fix(feishu): post into existing topic thread when reply_to is absent by konsisumer · Pull Request #16603 · NousResearch/hermes-agent

konsisumer · 2026-04-27T14:57:51Z

What changed and why

When a user sends a message inside a Feishu topic thread, the inbound event carries a thread_id. Hermes stored it in metadata and used it correctly when reply_to was set (via reply_in_thread=True on the message.reply API). However, the create path — reached when reply_to is None (progress/thinking messages, approval cards, or after the withdrawn-message fallback) — called message.create with receive_id_type="chat_id", completely ignoring thread_id. This caused every such message to start a brand-new topic instead of replying within the current one.

Fix in _send_raw_message (gateway/platforms/feishu.py):

When reply_to is None and thread_id is present in metadata, use receive_id_type="thread_id" with thread_id as receive_id.
The Feishu API accepts thread_id as a receive-id type and delivers the message inside the existing thread.
This also covers the fallback path (withdrawn message → active_reply_to = None) since metadata is forwarded unchanged.

How to test

Set up Hermes with a Feishu enterprise account and a topic group.
Open an existing topic thread and send a message to Hermes.
Verify that all Hermes responses (including intermediate thinking/progress messages and any approval cards) appear inside the same topic thread, not as new top-level topics.
Run the regression test: pytest tests/gateway/test_feishu.py::FeishuAdapterSendTests::test_send_uses_thread_id_receive_type_when_no_reply_to -v

What platforms tested on

macOS (unit tests only; Feishu enterprise tenant required for live verification)

Fixes #7734

alt-glitch · 2026-04-27T16:18:37Z

Likely duplicate of #11433 — same root cause: Feishu create path ignores thread_id when reply_to is absent.

konsisumer · 2026-04-27T22:22:22Z

Rebased onto origin/main. Local run of tests/gateway/test_feishu.py is green (190 passed). The 60 failing checks in CI are all in files outside this PR's scope (discord, slack, tui, hermes_cli, run_agent, tools) and appear to be pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-04-28T03:13:43Z

Rebased onto origin/main (3 new upstream commits since last push). Rebase completed with no conflicts — our two files (gateway/platforms/feishu.py, tests/gateway/test_feishu.py) were untouched by upstream. Local run of tests/gateway/test_feishu.py confirms 190 passed. The 21 CI failures from the previous run are all in files outside this PR's scope (hermes_cli, tui_gateway, run_agent, tools) and appear to be pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-04-28T07:20:43Z

Rebased onto origin/main (2 new upstream commits since last push: ec8243fe chore/release, 8081425a feat/security-redact). No conflicts — neither touches gateway/platforms/feishu.py or tests/gateway/test_feishu.py. Local run of tests/gateway/test_feishu.py confirmed 190 passed. The 19 CI failures from the previous run are all in files outside this PR's scope (hermes_cli, tui_gateway, run_agent, tools, other gateway tests) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-04-28T11:55:08Z

Rebased onto origin/main (5 new upstream commits since last push). No conflicts — neither gateway/platforms/feishu.py nor tests/gateway/test_feishu.py were touched by upstream. Local run of tests/gateway/test_feishu.py confirmed 190 passed. The 23 CI failures from the previous run are all in files outside this PR's scope (agent, hermes_cli, tui_gateway, run_agent, tools, and unrelated gateway tests) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-04-28T16:13:18Z

Rebased onto origin/main (44 new upstream commits since last push). No conflicts — neither gateway/platforms/feishu.py nor tests/gateway/test_feishu.py were touched by upstream. Local run of tests/gateway/test_feishu.py confirmed 190 passed. The 23 CI failures from the previous run are all in files outside this PR's scope (agent, hermes_cli, tui_gateway, run_agent, tools, and unrelated gateway tests) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-04-28T20:26:24Z

Rebased onto origin/main (5 new upstream commits since last push). No conflicts — neither gateway/platforms/feishu.py nor tests/gateway/test_feishu.py were touched by upstream. Local run of tests/gateway/test_feishu.py confirmed 190 passed. The CI failures are all in files outside this PR's scope (agent, hermes_cli, tui_gateway, tools, and unrelated gateway tests) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-04-29T01:04:41Z

Rebased onto origin/main. All 190 feishu tests pass locally (tests/gateway/test_feishu.py). The failing CI checks (test, nix, nix-lockfile-check) are pre-existing upstream issues: test_session.py imports normalize_whatsapp_identifier which does not exist in gateway/session.py on main, and the nix lockfile failures are infrastructure issues — neither is caused by this PR.

