_settings, start_preloader, stop_preloader, optimize_site, run_benchmark |
| Selector | site: "all", or an array of handles |
| What all covers | Every site this connection can see in one workspace, minus revoked ones |
| Pacing | One after another, with a 1 second pause between live calls |
| Result | summary (targets, succeeded, failed) and a row per site: ok, then data or error |
| Partial failure | Reported per site; the call counts as an error only when every site failed |
| Workspace argument | Optional. Omit it for the default workspace; name one by its handle or id from list_workspaces |
| Reach | Fixed when you approve the connection; list_workspaces shows reachable: false for the rest |
| Not fan-out | get_health, get_cache_status and every other tool take one site per call |
| Slow tools | optimize_site takes up to 2 minutes per site; a target_score run is capped at 12 rounds per call |
| Read-only | The connection token with Read-only everywhere on, or a Viewer sign-in, refuses every fan-out write; run_benchmark works. An OAuth sign-in gets the scopes the client asks for |
| Hub daily pass | Re-verifies sites, snapshots cache status and hit ratio, reads stored reports, alerts by Slack webhook or email |

## Rules

- Confirm once, then call once. The Hub has no confirmation step of its own, so your client's approval prompt and a read-only connection (the connection token with Read-only everywhere on, or a Viewer sign-in) are the gates that matter. The Hub's "this will touch all N sites" is an instruction to the agent for purge, toggle, settings and preloader calls, and it does not name optimize_site.
- "all" is one workspace, never every workspace. To act in another workspace, name it. A write to a non-default workspace must name it, and an array of handles is looked up only in the workspace the call names.
- Read every result per site. Treat some sites failing as a partial result, and never accept a blanket "done" from the agent.
- Try a behaviour change on one site before you send it to the fleet. toggle_cache and update_settings on every site are one call, and each site checks values against its own settings, so a Pro setting can fail on a Free site.
- optimize_site on many sites is slow, because each site can take up to two minutes inside one call. Use dry_run first, or name a short list of handles.
- The Hub has no scheduler for agent prompts. Its own daily pass and alerts run without an agent, and an agent can only act when you send a prompt or your AI client runs one on a timer.

## Good to know with Paperclip

The Hub publishes no tool annotations, so on the Connectors page Paperclip classifies its tools by name, and Paperclip says that fallback can under-classify. Start every Hub write Off, promote purge_cache or start_preloader to Ask first when a ticket needs them, and use Refresh actions with care, since it can widen what agents may call. For the adapter route, give routines a connection token with Read-only everywhere on: the Hub then refuses every fan-out write from that agent, and run_benchmark still works. User scope in Claude Code is shared by every Claude Code agent that OS user runs, so each of them reaches every site in the workspace. If only one agent should handle fleet work, run it as a dedicated OS user. Name the workspace on any ticket for a client workspace, because "all" covers one.

## More prompts for this job

They work in any client connected to xSpeed Hub.

- Use xSpeed Hub to list every site I can reach, and every workspace, and say which workspace you act in by default.
- With xSpeed Hub, purge the cache on every site in my workspace and tell me which ones failed.
- Use xSpeed Hub to run a benchmark on all my sites and rank them by cached response time.
- Use xSpeed Hub to start the preloader on blog, shop and docs only.
- With xSpeed Hub, turn page caching on for every site where it is off. Check each site first.
- Use xSpeed Hub to stop the preloader everywhere. The servers are busy.
- Ask xSpeed Hub to preview a safe optimize_site pass on every site with dry_run and summarize what it would change.
- Use xSpeed Hub to purge every site in the client-acme workspace, not my default one.

## Frequently asked questions

### Can a Paperclip routine run a fleet check every week?

Yes. A routine fires on a cron schedule and gives the task to an agent. Have it run run_benchmark across every site, which is a read, and post the per-site results on the task for a person to act on.

### Which route keeps fleet purges behind a person in Paperclip?

The Connectors route, with purge_cache and the other fan-out writes set to Ask first or Off. A server added with claude mcp add bypasses Paperclip's action permissions, and the Claude Code adapter skips prompts by default, so that route relies on a read-only connection.

### What does site "all" cover?

Every site the connection can see in one workspace. It never crosses workspaces, and revoked sites are skipped. If your connection is limited to some sites, "all" quietly means the sites you are allowed, with no error naming the rest. To act in another workspace, the agent names it with the workspace argument.

### What happens if one site fails?

The other sites still run. The reply has a summary and a row for each site with ok and its data or its error, so a site that was unreachable is named. A run counts as an error only when every site failed. Ask the agent to retry just the failed handles.

