Install any skill in seconds. Free to start, no credit card required.
Get Started Free →You are an expert in Clay, the data enrichment and outbound sales platform that pulls from 75+ data providers to build rich prospect profiles. You help teams automate lead discovery, enrich contacts with firmographic and technographic data, score leads against ICP criteria, and trigger personalized outreach sequences — replacing manual research with automated, data-driven prospecting.
.claude/skills/terminalskills-clay/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 14 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 150% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 47% | 0% |
| case-10 | ✗→✓ | ▲ Improved | 99% | 0% |
| case-14 | ✗→✓ | ▲ Improved | 224% | 0% |
| case-16 | ✗→✓ | ▲ Improved | 136% | 0% |
You are an expert in Clay, the data enrichment and outbound sales platform that pulls from 75+ data providers to build rich prospect profiles. You help teams automate lead discovery, enrich contacts with firmographic and technographic data, score leads against ICP criteria, and trigger personalized outreach sequences — replacing manual research with automated, data-driven prospecting.
markdown## Clay Table Structure A Clay table is a spreadsheet with superpowers. Each row is a lead. Columns can be: - **Imported** — CSV upload, CRM sync, or webhook - **Enriched** — Auto-populated from 75+ data sources - **AI-generated** — GPT/Claude processes enriched data into insights ## Example: ICP-Matched Lead Table | Column | Source | Description | |-----------------|-----------------|------------------------------------------| | Company Name | Import | From LinkedIn export or CRM | | Domain | Enrichment | Company website from Clearbit | | Employee Count | Enrichment | From LinkedIn/Crunchbase | | Funding Stage | Enrichment | Series A/B/C from Crunchbase | | Tech Stack | Enrichment | BuiltWith/Wappalyzer detection | | API Endpoints | Enrichment | Custom: count public API docs pages | | Decision Maker | Enrichment | VP Eng/CTO from LinkedIn + Apollo | | Email | Enrichment | Verified email from Hunter/Apollo | | Recent News | Enrichment | Latest press from Google News | | LinkedIn Post | Enrichment | Latest post from contact's LinkedIn | | ICP Score | AI Formula | 0-100 score based on all enrichment data | | Personalized Opener | AI Formula | GPT-generated first line for cold email |
typescript// Fetch enriched leads from Clay table const response = await fetch(`https://api.clay.com/v3/tables/${tableId}/rows`, { headers: { "Authorization": `Bearer ${process.env.CLAY_API_KEY}`, "Content-Type": "application/json", }, }); const { data: rows } = await response.json(); // Filter to qualified leads const qualified = rows.filter(row => ( row.icpScore >= 70 && row.emailVerified === true && row.employeeCount >= 20 )); // Add a new lead via API (triggers enrichment automatically) await fetch(`https://api.clay.com/v3/tables/${tableId}/rows`, { method: "POST", headers: { "Authorization": `Bearer ${process.env.CLAY_API_KEY}`, "Content-Type": "application/json", }, body: JSON.stringify({ rows: [ { "Company": "Acme Corp", "Domain": "acme.com" }, { "Company": "Beta Inc", "Domain": "beta.io" }, ], }), }); // Clay automatically enriches all configured columns
markdown## Available Data Sources (75+) **Company Data:** - Clearbit, Crunchbase, LinkedIn (company), BuiltWith, Wappalyzer - Google Maps, Glassdoor, G2, TrustRadius **Contact Data:** - Apollo, Hunter, Lusha, RocketReach, Snov.io - LinkedIn (person), Twitter/X profile **Technographic:** - BuiltWith, Wappalyzer, SimilarTech - Custom HTTP header checks, robots.txt analysis **Intent & News:** - Google News, Crunchbase funding alerts - Job posting analysis (hiring signals) - Website change detection **AI Enrichment:** - GPT/Claude for custom analysis - "Summarize this company's value proposition in one sentence" - "Score this lead 0-100 against our ICP: B2B SaaS, 20-100 employees, uses Node.js"
