Error catching telegram messages locally on my laptop, via Telegram connector

[artemu78]

Existing issues

• I searched open issues and this is not already reported.

KiroCrew version

kirocrew 0.1.3

How is it installed?

Desktop app

Platform

macOS

What happened

I've connected my own Telegram bot via Settings > Channels > Telegram.
Every time i type something in my telegram i face the response: "⚠️ Error — please try again"

Steps to reproduce

1 Open Kiro Crew Channels interface
2 type there telegram bot secret and user id
3 go to Telegram and try to say something to your bot

face the text: "⚠️ Error — please try again"

Relevant log output

2026-08-06 14:03:26,772 ERROR kiro_crew.telegram.transport_dispatch: Telegram transport_dispatch: error handling message
Traceback (most recent call last):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/telegram/transport_dispatch.py", line 370, in handle_message
    accumulated = await driver.run(full_message)
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/messaging/driver.py", line 354, in run
    async for event in self.provider.stream(message):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/providers/acp.py", line 984, in stream
    async for e in self._client.stream_events(message):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/acp/session_provider.py", line 220, in stream
    async for event in self._handle.prompt(message):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/acp/session_handle.py", line 457, in prompt
    async for event in self._dispatch_events(req_id, timeout):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/acp/session_handle.py", line 1204, in _dispatch_events
    _raise_acp_error(msg.error, self._advertised_model_ids())
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/acp/client.py", line 1085, in _raise_acp_error
    raise AcpError(formatted, transient=_is_transient_raw_error(error, available_models))
kiro_crew.acp.client.AcpError: Your account does not have access to model 'claude-opus-4.8'. Available to you: auto, claude-sonnet-4.5, claude-sonnet-4, claude-haiku-4.5, deepseek-3.2, minimax-m2.5, minimax-m2.1, glm-5 (+1 more). Pick one in the model picker, or set agent.model to 'auto' to let the backend choose a model your plan includes. Retrying will not help. (request_id: 4c0bca24-8a4c-4eea-b6d1-b96c400ac4e7)
2026-08-06 14:04:44,302 ERROR kiro_crew.telegram.transport_dispatch: Telegram transport_dispatch: error handling message
Traceback (most recent call last):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/telegram/transport_dispatch.py", line 370, in handle_message
    accumulated = await driver.run(full_message)
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/messaging/driver.py", line 354, in run
    async for event in self.provider.stream(message):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/providers/acp.py", line 984, in stream
    async for e in self._client.stream_events(message):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/acp/session_provider.py", line 220, in stream
    async for event in self._handle.prompt(message):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/acp/session_handle.py", line 457, in prompt
    async for event in self._dispatch_events(req_id, timeout):
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/acp/session_handle.py", line 1204, in _dispatch_events
    _raise_acp_error(msg.error, self._advertised_model_ids())
  File "/Applications/KiroCrew.app/Contents/Resources/backend-dist/kirocrew-backend-arm64/lib/python3.12/site-packages/kiro_crew/acp/client.py", line 1085, in _raise_acp_error
    raise AcpError(formatted, transient=_is_transient_raw_error(error, available_models))
kiro_crew.acp.client.AcpError: Your account does not have access to model 'claude-opus-4.8'. Available to you: auto, claude-sonnet-4.5, claude-sonnet-4, claude-haiku-4.5, deepseek-3.2, minimax-m2.5, minimax-m2.1, glm-5 (+1 more). Pick one in the model picker, or set agent.model to 'auto' to let the backend choose a model your plan includes. Retrying will not help. (request_id: 1b4c92ff-3219-42a4-89e7-0edc09af6e6a)

Anything else

I see it nagging about the model, but how I can change it, there are no tools for that, at least i don't see

[bolichen97]

🤖 Kiro Crew Auto-Pipeline [operator: bolichen97]: Thanks for pasting the log — it contains the whole answer, and the underlying cause is not Telegram. The turn fails with AcpError: Your account does not have access to model 'claude-opus-4.8', and the backend already produced a fully actionable message naming the models your plan does include and telling you to pick one in the model picker or set agent.model to auto. So the immediate unblock is to switch that session's agent off claude-opus-4.8. The defect this issue documents is that none of that text ever reaches you: handle_message catches the exception, logs it, and discards it (telegram/transport_dispatch.py:436-437), and the finally block's renderer.close() then finalizes the placeholder to the generic ⚠️ Error — please try again (telegram/renderer.py:767) — advice that is actively wrong here, since the error itself says retrying will not help. Any non-transient failure is therefore indistinguishable from a blip, and the user has no way to see the remedy without reading gateway logs. Starting an automated fix: carry a bounded, redacted failure reason from handle_message into the renderer's finalization so a non-transient error surfaces its own message instead of the generic retry line. Discord has the identical pattern at discord/renderer.py:541; that will be noted for a follow-up rather than widened here. This issue is now claimed by bolichen97's instance.

[bolichen97]

🤖 Kiro Crew Auto-Pipeline [operator: bolichen97]

Fix opened: #2565

A permanent turn failure (e.g. model entitlement) now surfaces its own bounded, sanitized reason in the Telegram reply instead of the generic '⚠️ Error — please try again'. Transient failures keep the retry wording. Note: your immediate unblock is switching the session's agent off the unavailable model — the fixed error message now tells you exactly that in chat.

[artemu78]

I've checked /Users/<myuser>/.kiro/crew/config.json
I mean there is "model": "auto" is set, still i face the error

config 2.json

{
  "meta": {
    "lastTouchedVersion": "0.1.3",
    "lastTouchedAt": "2026-08-06T11:05:35.450054+00:00"
  },
  "agent": {
    "approval_mode": "auto",
    "streaming": true,
    "model": "auto",

See full config.json in attachment.

[bolichen97]

🤖 Kiro Crew Auto-Pipeline [operator: bolichen97]

PR #2565 is review-ready: all 50 checks green, readiness:passed on head 17953db. Awaiting human merge.