konsisumer · 2026-04-29T08:55:11Z

Rebased onto origin/main (52 new upstream commits since last push). No conflicts — neither gateway/platforms/feishu.py nor tests/gateway/test_feishu.py was touched by upstream. Local run of tests/gateway/test_feishu.py confirmed 190 passed. The CI failures are all in files outside this PR’s scope (agent, hermes_cli, tui_gateway, run_agent, tools, and unrelated gateway tests) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-04-29T13:23:59Z

Rebased onto origin/main. No conflicts — neither gateway/platforms/feishu.py nor tests/gateway/test_feishu.py were touched by upstream. Local run of tests/gateway/test_feishu.py confirmed 190 passed. The CI failures are all in files outside this PR's scope (agent, hermes_cli, tui_gateway, run_agent, tools, and unrelated gateway tests) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-04-29T17:36:41Z

Rebased onto origin/main (41 new upstream commits since last push). No conflicts — neither gateway/platforms/feishu.py nor tests/gateway/test_feishu.py was touched by upstream. Local run of tests/gateway/test_feishu.py confirmed 190 passed. The CI failures are all in files outside this PR's scope (agent, hermes_cli, tui_gateway, run_agent, tools, and unrelated gateway tests) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-05-02T07:51:20Z

Rebased onto latest main (323 commits ahead). Rebase applied cleanly with no conflicts — upstream had also touched feishu.py in two commits but they didn't conflict with our change. All 195 feishu tests pass locally including the new test_send_uses_thread_id_receive_type_when_no_reply_to test.

The nix and nix-lockfile-check CI failures are unrelated to this PR (they affect infrastructure files not in our changeset) and appear to be pre-existing upstream issues.

konsisumer · 2026-05-02T10:12:47Z

Rebased onto origin/main (26 new upstream commits since last push). Rebase applied cleanly with no conflicts — the upstream feishu commit (38dd057e9) touched a different function and did not interact with our change. Local run of tests/gateway/test_feishu.py confirmed 196 passed. The 10 CI failures are all in files outside this PR's scope (tests/acp/test_server.py, tests/gateway/test_teams.py, tests/run_agent/test_concurrent_interrupt.py, tests/hermes_cli/test_update_gateway_restart.py, tests/test_tui_gateway_server.py, tests/tools/test_dockerfile_pid1_reaping.py) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-05-02T19:07:13Z

Rebased onto origin/main (1 new upstream commit: 5d3be898a docs(tts): mention xAI custom voice support). No conflicts — the upstream commit only touches hermes_cli/config.py, hermes_cli/setup.py, and TTS docs; neither gateway/platforms/feishu.py nor tests/gateway/test_feishu.py was affected. Local run of tests/gateway/test_feishu.py confirmed 196 passed. The 15 CI failures from the previous run are all in files outside this PR's scope (tests/acp/test_server.py, tests/gateway/test_teams.py, tests/hermes_cli/, tests/run_agent/, tests/test_tui_gateway_server.py, tests/tools/) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-05-02T23:14:20Z

Rebased onto latest origin/main (already current — 0 commits behind). Local run of tests/gateway/test_feishu.py confirms 196 passed. The 16 CI failures are all in files outside this PR's scope (tests/acp/test_server.py, tests/gateway/test_teams.py, tests/hermes_cli/, tests/run_agent/, tests/test_tui_gateway_server.py, tests/tools/) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-05-03T03:18:02Z

Rebased onto latest origin/main (already current — 0 commits behind). Local run of tests/gateway/test_feishu.py confirms 196 passed. The 15 CI failures are all in files outside this PR's scope (tests/acp/test_server.py, tests/gateway/test_teams.py, tests/hermes_cli/, tests/run_agent/, tests/test_tui_gateway_server.py, tests/tools/) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-05-03T06:05:10Z

Rebased onto origin/main (already current — 0 commits behind). Local run of tests/gateway/test_feishu.py confirmed 196 passed. The 15 CI failures are all in files outside this PR’s scope (tests/acp/test_server.py, tests/gateway/test_teams.py, tests/hermes_cli/, tests/run_agent/, tests/test_tui_gateway_server.py, tests/tools/) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-05-03T07:59:44Z

