clay-leads
Skill0 starsChanged 4 months agoMITView on GitHub
Use when Gera wants outbound leads — "find me leads", "build a lead list", "get me 50 leads for <avatar>", "cold outreach list", "enrich these companies", "who should I email for <offer>", or "/clay-leads". Sources + waterfall-enriches B2B leads via the official Clay plugin (verified emails/phones) and drafts personalized outreach from the business context pack, delivering a campaign-ready CSV. Costs Clay credits — never auto-invoke on a hunch; drafts only, nothing ever sends.
In this repository50 other files it gives its agents.
- aegisSkill
- agent-builderSkill
- agents-team-builderSkill
- ai-applySkill
- auditSkill
- bottleneckSkill
- buildSkill
- capability-radarSkill
- cowatchSkill
- deepforkSkill
- design-blueprintSkill
- evalSkill
- fable-modeSkill
- financesSkill
- forgeSkill
- framework-builderSkill
- grill-meSkill
- hooks-builderSkill
- level-upSkill
- loops-builderSkill
- meeting-hudSkill
- multi-brainSkill
- nate-watchSkill
- onboardSkill
- plugin-builderSkill
- ponytailSkill
- presentSkill
- prompt-evalSkill
- radar-harvestSkill
- rag-forgeSkill
- rlsgateSkill
- rnd-loopSkill
- rndSkill
- roastSkill
- routines-builderSkill
- servy-routineSkill
- session-handoffSkill
- skill-builderSkill
- skill-harvestSkill
- sparkSkill
- storm-researchSkill
- studySkill
- talkSkill
- taste-skillSkill
- verify-loopSkill
- verify-specificsSkill
- wheel-of-lifeSkill
- workanaSkill
- workaround-builderSkill
- workflow-builderSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 52
Never send, schedule, or buy anything (emails, domains, accounts, credit top-ups).
- Distinctive: line 53
Never spend credits without the Step-3 estimate + OK.
- Distinctive: line 54
Never use for SoftServe prospecting or route SoftServe data through Clay (policy §7) — Penguin Alley / personal only.
- Distinctive: line 55
Never fabricate or pad enrichment fields.
- Very specific: line 56
Never connect the hosted clay-for-reps MCP connector alongside the plugin (Clay's own guidance).
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
What this skill does
Runs the outbound half of the funnel (inbound = Penny → /lead-pipeline; marketplace = /workana): Claude Code orchestrates, Clay supplies the data. One run = a scoped ICP → sourced + enriched leads (verified emails, pain-point signals) → personalized subject/body drafts in the business's voice → one CSV Gera imports into Clay to campaign from. Method + verified platform limits: references/clay-leadgen.md. Connection + install: connections.md #28.
Hard law: this skill produces lists and drafts. It never sends anything, never buys anything, and never spends credits without stating the estimate and getting Gera's OK first.
Step 1 — Prereq gate (all three, in order)
- Plugin + auth: run
clay whoami(exit 0 = signed in; exit 3 = not). If theclayCLI is missing or unauthenticated, STOP and hand Gera the install block fromconnections.md#28 (marketplace installs +clay loginare his hands; remind him the post-login Claude Code restart is mandatory). Do not improvise another data source. - Credit balance:
clay credits— if the balance can't cover the estimate in Step 3, say so before anything else (Free plan = 100 Data Credits/mo; a Nate-scale 50-lead enriched run ≈ 172). - Context pack: confirm which identity is prospecting and load its context — Penguin Alley agency:
context/about-business.md+projects/REGISTRY.md; Cleia: its REGISTRY entry + pilot facts; services/workshops: the priced menu. Cold copy without a real offer + proof reads as slop — if the pack is thin for this ICP, tell Gera what's missing instead of padding with generalities.
Step 2 — Scope the run (with Gera)
Pin down, from $ARGUMENTS plus at most one round of questions: N (default 25 on a trial account, not 50), ICP/avatar (industry, size, geography, decision-maker titles), the CTA (e.g. "yes to a 90-second Loom"), and the offer the copy sells. Echo the scope back in one line before spending anything.
Step 3 — Cost gate (blocking)
Look up real prices before running: clay routines get <routine-id> → estimatedCreditCost for each enrichment in the plan, multiply by N, add the search cost. State: "~X credits ≈ $Y at plan rates; balance is Z. Proceed?" Default cap: 200 credits/run — exceeding it needs Gera's explicit raise. On a trial/free workspace also state how much of the monthly allotment this run consumes.
Step 4 — Source + enrich (CLI-first)
- Work through the
clayCLI / plugin skills (search → routines), not raw MCP calls, per CLI > API > MCP. - For N ≥ ~25, fan out subagents or a dynamic workflow by segment (Nate's pattern: one agent per city/vertical, then aggregate + dedupe + yield-check). Verify, don't trust: every lead needs a verified email status; a blank cell beats an invented one — never fabricate enrichment data.
- Remember platform limits (
references/clay-leadgen.md): searches = people/companies only; tables are read-only to agents; Workflows are Alpha. The CSV is built locally, not in a Clay table.
Step 5 — Copy pass
For each lead: personalization hook (from pain points / review signals / recent events, with the hook's source kept in its own column), subject, body. Voice rules are binding: curious-learner register, no hype, no fake familiarity, honest about where the business actually is (Nate's demo copy literally said "I'm still getting this off the ground" — that honesty is the model). If a proven cold-email strategy doc exists in knowledge/, use it as the style guide; don't freestyle structure.
Step 6 — Verify + deliver
Verification pass before delivery: dedupe, no blank required columns (email, subject, body), email_status = verified (or the lead moves to a needs-review sheet), copy actually references the lead's own data. Deliver to money/leads/<date>-<slug>/leads.csv + run-summary.md (scope, credits estimated vs actually spent, yield rate, anything skipped). Report the spend honestly even when it overshot.
Step 7 — Handoff checklist (Gera's hands, in Clay's UI)
Give Gera this list, done for him up to the send boundary: import CSV → new table → Campaign → add lead list → message from variables (email subject / email body rows) → preview → sender accounts (the in-UI "buy pre-warmed accounts" flow from the video is undocumented on clay.com — verify in-app; ~30 sends/day/account while warming) → 3-day follow-up slots (Servy can draft these). Launching the campaign is Gera's click, always — and agency outreach additionally respects the pa-outreach case-study gate.
Step 8 — Bookkeeping
Append the run to references/log.md (## [date] run | clay-leads — <slug>): N, credits, yield, what to fix next run. First real run: capture what worked/broke and update this skill same-session.
Never
- Never send, schedule, or buy anything (emails, domains, accounts, credit top-ups).
- Never spend credits without the Step-3 estimate + OK; never blow the 200-credit cap silently.
- Never use for SoftServe prospecting or route SoftServe data through Clay (policy §7) — Penguin Alley / personal only.
- Never fabricate or pad enrichment fields; never ship copy Gera hasn't seen for an external send.
- Never connect the hosted
clay-for-repsMCP connector alongside the plugin (Clay's own guidance).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

