How to connect Facebook to Hermes Agent

Pairing is a two-step handshake: run hermes clawlink begin, approve the link in your browser, then run hermes clawlink finish. If you ran finish before approving, or the approval link expired, run hermes clawlink begin again to get a fresh link. Confirm the plugin was installed with --enable, then verify with hermes clawlink test.

Reconnect Facebook from the dashboard, then start a fresh chat if the runtime still has the old tool catalog loaded.

Facebook tool schemas load on demand the first time a tool runs and are cached after that, so this error usually clears on its own: wait a few seconds and retry the same request. If every Facebook call keeps failing with it in a fresh chat, reconnect from the dashboard, and contact support if it still persists — that pattern points to a configuration problem on our side, not something you can fix by reconnecting again.

Two usual causes. The connected account may not have access to the specific workspace, inbox, store, or project in the request — check that first. If access looks right, the agent may have sent a placeholder value (like "YOUR_ID" or an example id from documentation) instead of a real one: ask it to run a list or search tool first, then retry the action with a real id from those results. Most failures at this stage are one of these two, not ClawLink bugs.

Almost always a placeholder Page id. An agent that has not listed anything yet will invent something like YOUR_PAGE_ID or a plausible numeric string, and because the shape is valid the call reaches Facebook and comes back as a missing object rather than a bad argument. The fix is to list first and act second: facebook_list_managed_pages returns the real ids and names, and facebook_get_page_details confirms you have the one you meant. The second cause is a Page you did not include at consent time, which produces the same error because the connection genuinely cannot see it. Reconnect from the dashboard and tick the Page. Note that on a manual build this same error usually means you passed a user token where a Page Access Token was needed; that cause does not apply here, because the hosted connection handles the token exchange.

Ask the agent to diagnose it:

Run facebook_list_managed_pages and show me every Page id and name you can see. Then run facebook_get_page_details on the one I name. Do not retry the failing call until both have returned.

Reading a Page and posting to it are different permissions, so a connection can list your Pages, pull insights, and then fail on the first publish. Two things cause it. The first is a partial approval: Facebook's consent screen lets you decline individual permissions, and declining the publishing one leaves a connection that looks healthy until a write is attempted. Reconnect and accept the full set. The second is your own role on the Page: Facebook requires a role with content permissions to publish, so an account with analyst access can read insights and never post. Check the Page's access settings in Meta Business Suite. facebook_get_page_roles shows what the connected identity holds.

Ask the agent to diagnose it:

Run facebook_get_page_roles for the Page I name and tell me what role the connected account has. Then report the exact error body from the failed publish, including the code and error_subcode.

Two different things produce this. If the name looks right, the tool catalog had not loaded when the first call went out; schemas load on demand, so retrying once normally fixes it. If the agent is reaching for something like metaads_get_insights or a campaign tool, the problem is that it is on the wrong integration. Ad campaigns, ad sets, creatives, and ad spend live on the separate Meta Ads connection, and the Facebook tools here cover Pages, posts, comments, photos, videos, and Messenger. Connect Meta Ads as well if you need both, and say which surface you mean in the prompt.

Ask the agent to diagnose it:

List the Facebook tools you actually have access to. Tell me which one posts to a Page, and whether you also have any Meta Ads tools available.

Try reconnecting Facebook and complete the consent flow in the same browser session. Partial OAuth approvals or switching accounts mid-flow can leave the connection incomplete.

Yes. ClawLink is the fastest way to connect Hermes to Facebook: link your Facebook account once in the browser and Hermes Agent can call the Facebook API through 39 ready-made tools — no custom code or token handling.

Install the plugin with hermes plugins install ClawLink-HQ/hermes-plugin --enable, then pair once: run hermes clawlink begin, approve the link in your browser, and run hermes clawlink finish. Connect Facebook in the dashboard and Hermes can use it from the next message — no config files or API keys to manage.

About two minutes. Sign in, click Connect next to Facebook in the dashboard, authenticate, and Hermes Agent can use it from the next chat message.

The alternative to ClawLink is usually manual OAuth app setup plus your own token handling, permission troubleshooting, and tool plumbing for Hermes Agent. That is fine if you want to build and maintain the integration yourself. Most teams just want Facebook working from chat.

Not for this connection. The requirement is real when you build the integration yourself, because Meta issues tokens to apps rather than to people, so a DIY route means registering an app, adding Facebook Login, requesting the permissions your calls need, and going through App Review before anyone outside your test list can use it. ClawLink connects through an app that is already registered, so what you do instead is approve access on Facebook's own consent screen and pick the Pages you want to include. The permission model does not change: Facebook still decides what the grant covers and still lets you revoke it from your account's Business Integrations settings.

No, and this is the part that trips up most manual builds. Publishing to a Page needs a Page Access Token rather than a user token, which is why so many Stack Overflow answers about error 100 end with "switch to a Page Access Token". That exchange happens inside the hosted connection. From chat you name the Page and call the tool; there is no token type for you to pick, and nothing expires that you have to notice and rebuild.

Only the ones you granted during sign-in, and only if your Facebook account has a role on them. Facebook's consent screen shows a Page picker, and unticking a Page there keeps it out of the connection permanently until you reconnect. Ask the agent to run facebook_list_managed_pages first: whatever comes back is the complete list of what it can touch, which is the fastest way to confirm the connection matches what you intended.

The mechanics worth checking are these. The grant is scoped to the Pages you selected, so a Page you did not tick is not reachable. Nothing about your login is stored: the connection holds an authorization, not a password, and Facebook lists it under Business Integrations in your settings where you can revoke it without involving us. The agent's reach is the 39 named tools on this page rather than an open API session, and the ones that write are individually named, including facebook_create_post, facebook_delete_post, and facebook_send_message. If you want a read-only period while you build trust, say so in the prompt and ask the agent to draft posts for approval rather than publish them; facebook_get_scheduled_posts and facebook_publish_scheduled_post make a review step easy to keep.

Pairing is a two-step handshake: run hermes clawlink begin, approve the link in your browser, then run hermes clawlink finish. If you ran finish before approving, or the approval link expired, run hermes clawlink begin again to get a fresh link. Confirm the plugin was installed with --enable, then verify with hermes clawlink test.