agentscope-ai / agentscope-ai/QwenPaw
[Feature]: 建议优化多步骤回复的消息聚合,避免连续发送多条碎片化消息 Feature Request — Aggregate Multi-Step Agent Responses to Avoid Message Spam
- Ngôn ngữ chính
- Python
- Star
- 34.9k
- Fork
- 3.1k
- Merge trung bình
- 1 ngày 15 giờ
- Pull request đã merge (30 ngày)
- 225
Mô tả
### 问题描述 / Problem Description
当前在对话中,当我向 Agent 提出一个任务后,Agent 在执行过程中会逐条发送大量消息卡片。比如一个操作涉及 10 个步骤,Agent 就依次弹出 10 条消息,导致聊天界面刷屏,体验非常差。
Currently, when I ask an Agent to perform a task, it sends multiple fragmented message cards one after another during execution. For example, a 10-step task results in 10 separate messages flooding the chat, which is a very poor user experience.
---
### 具体痛点 / Pain Points
- 一条对话中连续收到 5-10 条 Agent 消息,内容碎片化
5–10 agent messages arriving in rapid succession for a single request, totally fragmented
- 消息通知频繁弹出,干扰正常沟通
Notification spam disrupts normal communication flow
- 无法快速获取任务的整体结果,需要上下翻动才能看完
Hard to grasp the overall result — need to scroll up and down to read everything
- 如果有多个卡片消息,视觉效果混乱,不像一个智能助手
Visually cluttered, doesn't feel like a smart assistant
---
### 期望效果 / Desired Behavior
Agent 执行多步骤任务时,能够将过程消息智能聚合为一个整体回复。例如:
- 过程中不需要逐条汇报"正在进行第 N 步…"
- 全部完成后合并为一条结构清晰的结果消息
- 或者通过流消息 渐进式显示,或者通过编辑功能不断修改新增
- 或者至少显著减少消息条数,不要一题十答
When the Agent performs a multi-step task, it should intelligently aggregate the process messages into one coherent reply. For example:
- No need to report "Executing step N…" for each intermediate step
- Consolidate everything into a single well-structured final message
- Either through progressive display of streaming messages, or through continuous modification and addition through editing functions
- Or at minimum, significantly reduce message cards — don't send 10 replies for one question
---
### 截图 / Screenshots
Hướng dẫn đóng góp
Đánh giá
Issue này chưa được đánh giá.