catchup

Re-orient after a break — aggregates live work context from all sources and session memory into a concise briefing.

安装方式

默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

直接命令不会经过审查 Prompt；运行前请先检查来源。

npx skills add https://github.com/prassanna-ravishankar/reorient --skill catchup

命令会保持在同一行。复制前请横向滚动并检查完整内容。

想先保存到本地？可下载 SkillsMP 当前能够提供的文件。

决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。

SKILL.md

来源说明 · 只读预览

name	catchup
description	Re-orient after a break — aggregates live work context from all sources and session memory into a concise briefing.

Run the data aggregator (Linear, GitHub, Drive, Slack) from the project root:
```
uv run python scripts/catchup.py
```
This includes Slack mentions and unread DMs via the SDK. If Slack data is missing from the output (e.g. token expired), fall back to the slack-reader browser agent.
Query Granola MCP for recent meeting notes, decisions, and action items. If unavailable, skip.
Search session memory for recent context (decisions, open threads, things flagged for follow-up).
Synthesize all sources into a briefing with three sections:
- In progress — active work items across all sources
- Needs attention — unread mentions, review requests, DMs awaiting reply, meeting action items
- Since you were away — notable changes, merged PRs, closed issues, new threads, meetings held

Keep the briefing tight. Lead with the most time-sensitive items. Skip sources that returned no signal.