Daily Conversation Summary Assistant

Problem it solves

Engineers often collect daily context across chats, task discussions, meeting notes, and agent conversations, but by the end of the day the material is still fragmented: which conclusions were confirmed, which To-do items remain open, and which topics were raised but never followed up. Manual review is easy to skip, while searching through scattered history is noisy. This skill turns the end-of-day wrap-up into a repeatable workflow: it reads the day’s conversations and memory content, extracts key points, To-do items, and open topics, then pushes the report to Feishu.

How it works

It supports two trigger modes:

• Automatic trigger: generates the daily summary at 22:00 and pushes it to Feishu, which fits a fixed reporting cadence.
• Manual trigger: can be invoked with natural language for ad-hoc review, catch-up summaries, or adjusted framing.

The core output has three parts:

• Key information summary: extracts 3-5 major topics or conclusions from the day, answering what mattered most.
• To-do item cleanup: lists explicit To-do items separately for follow-up.
• Potential topic identification: flags unresolved topics worth tracking, so discussions are less likely to disappear.

The Feishu push closes the loop by delivering the report without manually copying it into a chat window.

Boundaries

It is best for knowledge management around text-based conversations and memory content, especially when engineers or teams want a stable daily digest. If the workflow depends on another IM, multi-timezone scheduling, deeper archival, or permission controls, verify whether the implementation supports those needs. The automatic time is fixed to 22:00, and manual summaries still rely on enough same-day context; with sparse input, the report will be thinner.