The first enrichment pass, and it is free. It reads a company's website alongside what your graph already knows, ranks what it finds across six signal classes, records a structured signal block onto the account, and saves a brief carrying the named variables your outreach later injects.
claude
One job, four steps. Website plus what your graph already knows, turned into the features your ICP model scores on.
One account, one domain, or a whole lead list in Nous. It scans the website with plain fetches, so this pass costs nothing.
claude
Stack, hiring, momentum, friction, intent and domain. Every finding belongs to exactly one class, and each one is scored one to ten.
claude
The concise signal block is written onto the account in Nous. Signals are the features your ICP model scores on, so this is what makes the score mean something.
A fuller brief lands in Notes, each signal carrying the named variables that outreach injects later. Raise size, founder handle, the post that proves it.
Every skill is a plain markdown file in your own folder. Read it, change it, and it is yours.
The six classes are scored against your GTM profile. What counts as momentum for you is not what counts for someone selling something else, and the file says so.
Change the class weights, add a class of your own, or cap the scan at the top hundred accounts in a list rather than running the whole thing.
This is the free first pass, so run it wide. Then run content-scan only on the accounts it qualifies, because that one costs money per profile.
A whole lead list scanned, ranked and recorded. No API keys beyond Nous, no cost.
Paste it into your terminal. Then use /signal-scan inside Claude Code.
curl -sL https://raw.githubusercontent.com/NousC/gtm-skills/main/.claude/skills/signal-scan/SKILL.md --create-dirs -o ~/.claude/skills/signal-scan/SKILL.md
Open Claude Code and type /signal-scan. Name an account or a lead list. It scans, ranks, and shows you the readout before it writes anything to the record.