[GH-ISSUE #303] feat: Auto tool approve should more closely follow the industry standard #216

New issue

Open

opened 2026-10-18 14:03:51 +00:00 by bsakaguchi · 1 comment

bsakaguchi commented

2026-10-18 14:03:51 +00:00

Owner

Originally created by @Classic298 on GitHub (Sep 22, 2026).
Original GitHub issue: https://github.com/open-webui/computer/issues/303

Right now the tool auto approve is relatively useless

it permanently blocks the chat, which hinders work of AI agents while the user is AFK
- if a tool call gets denied by the classifier/task model it should also be denied in chat
currently the task model gets zero context
- it should get a summary or a description of the main task or however else
- currently it too often blocks correct and harmless tool calls because it doesnt know what the task is
- in the light of this perhaps a small classifier model support like jev would also be cool but that is secondary of this issue

Originally created by @Classic298 on GitHub (Sep 22, 2026). Original GitHub issue: https://github.com/open-webui/computer/issues/303 Right now the tool auto approve is relatively useless - it permanently blocks the chat, which hinders work of AI agents while the user is AFK - if a tool call gets denied by the classifier/task model it should also be denied in chat - currently the task model gets zero context - it should get a summary or a description of the main task or however else - currently it too often blocks correct and harmless tool calls because it doesnt know what the task is - in the light of this perhaps a small classifier model support like jev would also be cool but that is secondary of this issue

bsakaguchi commented

2026-10-18 14:03:52 +00:00

Author

Owner

@owui-terminator[bot] commented on GitHub (Sep 22, 2026):

🔍 Related Issues Found

I found some existing issues that might be related. Please check if any of these are duplicates or contain helpful solutions:

🟣 #181 feat: make auto approve actually auto approve
Directly about auto-approve behavior and the task model deciding based on task context. The new issue expands on the same feature, asking that auto approve be less blocking and that the task model receive a task summary/context.
by Classic298
🟣 #13 feat: tool execution policies / allow once, always allow, "auto mode"
Describes the same 'auto mode' concept inspired by Claude Code, where a small task model decides whether a tool is safe and explicitly says it gets context from the conversation. This is closely aligned with the new request to supply task context and make auto approval work like industry-standard agent tooling.
by Classic298
🟢 #185 feat: approve and remember for tool calls
This open feature request is about approving tool calls and remembering the decision, which overlaps with the new issue's broader tool-approval UX and auto-approval flow. It is related enough to the approval-policy side of the request.
by Classic298
🟢 #213 bug: a stalled agent session hangs the chat task forever, and the whole transcript is lost if you kill it
Although broader, it describes an agent session hanging forever and the chat/task getting stuck, which matches the new issue's complaint that auto tool approve permanently blocks the chat and hinders agents while the user is AFK.
by ThalynLabs

💡 If your issue is a duplicate, please close it and add any additional details to the existing issue instead.

This comment was generated automatically. React with 👍 if helpful, 👎 if not.

@owui-terminator[bot] commented on GitHub (Sep 22, 2026):  🔍 **Related Issues Found** I found some existing issues that might be related. Please check if any of these are duplicates or contain helpful solutions: 1. 🟣 [#181](https://github.com/open-webui/computer/issues/181) **feat: make auto approve actually auto approve** *Directly about auto-approve behavior and the task model deciding based on task context. The new issue expands on the same feature, asking that auto approve be less blocking and that the task model receive a task summary/context.* *by Classic298* 2. 🟣 [#13](https://github.com/open-webui/computer/issues/13) **feat: tool execution policies / allow once, always allow, "auto mode"** *Describes the same 'auto mode' concept inspired by Claude Code, where a small task model decides whether a tool is safe and explicitly says it gets context from the conversation. This is closely aligned with the new request to supply task context and make auto approval work like industry-standard agent tooling.* *by Classic298* 3. 🟢 [#185](https://github.com/open-webui/computer/issues/185) **feat: approve and remember for tool calls** *This open feature request is about approving tool calls and remembering the decision, which overlaps with the new issue's broader tool-approval UX and auto-approval flow. It is related enough to the approval-policy side of the request.* *by Classic298* 4. 🟢 [#213](https://github.com/open-webui/computer/issues/213) **bug: a stalled agent session hangs the chat task forever, and the whole transcript is lost if you kill it** *Although broader, it describes an agent session hanging forever and the chat/task getting stuck, which matches the new issue's complaint that auto tool approve permanently blocks the chat and hinders agents while the user is AFK.* *by ThalynLabs* --- 💡 If your issue is a duplicate, please close it and add any additional details to the existing issue instead. *This comment was generated automatically.* React with 👍 if helpful, 👎 if not.

No labels

enhancement

pull-request

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

open-webui/computer#216

No description provided.

Rows
Columns