MCPOutils MCP/Connexions/get_account_capabilities

get_account_capabilities

Readconnections:read

What a connected account can do — and if it can't, why and what a person has to do. For one account (connectionId, from list_connected_accounts), for every account of a client (clientId), or for a platform in general (platform, no account). Per operation — publish, edit_post_caption, edit_post, delete_post, read_post_insights, read_comments, reply_comment / reply_message / reply_mention, like_comment, hide_comment, delete_comment, private_reply, read_reviews, reply_review, edit_review_reply, delete_review_reply — it gives `allowed`, and when not: `reason` (unsupported | needs_reconnect | not_connected | suspended | not_offered), `reasonText`, `prerequisite` (what to do first) and `humanHandoff` (a person has to act — reconnect, or finish an assisted publish). `publish.details` carries the post types, caption limit and media rule to write for. Derived from the same rules execution uses, so an allowed operation is one the tool for it will accept. Read-only; never returns credentials.

Paramètres
connectionIdstringfacultatif
One account (from list_connected_accounts)
clientIdstringfacultatif
Every account connected for this client Accepts the database ID or public cli- ID. IDs must belong to the authenticated agency and permitted client.
platformstringfacultatif
Just the network's own capabilities, without an account (e.g. instagram)
limitnumberfacultatif
Accounts per call when using clientId (1–100, default 50)
Exemple de requête
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_account_capabilities",
    "arguments": {}
  }
}
Exemple de réponse
{
  "jsonrpc": "2.0",
  "id": 1,
  "result": {
    "content": [
      {
        "type": "text",
        "text": "..."
      }
    ]
  }
}

La forme de la réponse est indicative — le contenu réel dépend de l'outil et de vos données.