ask_user marks an active app user unavailable, then continues with other actions
Nessuno ha ancora preso questa issue.
Valutazione
- Difficoltà
- 4/5
- Tempo stimato
- 3-5 giorni
- Idoneità per principianti
- 45/100
Direzione di ricerca
The issue is in the ask_user flow of the GitHub Copilot desktop app. Start by finding where ask_user is invoked and how the app determines user availability. Look at the activity log generation and the logic that auto-responds. Check the related issue #3606 for prior context. Testing will require reproducing the scenario in an active Windows session to see the fabricated response.
Scritto dal modello di indicizzazione a partire dal testo della issue.
Descrizione
Short summary
The app displays an ask_user question, immediately answers it on the user's behalf with “The user is not available to respond,” and lets the agent continue with other actions even though the user is actively present in the app.
Affected version or release
GitHub Copilot app with Copilot CLI 1.0.87-0
Installation context
GitHub Copilot desktop app on Windows; active interactive session.
What happened?
While I was actively in the app, the agent asked me to complete a Microsoft device-login step. Instead of allowing me to answer, the app generated this response on my behalf:
The user is not available to respond and will review your work later. Work autonomously and make good decisions.
That availability claim is false: I was in the app at the time. The agent then moved on to other work rather than waiting for the requested input.
This is a recurrence of #3606, reported on September 7, 2026. The attached screenshot was captured on September 22, 2026, so the problem remains reproducible two weeks later. In this occurrence, the question is visible in the activity log before the fabricated unavailable response.
Steps to reproduce
- Open and actively use a session in the GitHub Copilot app.
- Have the agent reach a step that requires user input, such as device authentication.
- Let the agent invoke
ask_user. - Observe that the question appears in the activity log.
- Observe that the app immediately supplies an unavailable-user answer and the agent continues with unrelated work, without giving the active user a chance to respond.
Expected behavior
The app should present the question to the active user and pause for a response. It must not infer that the user is unavailable while they are actively using the session, and it must not continue past a blocked authentication or approval step based on a fabricated answer.
Additional context
Related report: #3606
- Lingua principale
- Nessun dato sulla lingua
- Stelle
- 2.1k
- Fork
- 157
- Metriche di merge delle PR
- Nessuna PR unita negli ultimi 30g
Preparare l'ambiente
Come iniziare
- Leggi tutta la issue e poi la guida ai contributi del progetto.
- Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
- Fai un fork del repository e lavora su un branch.
- Apri una pull request che faccia riferimento al numero della issue.
Altre issue di github/app
-
triage
Difficoltà 2/5 1-3 ore Idoneità per principianti 70/100
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 65/100
-
When using GPT-5.Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 70/100
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 70/100
-
Difficoltà 2/5 1-3 ore Idoneità per principianti 70/100
Issue simili
-
Document an `openai-api` service example for uncensored-model endpoints (red-teaming use case)Aperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 72/100
UKGovernmentBEIS/inspect_ai#5592 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
[Bug] Choice-select parser accepts non-positive answer numbers, silently indexing the last nodeAperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 84/100
run-llama/llama_index#23273 ·
I maintainer di solito rispondono entro 2 giorni
-
hermes-plugin: agent_id lookup ignores agent_identity, so calls are attributed to the default agentAperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 78/100
TencentCloud/TencentDB-Agent-Memory#1545 · 1 commento ·
I maintainer di solito rispondono entro 1 giorno
-
fix(ai): avoid text tool call recovery middleware on native OpenAI responses/completions streamsAperta
Difficoltà 2/5 1-3 ore Idoneità per principianti 84/100
code-yeongyu/senpi#2203 ·
I maintainer di solito rispondono entro 1 giorno
-
area:ai-suggestions bug good first issue P2
Difficoltà 2/5 1-3 ore Idoneità per principianti 84/100
uttrflow/uttrflow-swift#1922 ·
I maintainer di solito rispondono entro 1 giorno