home/writing/calmfalcon-pipeline-prompts-v37

Calmfalcon Pipeline Prompts V37

GPTClaudeGemini··250 copies·updated 2026-07-14
calmfalcon-pipeline-prompts-v37.prompt
# CalmFalcon Pipeline Prompts — v37

> **Live document** — extracted from `calmfalcon-signals-pipeline-multi-agent.json`
> **Pipeline version:** v20 | **Prompt version:** pv-actwatch-v37 | **Date:** April 6, 2026
>
> **Architecture:** Split theme generation — 4 LLM agents in sequence:
> 1. **Theme Selector** — picks 8 clusters from signal data
> 2. **Theme Writer** (x8 parallel) — writes 1 theme per cluster
> 3. **Verifier Agent** — 12-rule QC pass on all themes
> 4. **Brief Writer** — generates weekly brief from scored themes
>
> All agents use GPT-4 Turbo | Temperature 0.3 | JSON mode where applicable

**Structure applied:** `<Role> → <Instruction> → <Task> → <Guardrails> → <Example>`

---

## Agent 1: Theme Selector

**n8n node:** `Theme Selector`
**Purpose:** Reads pre-computed signal clusters, selects exactly 8 for theme generation. Enforces max 3 ACT, min 5 WATCH, enterprise AI relevance gate, vendor dedup.
**Model:** GPT-4 Turbo | Temperature: 0.3

---

### SYSTEM PROMPT

when to use it

Community prompt sourced from the open-source GitHub repo kamalbhatiapm/cf_project_claude_code_n8n_supabase (no explicit license). A "Calmfalcon Pipeline Prompts V37" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.

tags

writingcommunitygeneral

source

kamalbhatiapm/cf_project_claude_code_n8n_supabase · no explicit license