agentfootprint

介紹

上下文工程，抽象化。構建 AI 代理，使每一次 LLM 呼叫皆可追溯至注入的內容、觸發者、時間以及快取方式。基於 footprintjs 建置。

AI 翻譯

取自儲存庫 README

README

正在載入完整 README。 以下為已儲存的摘錄。

Your agent gave an answer that looks right — and it's wrong. The logs can't tell you who influenced it. Agentfootprint can. The explainable AI agent framework for TypeScript: every read, write, decision, and tool call becomes connected evidence as your agent runs. When something goes wrong, you don't grep logs — you ask. Build agents — skills, steering, RAG, memory, control flow — and debug them like nothing else. Why is a query, not a guess. A real run, replayed — rendered with Story Lens ( npm i agentthinkingui ). Every frame is generated from the run's own trace; ▶ watch it live .