wtf is pre-meeting-intel?

Pre-meeting-intel is a Claude or Mira skill that puts together a research brief about a person before you meet them for the first time, built specifically for business development style meetings. Instead of relying on a static LinkedIn summary, it looks at what the person has actually been doing and saying recently, on the idea that knowing what someone shipped yesterday or what people are saying about them this week matters more than an outdated career overview. You run it with a single slash command, giving it a name, a company, and the context for your meeting, and it launches four research agents at the same time. One looks at community discussion about the person on places like Reddit, Hacker News, or social media replies, another looks at the person's own recent posts, talks, or writing, a third looks at coverage from outside media outlets, and a fourth gathers official self descriptions like a bio or company page. It then compares what the person says about themselves against how others describe them, since the authors consider that gap to be where the real insight is. The final output is a brief of around 2000 to 2500 words split into six sections: an identity summary, a short biographical narrative, a timeline of the last 90 days with direct quotes and confidence levels, a side by side comparison of self versus other descriptions, a deeper personality portrait, and a set of talking points meant to be spoken aloud in the actual meeting. There is also a data confidence section showing the underlying facts and how they were weighted. Two optional pieces of the pipeline rely on internal services from ByteDance, one for pulling verified social media post links instead of guessing them, and one for writing the finished brief into a Feishu or Lark document. Both are built to fail gracefully: if they are not available, which will be the case for anyone outside that company, the skill still produces the complete brief, just using a homepage link instead of a verified post link, and returning the brief as plain markdown text instead of a hosted document. To install it for Claude Code, you clone the repository into your Claude skills folder. It works out of the box using web search, web fetch, and Python, with no extra configuration required unless you have access to the internal enhancers.