[roomote-roomote]

​Opened on behalf of @daniel-lxs. Follow up by mentioning @roomote-roomote, in the web UI, or in Telegram.

Related issue

Internal maintainer-requested developer reliability work; no issue was opened.

Why this PR exists

• A maintainer explicitly invited this PR in the linked issue or discussion
• I am a maintainer / this is internal Roomote work

The supported Telegram harness replaces its stored inbound message on duplicate delivery or edit. That reset already-recorded bot reactions, making a correctly deduplicated application update look like a lost acknowledgement when inspecting /mock/state.

What changed

Preserve the bot's existing reaction state when recording another delivery of the same chat/message pair. Incoming text still updates; new messages start without reactions, different chats remain isolated, and explicit Bot API reaction replacement/clearing still works. This is a two-line mock-recorder correction with protocol-level regression coverage; live Telegram behavior is unchanged.

How it was tested

• Current-develop baseline c97173a7: real local API, isolated PostgreSQL/Redis and the checked-in Telegram harness received a follow-up for a factory-created active run. The first delivery queued one message and recorded 👀; the exact duplicate correctly returned duplicate:true but the harness erased 👀.
• Exact head 6e32a383: restarted the mock provider and reset only the disposable Redis fixture, then replayed the same updates. Both observations retain 👀; PostgreSQL still has one task/run and Redis has exactly one queued follow-up.
• The new regression failed on develop, then passed with the fix. It also verifies edited text, identical message IDs in different chats, and explicit Bot API reaction clearing.
• All 607 communication-package tests, affected typecheck, fast workspace lint/types and Knip passed. An initial full-package attempt omitted the test environment and could not reach the default database port; the normal dotenv-wrapped rerun passed.
• Browser proof is not applicable to this JSON/protocol harness. The provider is mocked; API, protocol and persistence are real. No real bot messages, model inference, actual worker task execution or hosted/live-provider parity is claimed.

Related PRs

• Worker release packaging recovery #3629 — independent developer/operator reliability change.
• Runtime investigation source map #3635 — independent developer skill correction.

Checklist

• The PR title follows the repo convention: [Fix], [Feat], [Improve], [Refactor], [Docs], or [Chore] followed by a user-facing description
• This PR is small and scoped to one change
• pnpm lint and pnpm check-types pass locally
  • Normal fast lint/type gates and changed-file formatting passed; full aliases were not rerun for this mock-only diff.
• I added tests or included a clear manual validation note above
• I removed secrets, tokens, private keys, and customer data from code, logs, and screenshots
• If this change should appear in the changelog, I ran pnpm changeset

No changeset: mock-provider developer tooling only.

[roomote-community]

No code issues found. Reviewed reaction preservation across replayed messages and edits, chat/message isolation, explicit reaction replacement/clearing, and regression coverage. Current-head CI tests, lint, typecheck, Knip, and CodeQL passed. See session

_{Reviewed 6e32a38}