Rebased onto origin/main (2 new upstream commits: fix(tools): reconfigure enabled unconfigured toolsets and fix(model): avoid bedrock credential probe in provider picker). No conflicts — neither touched gateway/platforms/feishu.py or tests/gateway/test_feishu.py. Local run of tests/gateway/test_feishu.py confirmed 196 passed. The 17 CI failures are all in files outside this PR’s scope (tests/acp/test_server.py, tests/gateway/test_teams.py, tests/hermes_cli/, tests/run_agent/, tests/test_tui_gateway_server.py, tests/tools/) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-05-03T08:59:21Z

Rebased onto origin/main (10 new upstream commits since last push). No conflicts — neither gateway/platforms/feishu.py nor tests/gateway/test_feishu.py was touched by upstream. Local run of tests/gateway/test_feishu.py confirmed 196 passed. The 15 CI failures are all in files outside this PR's scope (tests/acp/test_server.py, tests/gateway/test_teams.py, tests/hermes_cli/, tests/run_agent/, tests/test_tui_gateway_server.py, tests/tools/) and are pre-existing upstream failures unrelated to the Feishu threading fix.

konsisumer · 2026-05-03T17:07:14Z

Closing — deferring to #11433 by @youttiao which addresses the same. Reopen if that PR stalls.

alt-glitch added type/bug Something isn't working P2 Medium — degraded but workaround exists platform/feishu Feishu / Lark adapter comp/gateway Gateway runner, session dispatch, delivery labels Apr 27, 2026

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from cbc3fbf to 5ee633d Compare April 27, 2026 22:22

konsisumer mentioned this pull request Apr 27, 2026

[Bug]: Title: Two issues with the Feishu plugin #7734

Open

1 task

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from 5ee633d to 1c3cca1 Compare April 28, 2026 03:13

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from 1c3cca1 to 7472b87 Compare April 28, 2026 07:20

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from 7472b87 to a8ef800 Compare April 28, 2026 11:55

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from a8ef800 to 91cbf10 Compare April 28, 2026 16:13

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from 91cbf10 to 28e65be Compare April 28, 2026 20:26

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from 28e65be to ea3db18 Compare April 29, 2026 01:04

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from ea3db18 to df0c82a Compare April 29, 2026 08:55

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from df0c82a to e179810 Compare April 29, 2026 13:23

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from e179810 to fc31dc6 Compare April 29, 2026 17:36

jlcbk mentioned this pull request Apr 30, 2026

Feishu topic progress messages can still create new topics #17875

Closed

highland0971 mentioned this pull request Apr 30, 2026

fix(feishu): comprehensive topic thread reply fix (A+B+C+D) #18121

Closed

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from fc31dc6 to 98fbb90 Compare May 2, 2026 07:51

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch 2 times, most recently from d12d79d to a16fc4f Compare May 2, 2026 19:07

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from a16fc4f to 887cc39 Compare May 2, 2026 23:14

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from 887cc39 to 11d118c Compare May 3, 2026 03:18

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch 2 times, most recently from f195451 to a9cc626 Compare May 3, 2026 07:59

fix(feishu): post into existing topic thread when reply_to is absent

9b94e88

konsisumer force-pushed the fix/feishu-topic-thread-create-path branch from a9cc626 to 9b94e88 Compare May 3, 2026 08:59

konsisumer closed this May 3, 2026

Conversation

konsisumer commented Apr 27, 2026

What changed and why

How to test

What platforms tested on

Uh oh!

alt-glitch commented Apr 27, 2026

Uh oh!

konsisumer commented Apr 27, 2026

Uh oh!

konsisumer commented Apr 28, 2026

Uh oh!

konsisumer commented Apr 28, 2026

Uh oh!

konsisumer commented Apr 28, 2026

Uh oh!

konsisumer commented Apr 28, 2026

Uh oh!

konsisumer commented Apr 28, 2026

Uh oh!

konsisumer commented Apr 29, 2026

Uh oh!

konsisumer commented Apr 29, 2026

Uh oh!

konsisumer commented Apr 29, 2026

Uh oh!

konsisumer commented Apr 29, 2026

Uh oh!

konsisumer commented May 2, 2026

Uh oh!

konsisumer commented May 2, 2026

Uh oh!

konsisumer commented May 2, 2026

Uh oh!

konsisumer commented May 2, 2026

Uh oh!

konsisumer commented May 3, 2026

Uh oh!

konsisumer commented May 3, 2026

Uh oh!

konsisumer commented May 3, 2026

Uh oh!

konsisumer commented May 3, 2026

Uh oh!

konsisumer commented May 3, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants