Feature Request: Auto-resolve Discord user mentions from <@id> to @username

Problem

When a Discord user sends a message like @username message, the bot receives <@USER_ID> message in the content field. While Discord API provides resolved user data in the mentions array, OpenClaw currently passes the raw <@id> format to the LLM, making it hard for agents to understand who is being mentioned.

Current Behavior

{
  "content": "<@USER_ID_PLACEHOLDER> hello",
  "mentions": [{
    "id": "USER_ID_PLACEHOLDER",
    "username": "username",
    "global_name": "Display Name"
  }]
}

LLM receives: <@USER_ID_PLACEHOLDER> hello

Desired Behavior

LLM should receive: @Display Name hello (using global_name or username)

Proposed Solution

Add a Discord message preprocessor that:

1. Parses the mentions, mention_roles, and mention_channels arrays
2. Replaces <@id>, <@&role_id>, and <#channel_id> with human-readable names
3. Preserves the original format for sending messages back to Discord

Example mapping:

• <@USER_ID> → @DisplayName
• <@&ROLE_ID> → @RoleName
• <#CHANNEL_ID> → #channel-name

Use Cases

1. Multi-agent conversations where agents need to understand who is being mentioned
2. Role-based triggers (related to Feature Request: Discord role-mention triggers #7717)
3. Better context understanding for AI agents

Workaround

Currently, we have to manually configure user ID mappings in each agent's SOUL.md:

- `<@USER_ID>` = @username
- `<@USER_ID_2>` = @username2

This is tedious and error-prone when adding new users or agents.

Related Issues

• Feature Request: Discord role-mention triggers #7717: Feature Request: Discord role-mention triggers
• [Bug]: Discord - multiple agent bots are considered to be the same bot and filtered out when talking to each other #11199: Multiple agent bots communication issues