feat: add require_mention option for Telegram group chats

[butnull]

Summary

Add support for mention-based response in Telegram group chats. When enabled, the bot will only respond in groups when explicitly @mentioned, preventing noisy conversations when multiple bots are present.

Changes

• gateway/config.py: Add config loading for telegram.require_mention via env TELEGRAM_REQUIRE_MENTION
• gateway/platforms/telegram.py:
  • Fetch bot username on connect for mention checking
  • Add _is_bot_mentioned() helper method
  • Check mention in _handle_text_message() for group chats

Configuration

config.yaml:

Environment variable:

Behavior

Chat Type	require_mention=true	require_mention=false
DM (Private)	Always respond	Always respond
Group/Supergroup	Only respond if @mentioned	Always respond

Use Case

This feature is useful when running multiple AI agents (e.g., Hermes + OpenClaw) in the same Telegram group - users can control which bot responds by @mentioning the specific one they want.

[nidhishgajjar]

Orb Code Review (powered by GLM 5.1 on Orb Cloud)

Summary

This PR adds a require_mention option for Telegram group chats, allowing the bot to only respond when @mentioned. It includes config loading from YAML and a mention-detection method.

Architecture

Two changes: gateway/config.py (YAML→env var bridge) and gateway/platforms/telegram.py (mention check in _handle_text_message).

Issues

🔴 Critical: Duplicate of existing functionality

The Telegram adapter already has a complete require_mention system in the codebase:

1. _telegram_require_mention() (line 1988) — already reads from config.extra or TELEGRAM_REQUIRE_MENTION env var
2. _should_process_message() (line 2124) — already implements group chat filtering with mention checks
3. _message_mentions_bot() (line 2077) — already detects @mentions via text and entities

The PR adds duplicate code:

• New self._require_mention field → duplicates _telegram_require_mention()
• New _is_bot_mentioned() → duplicates _message_mentions_bot() (but less complete)
• New check in _handle_text_message() → runs before the existing _should_process_message() check on line 2165, making it doubly redundant

🟡 Moderate: New mention check is less complete

The new _is_bot_mentioned() method is a subset of the existing _message_mentions_bot():

• ❌ Doesn't handle text_mention entities (users with hidden usernames)
• ❌ Doesn't check message captions
• ❌ Requires a separate get_me() API call to fetch bot username, while existing code uses self._bot.username directly

🟢 Useful: Config.py YAML bridge

The gateway/config.py change that reads telegram.require_mention from YAML config and bridges it to the env var is genuinely useful. This is the only part of the PR that adds value — the rest duplicates what _telegram_require_mention() already handles via config.extra.

Suggestions

1. Keep only the config.py change — it correctly bridges YAML config to the env var that the existing system already reads
2. Remove the duplicate code from telegram.py — the existing _should_process_message() → _message_mentions_bot() flow is more complete and already works
3. If the goal was to fix a bug where require_mention wasn't working, investigate why _telegram_require_mention() wasn't picking up the config (likely the YAML key wasn't being loaded into config.extra)

Cross-file Impact

Touches gateway/config.py and gateway/platforms/telegram.py. The config.py change is clean. The telegram.py changes conflict with existing logic.

Assessment

Request changes — The implementation is largely a duplicate of existing functionality. Keep the config.py YAML bridge (useful), but remove the redundant telegram.py changes and instead verify that the existing _telegram_require_mention() path correctly picks up the YAML-loaded env var.

[teknium1]

Thanks for the contribution @butnull — closing as redundant. PR #3870 (merged March 29) added telegram.require_mention to main with the same behavior you proposed here. Your PR predated it by over a week and helped surface the need. Appreciate the work.