[Bug]: RemoteConversation.send_message(sender=...) silently drops the sender that LocalConversation records
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Newbie friendliness
- 35/100
- Issue type
- Bug
- Clarity
- Clearly specified
- Activity status
- Active
- Tech stack
- python
- Domain
- backend-api-design
Research direction
Read the RemoteConversation.send_message path and the server's SendMessageRequest for POST /api/conversations/{id}/events, then compare them with LocalConversation's sender handling. The issue identifies a product decision about free-form labels versus verified conversation senders that must be resolved before implementation. Done means the remote path records the sender or raises, with live-server coverage and the goal runner, TaskManager, and delegate behavior tested or documented.
Written by the indexing model from the issue text.
Description
I'm an AI agent helping Engel Nyst (@enyst) with project work.
Is there an existing issue for the same bug?
- I have searched existing issues and this is not a duplicate.
Bug Description
Conversation.send_message(message, sender=...) behaves differently on the two conversation kinds. LocalConversation saves the user MessageEvent with that sender; RemoteConversation posts {"sender": ...} to POST /api/conversations/{id}/events, whose body (SendMessageRequest) has no sender field (only sender_conversation_id, added in #6960). The server answers 200 and saves the message with sender: null, so the label is lost silently.
SDK callers that rely on it: the goal runner (send_message(..., sender="goal")), TaskManager and the delegate tool (sender=parent_name). On a remote conversation their messages are indistinguishable from the user's own.
This is a seam bug between the SDK and the server: the SDK map's S04.conversation-sender (local, passes) and the agent-server map's F06.send-sender-conversation (REST, passes for sender_conversation_id) both hold, but the remote client path between them drops the sender. Found by the daily feature-map pass; not a regression from 0275e714..1414239 (the server never had a sender field and the client code is unchanged in that range).
This needs a maintainer's product decision. The server deliberately only records senders it can vouch for (conversation:<hex> from an existing conversation, with a "[Sent by …]" header). Options I see:
RemoteConversationmapsconversation:<hex>senders tosender_conversation_idand raisesValueErrorfor any other sender (no silent drop; callers like the goal runner would need another way to mark their messages remotely).- The server accepts a free-form
senderlabel from authenticated clients and records it as-is (distinct from the vouchedconversation:form). RemoteConversationkeeps dropping it but warns, and the docstring says so.
Expected Behavior
A sender passed to RemoteConversation.send_message is either recorded on the saved message as it is locally, or refused with an error; it is never silently discarded.
Actual Behavior
Local: [["goal", "QA_SEAM_LOCAL"]]. Remote, same program and SDK: [[null, "QA_SEAM_REMOTE"]]. The REST call returns 200.
Steps to Reproduce
On main@141423921fdaeb08cd5f53464a69e860da70b66f, with the verify-agent-server skill's CLI on PATH (.agents/skills/verify-agent-server/scripts):
uv sync --dev
export AGENT_SERVER_VERIFY_RUN=$(control-agent-server launch --new --name sender-seam --print-run)
export F=$AGENT_SERVER_VERIFY_RUN/fixtures/seam; mkdir -p $F
curl -sL https://raw.githubusercontent.com/enyst/openhands-evidence/128dce7746a4a625f149347e3f3075627d95dac2/software-agent-sdk/bugs/2026-11-19/remote-sender/seam.py -o $F/seam.py
control-agent-server exec --env F="$F" --save sender-seam/probe -- .venv/bin/python $F/seam.py # prints SEAM_RESULT {...}
control-agent-server stop
REST-only form: control-agent-server api POST /api/conversations/$CID/events --json '{"role":"user","content":[{"type":"text","text":"x"}],"sender":"qa-x"}' returns 200, and GET /api/conversations/$CID/events/search shows the message with sender: null.
Acceptance Criteria
Once the option is chosen:
-
RemoteConversation.send_message(msg, sender=X)never results in a saved message withsender: nullwhile returning normally: it records X (or itsconversation:mapping) or raises. - A test runs the same
send_message(..., sender=...)throughLocalConversationandRemoteConversationagainst a live server and asserts the same outcome. - The goal runner,
TaskManagerand delegate tool behave as decided on remote conversations (covered by a test or documented).
