name: submit-github-issue description: Submit a GitHub issue to the openclaw/openclaw repository

GitHub Issue Report: Telegram Bot Own Message Triggers Pairing

Summary

When a Telegram bot sends a message in a DM (direct message) conversation, the bot's own message is processed by the Gateway and triggers a pairing request because the bot's user ID is not in the allowFrom list. This creates an unnecessary pairing challenge from the bot to itself.

Bug Details

Root Cause

In extensions/telegram/src/bot-message-context.ts, the buildTelegramMessageContext function unconditionally calls enforceTelegramDmAccess for all messages, including messages sent by the bot itself:

if (
  !(await enforceTelegramDmAccess({
    isGroup,
    dmPolicy: effectiveDmPolicy,
    msg,
    chatId,
    effectiveDmAllow,
    accountId: account.accountId,
    bot,
    logger,
    upsertPairingRequest,
  }))
) {
  return null;
}

The enforceTelegramDmAccess function (in dm-access.ts) checks if the sender is in the allowFrom list. Since the bot's own ID is typically not in this list (only human users are), it triggers a pairing request:

if (dmPolicy === "pairing") {
  // ... creates pairing challenge
}

Evidence

1. Pairing store record showing bot's own ID triggered pairing:

{
  "id": "8492386395",
  "code": "PK2753K5",
  "createdAt": "2026-03-24T23:43:30.448Z",
  "lastSeenAt": "2026-03-24T23:45:18.496Z",
  "meta": {
    "username": "lancelot_clawdbot",
    "firstName": "Lancelot",
    "accountId": "default"
  }
}

1. Code analysis shows:
   • Reaction handler has bot filtering: if (user?.is_bot) return; (line 790 in bot-handlers.runtime.ts)
   • Regular message handler does NOT have bot filtering
   • buildTelegramMessageContext unconditionally calls enforceTelegramDmAccess

Similar Issues

Related to #32256 (forum topic system messages) and #46567 (pairing persistence), but this is a distinct issue specifically about bot's own messages triggering pairing.

Expected Behavior

Bot's own messages should be filtered out and not processed for pairing checks, similar to how reaction events are filtered:

// In bot-handlers.runtime.ts, around line 1726
bot.on("message", async (ctx) => {
  const msg = ctx.message;
  if (!msg) return;
  
  // Add bot message filtering
  if (msg.from?.id === ctx.me?.id) {
    return; // Ignore bot's own messages
  }
  
  // ... rest of handler
});

Actual Behavior

Bot's own messages are processed through the full pipeline, triggering unnecessary pairing requests when dmPolicy is set to "pairing".

Environment

• OpenClaw version: 2026.3.23-2
• OS: Linux 5.15.102-1-pve
• Node: v22.22.0
• Telegram Bot: @lancelot_clawdbot

Configuration

{
  "telegram": {
    "enabled": true,
    "dmPolicy": "pairing",
    "allowFrom": ["426518660"]
  }
}

Impact

• Low severity: pairing request is created but doesn't cause infinite loop due to existingIdx check in pairing store
• Creates unnecessary pairing entries
• May confuse users who see pairing messages from the bot

Proposed Fix

Add bot message filtering in the message handler:

// In extensions/telegram/src/bot-handlers.runtime.ts
bot.on("message", async (ctx) => {
  const msg = ctx.message;
  if (!msg) return;
  
  // Filter out bot's own messages
  if (msg.from?.is_bot && msg.from.id === ctx.me?.id) {
    return;
  }
  
  // ... existing logic
});

Alternatively, add the check in buildTelegramMessageContext before calling enforceTelegramDmAccess.

Additional Context

• The issue only occurs in DM conversations (not groups)
• The issue only occurs when dmPolicy is set to "pairing"
• The pairing store's existingIdx mechanism prevents infinite loops, but the first occurrence still creates a pairing request