linkedin-start

name	linkedin-start
description	The front door to everything Founder OS can do with your LinkedIn data. Trigger on "help me with my LinkedIn", "what can I do with my LinkedIn", "give me my LinkedIn data", "do something with my LinkedIn export", "I downloaded my LinkedIn data", or any open-ended LinkedIn request where the user has not yet picked a specific job. You bring one file (the LinkedIn data export); the OS shows the baseline of your network as it stands today and routes you to the outcome you want: more leads, a better job, a louder brand, or a healthier network. One honest disclaimer, no full onboarding, and never a claim the data cannot support. Routes to linkedin-network-scan, linkedin-power-audit, linkedin-brand-direction, or linkedin-warm-revival depending on the outcome chosen.
why	A new user does not know which of six LinkedIn skills they need - they know the outcome they want (clients, a job, a name, a healthier network). This is the one entry that turns one file plus one choice into the right path, so nobody has to learn the skill list first.
enhance	Have your LinkedIn export downloaded before you start (Settings -> Data privacy -> Get a copy of your data -> Download larger data archive). With the file in hand the OS can show your real baseline in the same session instead of waiting on the email.
summary	Pick your LinkedIn outcome; the OS aims your own data at it.
allowed-tools	["Read","Bash"]
mcp_requirements	[]

LinkedIn Start

Runs on: local-exec - the happy path runs one local Python entrypoint against the user's export and writes the completed private bundle outside the repo. On a read-only or cloud surface, explain the same route without claiming the run happened.

This is the front door. You have a LinkedIn data export (or can request one). Underneath that one file sit several outcomes. This skill shows you the honest baseline of your network and aims the same data at the outcome you actually want, without making you learn which skill does what first.

The principle: do more for you than you asked, in priority order, and never invent a fact your data does not hold. If the export cannot prove something (seniority, buying power, headcount, where someone is based), the OS says so rather than guessing.

The four outcomes

You want	What it means	Where it routes
Leads	more or better revenue relationships	linkedin-network-scan with the sales ICP (icp.example.yaml)
A better job	who in your network can refer or hire you	linkedin-network-scan with the career ICP (icp.career.example.yaml)
A louder brand	what your network already rewards, and the content lane to own	linkedin-power-audit then linkedin-brand-direction
A healthier network	over and under-represented roles versus your goal	linkedin-power-audit (the network-gap read)

Not sure which one? Pick "show me all lightly" and the OS runs the cheapest read first (the scan) and names what each other outcome would add, so you choose with your own numbers in front of you.

There is a fifth path the data also supports: reviving dormant-but-valuable contacts (linkedin-warm-revival). It needs the power audit's audit.json first, so if a user asks for it directly, run linkedin-power-audit before routing there. Never route to warm-revival without that prerequisite.

Read the state before routing

The route is only as good as what it knows. Before choosing, read brain/.snapshot.md. If it is missing, or its date: line is more than 3 days old, run python scripts/brain-snapshot.py --write first and read the fresh one. The open-flags block tells you what is stuck, the must-do block tells you what this week is actually about, and the patterns block tells you what keeps happening - a front door that routes without them is a cold start at the exact moment the user is deciding whether this OS knows them. If Python is unavailable, route from the files you can read and say so.