PR #15227: Fix Feishu group mention config

• Repository: NousResearch/hermes-agent
• Author: systemoutprintlnnnn
• State: open | merged: False
• Link: https://github.com/NousResearch/hermes-agent/pull/15227

Description (problem / solution / changelog)

Fixes #15226

Summary

• Bridge Feishu group reply settings from config.yaml into runtime PlatformConfig.extra.
• Add require_group_mention support in the Feishu adapter with a secure default of true.
• Coerce common boolean representations so string values like "false" work as expected.
• Add regression tests for config bridging and unmentioned group message routing.

Verification

pytest tests/gateway/test_feishu.py tests/gateway/test_config.py -q
218 passed

Notes

This keeps group policy checks in place before accepting unmentioned group messages. Disabling mention requirements only routes messages when the existing sender/chat policy allows them.

Changed files

• gateway/config.py (modified, +6/-0)
• gateway/platforms/feishu.py (modified, +23/-0)
• tests/gateway/test_config.py (modified, +20/-0)
• tests/gateway/test_feishu.py (modified, +11/-0)

PR #15525: fix(gateway): support require_mention toggle for Feishu group chats

• Repository: NousResearch/hermes-agent
• Author: foreverxdord
• State: open | merged: False
• Link: https://github.com/NousResearch/hermes-agent/pull/15525

Description (problem / solution / changelog)

Summary

Feishu is the only platform adapter that ignores the require_mention config option. This PR adds support for it, allowing users to configure the bot to respond in group chats without requiring an @mention.

Problem

Setting require_mention: false in config.yaml under the Feishu platform block has no effect because feishu.py never reads that config value — _should_accept_group_message() is hardcoded to always require @mention. All other platforms (Telegram, Discord, Slack, WhatsApp, DingTalk, Matrix) implement this feature.

Related open issues: #15226, #5465, #10275, #9835

Changes

gateway/platforms/feishu.py — 24 lines added, 1 line modified:

• Modified _should_accept_group_message() (line 3628): Added early-return bypass before the mention-check logic — if _feishu_require_mention() returns False or the chat_id is in _feishu_free_response_chats(), the message is accepted immediately.

• Added _feishu_require_mention(): Reads self.config.extra.get("require_mention"), handles string/bool types, defaults to True for backward compatibility.

• Added _feishu_free_response_chats(): Reads self.config.extra.get("free_response_channels") with FEISHU_FREE_RESPONSE_CHATS env var fallback. Supports both list and comma-separated string formats for per-chat granularity.

Design decisions

• Pattern aligns with telegram.py implementation exactly
• DMs (chat_type == "p2p") are unaffected — the mention gate only applies to group messages
• @_all (Feishu's @everyone) passthrough is preserved
• Defaults to True (require mention) for backward compatibility — existing users are not affected
• No changes to config.yaml or gateway/config.py needed — the generic bridging loop already passes require_mention into feishu's config.extra

Test Plan

• Running in production with require_mention: false — bot responds to all group messages without @mention
• DMs continue to work normally
• Unit tests for new methods (happy to add if maintainers want)

Closes #15226

Changed files

• gateway/platforms/feishu.py (modified, +24/-1)