Claude (Sonnet 4.6) on AUTOPILOT seems wrongly integrated, doesnt terminate, stop or do anything, keeps infinate internal conversation
- 主要语言
- Shell
- 星标
- 11.2k
- 派生
- 1.9k
- 平均合并
- 14 小时 16 分钟
- 30 天内合并 PR
- 6
描述
### Describe the bug
Since this was working a month ago something fundamentally seem to have been broken.
When select the mentioned model then nothing results from it, it just keeps infinitely looking on internal conversation, this wasnt the case when i used copilot last month, so something must have been broken in the Anthropic model integration.
### Affected version
GitHub Copilot CLI 1.0.40
### Steps to reproduce the behavior
1. Select Sonnet 4.6 (maybe also other, i only tested this)
2. set on autopilot maybe
3. ask it to do literally anything
4. wont reach a result
### Expected behavior
just reach a conclusion, result, fix whatever, instead of infinitely circling in internal conversation and draining tokens
### Additional context
$PSVersionTable
Name Value
---- -----
PSVersion 7.6.1
PSEdition Core
GitCommitId 7.6.1
OS Microsoft Windows 10.0.26200
Platform Win32NT
PSCompatibleVersions {1.0, 2.0, 3.0, 4.0…}
PSRemotingProtocolVersion 2.4
SerializationVersion 1.1.0.1
WSManStackVersion 3.0
贡献指南
调研方向
首先,在 Windows 10 上的 PowerShell 7.6.1 中选择 autopilot,并在其中选择 Claude Sonnet 4.6,复现 GitHub Copilot CLI 1.0.40 中的问题。将其行为与禁用 autopilot 或使用其他模型时的行为进行比较;当 CLI 得出结论或停止运行,而不是无限循环并耗尽 token 时,即可认为问题已解决。
由索引模型根据 Issue 内容生成。
评估
- 技术栈
- powershell
- 领域
- ai, cli
- Issue 类型
- 缺陷
- 难度
- 4/5
- 预计耗时
- 3-5 天
- 活跃度
- 冷清
- 描述清晰度
- 基本清楚
- 新手友好度
- 45/100