Menu
Use when user says "mcp:pipeline", "kompletter mcp durchlauf", "leads komplett verarbeiten", "qualify research generate", "alles in einem lauf mcp", "full pipeline", or triggers /mcp:pipeline.
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Use when user says "mcp:campaign", "erstelle kampagne via mcp", "kampagne per mcp anlegen", "campaign blueprint erstellen", "bearbeite kampagne via mcp", or triggers /mcp:campaign.
Use when user says "mcp:generate", "generiere emails", "email generation", "MCP workflow", "generate emails for campaign", or triggers /mcp:generate.
Use when user says "mcp:import", "importiere leads via mcp", "leads hochladen mcp", "csv leads importieren", "outscraper import mcp", or triggers /mcp:import.
Use when user says "mcp:qualify", "qualifiziere leads", "lead qualifizierung", "qualify leads", "leads bewerten", or triggers /mcp:qualify.
Use when user says "mcp:research", "recherchiere leads", "lead research", "research leads", "leads recherchieren", or triggers /mcp:research.
Use when user says "mcp:verify", "verify emails", "email verification", "pruefe emails", "email review", "emails pruefen", or triggers /mcp:verify.
| name | mcp-pipeline |
| description | Use when user says "mcp:pipeline", "kompletter mcp durchlauf", "leads komplett verarbeiten", "qualify research generate", "alles in einem lauf mcp", "full pipeline", or triggers /mcp:pipeline. |
Dieser Skill orchestriert die komplette Lead-Verarbeitung einer Kampagne als einen Lauf ueber die Phasen-Skills — die Claude-Subagent-Variante von "Qualifying + Research + Emails in einem Lauf" :
Phase 1: /mcp:qualify — list_leads(qualification_status="pending") -> Subagents -> write_lead_details
Phase 2: /mcp:research — list_leads(research_status="pending", fit_level="qualified") -> Subagents
Phase 3: /mcp:generate — list_leads(campaign_status="processing") -> Subagents -> save_lead_variables
(danach manuell: /mcp:verify — Review & Approve)
| Eingabe | Verhalten |
|---|---|
/mcp:pipeline 80 | Voller Lauf fuer Kampagne 80 (alle 3 Phasen) |
/mcp:pipeline 80 --bis research | Nur Phase 1+2 |
/mcp:pipeline | Kampagne via list_campaigns waehlen |
Vorab abfragen: Batch-Groesse (Default 10) und welche Phasen (Default: alle 3). KI-Variablen (Phase 3) ist optional zuschaltbar/abschaltbar — Review/Approve gehoert bewusst NICHT in die Pipeline (Vier-Augen-Prinzip via /mcp:verify).
mcp-qualify, mcp-research, mcp-generate) — Prompts, Regeln und Fehlerbehandlung von dort uebernehmen, keine abweichende Logik.Nach jeder Phase einen Zwischen-Report:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Phase {n}/{3} ({name}) abgeschlossen — Kampagne {campaign.name}
Verarbeitet: X | Erfolg: Y | Fehler: Z | Nicht qualifiziert: Q (nur Phase 1)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Abschluss-Report nach der letzten Phase: Gesamtzahlen pro Phase + "Naechster Schritt: /mcp:verify — Variablen pruefen und freigeben."
/mcp:qualify, /mcp:research, /mcp:generate · Review: /mcp:verify/mcp:campaign, /mcp:import