A free Claude Code skill that scores and ranks any list of target accounts by hiring, funding, tech stack and buying-intent signals, so you know exactly who to contact first.
Signal Scoring is a Claude Code skill: a markdown playbook that gives your agent a weighted rubric for ranking target accounts. It scores four signal categories (hiring, funding, tech stack changes, buying intent) out of 100 points, then sorts accounts into four tiers: Hot, Warm, Cool, or Discard, each with its own recommended action and response window.
The video's claim is that it "scores your leads by real buying signals, so you know exactly who to call first." That held up in testing. Given 8 invented target accounts with deliberately different signal profiles, one maxed out on every category and one with zero signals at all, the skill produced 8 distinct scores in the correct order with no ties, and correctly identified both extremes. One honest caveat: this skill does not go out and find hiring posts, funding rounds, tech stack changes or website visits on its own. You (or another connected tool) still have to supply those facts. This is a scoring and ranking rubric applied to information you already have, not a live data feed.
Two more things worth knowing before you rely on it. First, its own output checker only confirms a handful of keywords appear in your write-up and that it isn't too short, it does not check whether your math or your tiers are correct. Second, the default weights genuinely under-rank accounts that only show one very strong signal (in testing, an account with a hot, days-old intent signal but nothing else still scored below the "Discard" floor). The skill tells you to recalibrate the weights against your own closed-won data before trusting the tiers blindly, and testing backed that advice up.
Anyone sitting on a list of target accounts with scattered signals, a few hiring posts here, a funding announcement there, some pricing page visits somewhere else, who has no consistent way to decide who gets called today versus who waits. This replaces the gut-feel call of "which lead looks hottest" with an actual weighted score and a tier, so the same account gets the same priority no matter who on the team is looking at it.
This copies the whole gtm-skills library onto your machine. You only need one folder out of it.
git clone https://github.com/LeadMagic/gtm-skills.gitThis drops just the signal-scoring skill into the folder Claude Code checks for project skills.
mkdir -p .claude/skills
cp -r gtm-skills/skills/prospecting/signal-scoring .claude/skills/signal-scoringNote: the repo also documents a one-line ./install.sh and a gh skill install command. Both rely on a gh skill extension that was not present on the test machine, and the one-line installer failed outright instead of falling back to a plain copy. The cp -r above is the path that actually worked in testing, and it is also the exact fallback method the repo's own installer uses internally.
Sample input used in testing: 8 invented target accounts, each with a different mix of hiring, funding, tech stack and intent signals (none real companies or people). Abbreviated real output, ranked highest to lowest:
1. Ferrous Metals Group 98 Tier A - Hot
3 new AE roles + new VP Sales this week, Series C $60M
last week, added Gong + Outreach this month, demo
request + 6x pricing visits this week
2. Northwind Robotics 84 Tier A - Hot
3. Meridian Health Partners 70 Tier B - Warm
4. Vantage Analytics 41 Tier C - Cool
5. Harbor & Stone Legal 30 Discard
6. Cascade Freight Co 28 Discard (see caveat below)
7. Ridgeline Manufacturing 14 Discard
8. Solace Health 0 Discard (no signals, control case)All 8 scores landed in the correct order on the first pass, including a deliberate zero-signal account (Solace Health) ranking exactly last.
./install.sh or gh skill install fails with unknown command "skill" for "gh". Your GitHub CLI does not have the gh skill extension the repo assumes. Skip it and use the plain cp -r copy in Step 2 instead, it lands in the same place.
scripts/cc-gtm.py (the "curated installer") crashes with a SyntaxError about an f-string and a backslash. That script needs Python 3.12 or newer. If you're on an older Python 3, use the cp -r copy method instead.
scripts/check-output.py says FAIL - missing terms on a deliverable that looks correct. It is checking for a handful of literal keywords (like "prospecting" and "signal") somewhere in your write-up, not checking your scoring. Add a sentence that naturally uses the missing word and it will pass, this is a minimum-completeness check, not proof the scores are right.
An account with one very strong signal keeps landing in Discard. That is the default weights doing what they are built to do, they need signals spread across at least two categories to clear the Cool tier. If a single-category signal (usually intent, like repeated pricing page visits) genuinely predicts a sale in your business, raise that category's weight yourself, the skill expects you to calibrate against your own closed-won data rather than trust the out-of-the-box numbers forever.
Built by LeadMagic (https://github.com/LeadMagic/gtm-skills), MIT license, 53 stars at test time. Correction to how the video describes it: it scores and ranks leads exactly as claimed, but it works on signal facts you already have, not signals it goes out and finds itself, and its own output checker only confirms a few keywords are present, it does not verify the scoring is correct.
This is one agent. The 7-Day AI Native Challenge is how you learn to build the next ten yourself, live, with Dany.
Join the Challenge →Too busy to build it yourself? Focentra AI will build, install and train your team on it. Priced on the call.
Talk to Focentra AI