reworkd/AgentGPT

🌎 Manually source links in

Open

#773 建立於 2023年6月12日

在 GitHub 查看
 (0 留言) (0 反應) (0 負責人)TypeScript (34,594 star) (9,446 fork)batch import
help wanted

描述

When AgentGPT sources web links, we are manually writing a source via markdown. This is expensive in terms of tokens. We could instead just write the source, and then provide the links afterwards without using completions

Image

貢獻者指南

技術棧
typescript
領域
frontend
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-3 hours
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
stale
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
clear
前置要求
ReactTypeScriptMarkdown rendering
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
70
研究方向
Locate the component(s) responsible for rendering agent output, particularly where markdown sources are displayed. This is likely in the chat interface or message display area. The goal is to modify the rendering so that source links are listed after the main content instead of being inline, thus saving tokens. Look for files like `ChatMessage.tsx` or similar, and review how the markdown is currently generated to identify where to separate the sources.