typescript// Clay fires webhooks when rows change or meet conditions // Example: trigger outbound sequence when ICP score reaches 80+ // Webhook payload from Clay interface ClayWebhook { table_id: string; row_id: string; trigger: string; // "row_updated" | "filter_matched" data: { company: string; contactEmail: string; contactName: string; icpScore: number; personalizedOpener: string; }; } // Handle webhook — start email sequence app.post("/api/clay-webhook", async (req, res) => { const payload: ClayWebhook = req.body; if (payload.data.icpScore >= 80) { await startEmailSequence({ to: payload.data.contactEmail, name: payload.data.contactName, company: payload.data.company, opener: payload.data.personalizedOpener, }); } res.json({ ok: true }); });
markdown## Setup 1. Create account at https://clay.com 2. Create a table with your ICP columns 3. Configure enrichment providers (drag-and-drop) 4. Set up AI formulas for scoring and personalization 5. API key: Settings → API → Generate key ## Pricing - Free: 100 enrichments/month - Starter: $149/month (2,000 enrichments) - Explorer: $349/month (10,000 enrichments) - Pro: $800/month (50,000 enrichments)
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 35,502 | 10,428 | -71% | 1 | 1 | 0% | 1,514 | 3,790 | +150% | 0 | 0 | — |
case-02 | pass→pass | 18,848 | 15,958 | -15% | 1 | 1 | 0% | 2,363 | 4,065 | +72% | 0 | 0 | — |
case-03 | pass→pass | 8,163 | 5,856 | -28% | 1 | 1 | 0% | 1,239 | 2,464 | +99% | 0 | 0 | — |
case-04 | fail→pass | 8,810 | 5,556 | -37% | 1 | 1 | 0% | 1,668 | 2,445 | +47% | 0 | 0 | — |
case-05 | fail→fail | 15,011 | 14,659 | -2% | 1 | 1 | 0% | 2,458 | 4,022 | +64% | 0 | 0 | — |
case-06 | pass→pass | 9,975 | 8,254 | -17% | 1 | 1 | 0% | 1,621 | 2,858 | +76% | 0 | 0 | — |
case-07 | pass→pass | 15,135 | 10,478 | -31% | 1 | 1 | 0% | 2,563 | 3,302 | +29% | 0 | 0 | — |
case-08 | pass→pass | 13,936 | 9,523 | -32% | 1 | 1 | 0% | 2,355 | 3,155 | +34% | 0 | 0 | — |
case-09 | pass→pass | 11,906 | 10,440 | -12% | 1 | 1 | 0% | 1,951 | 3,172 | +63% | 0 | 0 | — |
case-10 | fail→pass | 11,989 | 14,324 | +19% | 1 | 1 | 0% | 1,990 | 3,956 | +99% | 0 | 0 | — |
case-11 | pass→pass | 12,828 | 12,680 | -1% | 1 | 1 | 0% | 1,888 | 3,628 | +92% | 0 | 0 | — |
case-12 | pass→pass | 15,988 | 15,917 | -0% | 1 | 1 | 0% | 2,639 | 4,131 | +57% | 0 | 0 | — |
case-13 | pass→pass | 3,860 | 1,561 | -60% | 1 | 1 | 0% | 511 | 1,686 | +230% | 0 | 0 | — |
case-14 | fail→pass | 3,359 | 1,833 | -45% | 1 | 1 | 0% | 553 | 1,790 | +224% | 0 | 0 | — |
case-15 | pass→pass | 4,501 | 1,650 | -63% | 1 | 1 | 0% | 739 | 1,762 | +138% | 0 | 0 | — |
case-16 | fail→pass | 6,478 | 6,131 | -5% | 1 | 1 | 0% | 1,047 | 2,474 | +136% | 0 | 0 | — |
case-17 | fail→fail | 13,805 | 15,281 | +11% | 1 | 1 | 0% | 2,336 | 4,409 | +89% | 0 | 0 | — |
case-18 | pass→pass | 6,679 | 3,653 | -45% | 1 | 1 | 0% | 1,107 | 2,140 | +93% | 0 | 0 | — |
case-19 | pass→pass | 14,537 | 16,818 | +16% | 1 | 1 | 0% | 2,952 | 4,676 | +58% | 0 | 0 | — |
case-20 | pass→pass | 12,772 | 13,046 | +2% | 1 | 1 | 0% | 2,584 | 3,974 | +54% | 0 | 0 | — |
case-21 | pass→pass | 6,560 | 5,327 | -19% | 1 | 1 | 0% | 1,130 | 2,443 | +116% | 0 | 0 | — |
case-22 | pass→pass | 17,205 | 12,822 | -25% | 1 | 1 | 0% | 3,194 | 3,751 | +17% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 21 counted toward the lift figure. The other 1 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +23 percentage points is the difference between those two pass rates over the 21 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.