How NanoClaw handles this
NanoClaw is handy for the Cloudflare question that arrives in a team chat. Someone in Slack says the pricing page looks old; you reply to the agent with its trigger word, @Andy by default, and name the site. The agent group that has xSpeed Hub calls cloudflare_verify, which asks Cloudflare whether the token and Zone ID stored in xSpeed still resolve, and answers in the same thread, so everyone sees the result.
If the link works and the page is stale, the agent can purge in the right order: purge_cache for the server page cache, then purge_cloudflare for the edge, so the edge does not refill from the old copy. set_cloudflare_dev_mode with enabled true bypasses the edge for about three hours, which helps a developer testing changes but slows the site for every visitor, so it is rarely the right answer to a message in a busy channel. None of the three Cloudflare tools accepts site: "all".
A scheduled task covers the dull half. ncl tasks create with a cron recurrence can run cloudflare_verify on each site every morning and post only the failures to a chat you name in the prompt, because a scheduled task has no chat attached. That check only reads, so it suits a run nobody watches.
Set up NanoClaw once
Already connected? Skip to the prompts. Alternatives and troubleshooting are on the NanoClaw guide.
1Put 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.
2Add xSpeed Hub to your agent group
Copy the connection token from Connect AI in the Hub (only the workspace owner sees it). Then run this on the machine that hosts NanoClaw, with your agent group's ID from ncl groups list. The header goes in as a JSON object, because NanoClaw refuses credentials inside the URL. NanoClaw may ask an admin to approve the change.
ncl groups config add-mcp-server --id <group-id> --name xspeedhub \
--url https://app.xspeedcache.com/xspeed/mcp \
--headers '{"Authorization":"Bearer <your connection token>"}' 3Restart the group and ask
The new server takes effect after a restart. Then message your agent with its trigger word, @Andy by default, and ask something that names xSpeed Hub.
ncl groups restart --id <group-id>
Full NanoClaw setup, sign-in options and FAQ
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. Whether NanoClaw asks you before an MCP tool runs depends on how your fork and its admins are set up; its documentation gates adding a server behind approval but describes no per-tool prompt for calls. xSpeed Hub has no confirmation step of its own, so a write runs as soon as the agent calls it and your connection allows writes. The Hub tells the agent to confirm the target site and to ask once before a write that touches every site, which in NanoClaw shows up as a question in the chat. The firm stop is the token: with Read-only everywhere on, the Hub refuses every write tool. Keep it on for scheduled tasks.
What do I ask?
Three prompts written for NanoClaw. More for this job are below.
@Andy use xSpeed Hub to verify the Cloudflare link on shop and tell me what Cloudflare said.
@Andy with xSpeed Hub, purge the page cache on blog and then its Cloudflare edge.
@Andy using xSpeed Hub, check Cloudflare on every site one at a time and list any that fail.
What happens, step by step
When Cloudflare sits in front of a WordPress site, it keeps its own copy of your files at the edge. Through xSpeed Hub an agent can check that the site's Cloudflare credentials still work, purge the edge cache, and turn development mode on or off. All three need Cloudflare connected in xSpeed on that site, and each works on one site at a time.
01
Verify the connection
cloudflare_verify calls Cloudflare's API with the credentials stored on the site and checks that the token and zone still resolve. It purges nothing and works on a read-only connection. If Cloudflare is not connected on that site, the call comes back as an error and the agent should tell you rather than retry.
02
Purge the edge
purge_cloudflare empties the Cloudflare edge cache for the site's zone. It does not touch the copies on your server. When Cloudflare is in front of a site, a content change usually needs both, so the agent runs purge_cache first and purge_cloudflare after it, or one purge_cache of type all, which reaches the edge when Cloudflare is connected.
03
Turn development mode on to see origin changes
set_cloudflare_dev_mode with enabled true bypasses the edge for about three hours, so changes on the origin show immediately. It slows the site for real visitors the whole time. Your client's prompt is the gate here, because the Hub's instructions do not name this tool in their confirm list.
04
Turn it off when you finish
set_cloudflare_dev_mode with enabled false switches development mode off. Cloudflare also ends it on its own after about three hours. No Hub tool reads whether it is currently on, so note when you turned it on.