Installation Method
From source (uv)
SDK Version
main@141423921fdaeb08cd5f53464a69e860da70b66f (openhands-sdk 1.62.0)
Version Confirmation
- I have confirmed this bug exists on the LATEST version of OpenHands SDK
Python Version
3.13
Model Name (if applicable)
None (no model call; dummy LLM config)
Operating System
Linux
Logs and Error Messages
Saved exchange and program: [probe.json](https://github.com/enyst/openhands-evidence/blob/128dce7746a4a625f149347e3f3075627d95dac2/software-agent-sdk/bugs/2026-11-19/remote-sender/probe.json), seam.py, [README](https://github.com/enyst/openhands-evidence/blob/128dce7746a4a625f149347e3f3075627d95dac2/software-agent-sdk/bugs/2026-11-19/remote-sender/README.md).
Minimal Code Sample
remote = Conversation(agent=agent, workspace=RemoteWorkspace(host=url, api_key=key, working_dir=d))
remote.send_message("hi", sender="goal")
[e.sender for e in remote.state.events if isinstance(e, MessageEvent) and e.source == "user"] # [None]; locally ["goal"]
Generated by Claude Code
- Dominant language
- Python
- Stars
- 1.4k
- Forks
- 797
- Avg merge
- 1d 17h
- Merged PRs (30d)
- 354
Getting set up
- No Dockerfile or Docker Compose file
- Has a pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from OpenHands/software-agent-sdk
-
[Bug]: `list_directory` reports a complete listing as truncated when a directory has exactly 500 entriesPossibly taken @balabommablock-cpu claimed this 29 days ago. Openbug priority:low ready-for-dev
Difficulty 2/5 1-3 hours Newbie friendliness 87/100
OpenHands/software-agent-sdk#6133 ·
Maintainers usually reply within 1 day
-
[Bug]: Command hook that exits 1 is treated as blocking when its stdout is a JSON `deny` or `continue: false`Possibly taken A pull request linked to this issue is open or already merged. Openbug priority:low ready-for-dev
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
OpenHands/software-agent-sdk#6116 · 3 comments ·
Maintainers usually reply within 1 day
-
[Bug]: TypeScript GET-with-body requests remain pending after a truncated responsePossibly taken @BlueRaddish claimed this 34 days ago. Openbug priority:low ready-for-dev
Difficulty 2/5 1-3 hours Newbie friendliness 85/100
OpenHands/software-agent-sdk#6059 ·
Maintainers usually reply within 1 day
-
[Bug]: Document whether `prompt_tokens` includes `cache_read_tokens` in accumulated token usagePossibly taken @Jaxon1216 claimed this 36 days ago. Openpriority:low ready-for-dev
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
OpenHands/software-agent-sdk#5975 ·
Maintainers usually reply within 1 day
-
priority:low
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
OpenHands/software-agent-sdk#5708 · 1 comment ·
Maintainers usually reply within 1 day
All issues in OpenHands/software-agent-sdk
Similar issues
-
[Bug]: reme CLI crashes when the running server used a relative config pathPossibly taken @nurlanmnn claimed this today. Open
Difficulty 2/5 1-3 hours Newbie friendliness 88/100
agentscope-ai/ReMe#770 ·
Maintainers usually reply within 1 day
-
Pi: paused model calls are reported as "not signed in" when the model's provider comes from an extensionPossibly taken @balcsida claimed this today. Open
Difficulty 2/5 1-3 hours Newbie friendliness 73/100
microsoft/TokenEconomizer#288 ·
Maintainers usually reply within 1 day
-
Difficulty 1/5 Under an hour Newbie friendliness 67/100
sickn33/agentic-awesome-skills#2437 ·
Maintainers usually reply within 1 day
-
site-bug triage
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
ethereum/consensus-specs#5846 ·
Maintainers usually reply within 1 day