AT Protocol MCP Server

Appearance

block_user ​

Block a user to prevent them from seeing your content and interacting with you.

Authentication ​

Required: Yes (Private tool)

Parameters ​

actor (required) ​

  • Type: string
  • Description: User identifier (DID or handle) to block. Handles are resolved to a DID before the block record is created, because a block record's subject must be a DID.

Response ​

Tool results are returned as stringified JSON text. The illustrative shape is:

typescript
{
  success: boolean;
  message: string;
  blockedUser: {
    actor: string;   // echoes the actor you passed in
    did?: string;
    uri?: string;    // AT-URI of the created block record
  };
}

Examples ​

Block a User ​

json
{
  "actor": "harasser.bsky.social"
}

Response (illustrative):

json
{
  "success": true,
  "message": "User harasser.bsky.social has been blocked. They cannot see your content or interact with you.",
  "blockedUser": {
    "actor": "harasser.bsky.social",
    "uri": "at://did:plc:myuser/app.bsky.graph.block/block123"
  }
}

What Blocking Does ​

  • The blocked user cannot see your posts, and you cannot see theirs.
  • The blocked user cannot reply to your posts or follow you.
  • Existing follow relationships between you are removed.

Error Handling ​

An actor that is neither a DID (did:...) nor a handle (user.domain) raises a VALIDATION_ERROR before any network call:

json
{
  "error": "Actor must be a valid DID (did:...) or handle (user.domain.com)",
  "code": "VALIDATION_ERROR"
}

A handle that cannot be resolved, or other API failures, surface as the underlying AT Protocol error. This tool does not deduplicate or special-case self-blocks; those conditions are handled by the AT Protocol service and returned as its error.

Block vs Mute ​

  • Block: complete separation; the user can tell they are blocked.
  • Mute: soft filter that only affects your own feeds; the user is unaware.

See Also ​