05
Read a failure
A rejected token, an empty zone ID or a refused call comes back as an error that names the action, the HTTP status and Cloudflare's message when it gave one. It is not reported as a success. The credentials themselves are set in the xSpeed Cache panel in wp-admin, not through the Hub.
Reference
| Read tool | cloudflare_verify; purges nothing and works on a read-only connection |
|---|---|
| Write tools | purge_cloudflare and set_cloudflare_dev_mode (enabled true or false) |
| Needs | Cloudflare connected in xSpeed on that site: integration on, an API token (or the legacy key and email) and a Zone ID |
| Edge purge scope | Everything Cloudflare holds for the site's zone; the server copies stay |
| Development mode | Bypasses the edge for about 3 hours, then Cloudflare switches it off |
| Token permissions | Zone Cache Purge, and Zone Settings Edit for development mode |
| What verify proves | The token and zone resolve; a token without Zone Settings Edit still verifies |
| Through purge_cache | Type all also clears the edge when Cloudflare is connected; page, assets, object and rest do not |
| Every site at once | None of the three accept site: "all"; one site per call |
| Failure shape | An error with the action, the HTTP status and Cloudflare's message |
| Read-only connection | The connection token with Read-only everywhere on, or a Viewer member's sign-in; purge_cloudflare and set_cloudflare_dev_mode are refused and cloudflare_verify works. An OAuth sign-in gets the scopes the client asks for, so the switch does not make it read-only |
| Confirmation | None on the Hub; your client's prompt is the gate |
Rules worth keeping
- Run cloudflare_verify first, but do not read a pass as proof that every call will work. A token that lacks Zone Settings Edit verifies and purges fine, then fails the first time development mode is switched.
- Purge the server and the edge together after a content change. Purging only the site leaves visitors on the old page until the edge copy expires.
- Turn development mode off as soon as you are done. It slows the site for real visitors for up to about three hours, and no tool reports its state.
- Never paste a Cloudflare token or Global API Key into the chat. Connect Cloudflare in the xSpeed Cache panel in wp-admin, and use a scoped API token rather than the Global API Key.
- purge_cloudflare and set_cloudflare_dev_mode are write tools. The Hub runs them as soon as your connection allows writes, so your client's approval prompt and a read-only connection (the connection token with Read-only everywhere on, or a Viewer member's sign-in) are the gates that matter.
Good to know with NanoClaw
NanoClaw gates adding a server behind approval but describes no per-tool prompt for calls, so whether the agent stops before purge_cloudflare or set_cloudflare_dev_mode depends on your fork. The Hub has no confirmation step of its own, and its instructions to the agent do not list set_cloudflare_dev_mode among the tools to confirm. If the agent answers a shared channel, a request from anyone it responds to there can reach development mode and slow the site for about three hours. For an agent group in a busy channel, send a connection token with Read-only everywhere on, which leaves cloudflare_verify working and refuses both writes. Never post a Cloudflare token in the chat; it is set in the xSpeed Cache panel in wp-admin. The Hub token itself sits in the group's container configuration on your host, so keep that machine private.
More prompts for this job
They work in any client connected to xSpeed Hub.
Use xSpeed Hub to check that the Cloudflare connection on shop still works.
With xSpeed Hub, purge the Cloudflare edge cache on blog.
I changed the checkout page on shop. Use xSpeed Hub to purge the site cache and the Cloudflare edge.
Use xSpeed Hub to turn Cloudflare development mode on for shop. I am editing the theme.
With xSpeed Hub, turn Cloudflare development mode off on shop.
Ask xSpeed Hub to verify Cloudflare on shop and tell me what is wrong if it fails.
Using xSpeed Hub, which of my sites have a working Cloudflare connection? Check them one at a time.
Frequently asked questions
Keep going
Manage Cloudflare with other agents
More with NanoClaw
Documentation
- How to connect Cloudflare
- How to serve assets from a CDN
- How to enable Page Cache
- How to write prompts for xSpeed Hub
- NanoClaw + xSpeed
- Every AI agent that works with xSpeed
From the blog