Back to the catalog

brandketing

Marketing intelligence plugin that builds a living context store from product docs, customer calls, CRM data, and competitor research - then

Open source Repository Open in the app JSON README (API)

About

Marketing intelligence plugin that builds a living context store from product docs, customer calls, CRM data, and competitor research - then compiles structured briefs and generates copy through a 6-step strategy workflow with a built-in critic pass. Separates evidence collection, human approval, brief compilation, and copy generation into distinct steps so the model never does extraction, positioning, and writing in one jump.

Details

Kind
Plugins
Topic
Marketing & analytics
Publisher
entro314-labs
Origin
marketplace
Category
ferramentas
Last push
2026-04-16T06:04:49Z
Repository state
ativo
License
MIT
Added
2026-08-30 01:48:58
Updated
2026-08-30 01:48:58
Origin id
entro314-labs/brandketing/brandketing

README

# brandketing

A Claude Code plugin that turns raw product and customer data into a living marketing intelligence context store - and compiles it into briefs and copy on demand.

## What it does

Instead of filling out a brief from scratch every time, this plugin maintains a **context store** that accumulates structured marketing objects over time. Each asset you generate pulls from approved, sourced, scored objects - not from memory or imagination.

The two source documents behind this system:
- `brief-harness.md` - automated brief compilation from 5 data pipelines
- `marketing-harness.md` - 6-step strategy workflow + critic pass for copy generation

## Install

```
/plugin install brandketing@claude-plugins-official
```

Or from GitHub:
```
/plugin install brandketing@github:entro314-labs/brandketing
```

## Commands

| Command | What it does |
|---|---|
| `/brandketing:ingest <source> <pipeline>` | Normalize a raw source file into structured context objects |
| `/brandketing:review [type]` | Approve, reject, or edit pending objects |
| `/brandketing:brief <deliverable> [icp]` | Compile a brief from approved objects |
| `/brandketing:generate [deliverable] [icp]` | Run strategist + critic, output full copy hierarchy |
| `/brandketing:status` | Show context store freshness and object counts |

## Pipelines

| Pipeline | Source types | Output objects |
|---|---|---|
| `product-truth` | Docs, PRDs, changelogs, demos | Capability |
| `customer-truth` | Call transcripts, tickets, reviews | Pain, Quote, UseCase |
| `commercial-truth` | CRM, won/lost, account data | Persona |
| `market-truth` | Competitor pages, comparison content | Competitor, Cliche |
| `proof-truth` | Analytics, case studies, wins | Proof |

## Agents

- **normalizer** - converts raw sources into typed, scored objects
- **compiler** - assembles brief sections from approved objects, filtered by ICP
- **strategist** - runs the 6-step strategy workflow to produce a messaging hierarchy
- **critic** - red-teams every output line and rewrites what's weak

## Context store

Each project gets its own store at `.brandketing/context/store/` (not committed to git).

Objects: `capabilities`, `pains`, `quotes`, `use_cases`, `personas`, `competitors`, `cliches`, `proofs`, `brand`

Each object carries: confidence score, source references, freshness date, and approval status.

## Workflow

```
1. /brandketing:ingest call-transcripts.txt customer-truth
2. /brandketing:ingest product-docs.md product-truth
3. /brandketing:review
4. /brandketing:generate homepage-hero icp-1
```

For a lightweight version without the context store, see **brandketing-quick**.

## Design principles

- **Machine assembles, human approves.** The system proposes; a human decides ICP, category, and positioning bet.
- **Every claim has a source.** Objects without source references are not created.
- **Freshness is tracked.** Static / slow-moving / fast-moving layers refresh on different cadences.
- **Copy goes through a critic.** Every generation runs a red-team pass before output.

More