### Can the agent check the health of every site in one call?

No. get_health and get_cache_status take one site per call, so the agent loops. The tools that fan out are purge_cache, toggle_cache, update_settings, start_preloader, stop_preloader, optimize_site and run_benchmark. The Hub dashboard has a fleet overview for health.

### Does the Hub watch my fleet when no agent is running?

Yes, on its own. A daily pass re-verifies each site, retrying a failed probe once, snapshots cache status and hit ratio, and reads the PageSpeed and GTmetrix reports your sites already stored. It can alert you by Slack webhook or email when a site becomes unreachable, caching is switched off, the hit ratio falls below a floor you set, or a score regresses. Each alert fires once per incident, not every day.

### Does the Hub ask before it touches every site?

The Hub has no confirmation step. Its instructions tell the agent to confirm once with you, saying how many sites a call will touch, but that is guidance and not a lock. Your AI client's approval prompt and a read-only connection (the connection token with Read-only everywhere on, or a Viewer sign-in) are what stop a write. The instruction covers purge, toggle, settings and preloader calls, not optimize_site. The dashboard's own Purge all button does confirm before it runs.

### Can the agent add sites, remove sites or change who has access?

No. It cannot add, remove or rename sites, change members, roles or billing, mint or rotate tokens, or widen the workspaces this connection reaches. Those stay in the Hub dashboard.

## Manage every site at once with other agents

[Claude Code](https://xspeedcache.com/agent/claude-code/fleet/) · [Claude](https://xspeedcache.com/agent/claude/fleet/) · [Claude Cowork](https://xspeedcache.com/agent/claude-cowork/fleet/) · [ChatGPT](https://xspeedcache.com/agent/chatgpt/fleet/) · [Codex](https://xspeedcache.com/agent/codex/fleet/) · [Cursor](https://xspeedcache.com/agent/cursor/fleet/) · [GitHub Copilot in VS Code](https://xspeedcache.com/agent/github-copilot/fleet/) · [Windsurf](https://xspeedcache.com/agent/windsurf/fleet/) · [Gemini CLI](https://xspeedcache.com/agent/gemini-cli/fleet/) · [Antigravity](https://xspeedcache.com/agent/antigravity/fleet/) · [Zed](https://xspeedcache.com/agent/zed/fleet/) · [Kiro](https://xspeedcache.com/agent/kiro/fleet/) · [OpenCode](https://xspeedcache.com/agent/opencode/fleet/) · [OpenClaw](https://xspeedcache.com/agent/openclaw/fleet/) · [Hermes Agent](https://xspeedcache.com/agent/hermes-agent/fleet/) · [Grok Build](https://xspeedcache.com/agent/grok-build/fleet/) · [ChatGPT dots](https://xspeedcache.com/agent/chatgpt-dots/fleet/) · [Grok](https://xspeedcache.com/agent/grok/fleet/) · [Grok Bot](https://xspeedcache.com/agent/grok-bot/fleet/) · [Muse](https://xspeedcache.com/agent/muse/fleet/) · [Manus](https://xspeedcache.com/agent/manus/fleet/) · [Kimi Code](https://xspeedcache.com/agent/kimi/fleet/) · [NanoClaw](https://xspeedcache.com/agent/nanoclaw/fleet/)

## More with Paperclip

- [Purge the WordPress cache with Paperclip](https://xspeedcache.com/agent/paperclip/purge-cache/)
- [Find out why WordPress pages are not cached with Paperclip](https://xspeedcache.com/agent/paperclip/troubleshoot-cache/)
- [Scan a website for speed problems with Paperclip](https://xspeedcache.com/agent/paperclip/speed-scan/)
- [Run and track PageSpeed tests with Paperclip](https://xspeedcache.com/agent/paperclip/pagespeed-tests/)
- [Raise a WordPress site's PageSpeed score with Paperclip](https://xspeedcache.com/agent/paperclip/optimize-site/)
- [Tune WordPress cache settings with Paperclip](https://xspeedcache.com/agent/paperclip/cache-settings/)
- [Warm the WordPress cache with Paperclip](https://xspeedcache.com/agent/paperclip/preload-cache/)
- [Set up the Redis object cache with Paperclip](https://xspeedcache.com/agent/paperclip/object-cache/)
- [Manage Cloudflare caching with Paperclip](https://xspeedcache.com/agent/paperclip/cloudflare/)

## Documentation

- Managing your fleet with xSpeed Hub: https://xspeedcache.com/docs/managing-your-fleet-with-xspeed-hub/
- Teams, roles and workspaces in xSpeed Hub: https://xspeedcache.com/docs/hub-teams-and-workspaces/
- Scoped AI connections in xSpeed Hub: https://xspeedcache.com/docs/hub-connection-scopes/
- Connecting to xSpeed Hub: https://xspeedcache.com/docs/connecting-to-xspeed-hub/