Ralph Wiggum Copywriter
ActiveClaude Code plugin: autonomous multi-agent copywriting loop. Learns your voice, self-critiques, and rewrites until acceptance criteria pass.
Overview
Ralph Wiggum Copywriter is a Claude Code plugin that runs an autonomous multi-agent copywriting loop. It learns your voice, critiques its own work, and rewrites until acceptance criteria pass.

Built on the Ralph Wiggum pattern: each iteration runs in a fresh context window. Memory persists only through git commits and text files. No state leaks, no context drift.
Architecture
The system coordinates four specialized agents feeding into a shared SQLite database:
- TrendScout Agent — scans for emerging topics and scores relevance
- Research Agent — gathers data points, key findings, and summaries
- Product/Marketing Agent — captures key messages and target audience context
- Ralph (Copywriter) — reads all inputs, plans content, writes drafts, self-critiques, and iterates until acceptance criteria pass
The self-critique loop is the key part. Instead of generating and stopping, Ralph evaluates its own output against acceptance criteria, logs learnings to progress.txt, and iterates. Patterns compound across iterations.
The Ralph Loop
An 8-step cycle that repeats until all tasks are complete:
- Read task definitions from
prd.json - Check
progress.txtfor accumulated learnings - Pick the highest-priority incomplete task
- Execute following explicit acceptance criteria
- Verify quality through automated tests
- Commit progress to git
- Log learnings and mark task done
- Repeat
Each iteration is a fresh context window. Memory persists through files, not state. Small stories with explicit criteria keep scope tight—every task must complete in a single iteration.
How to Use
Install as a Claude Code plugin, then four commands:
/ralph-init— initialize a content project in the current directory/ralph-marketer— start the autonomous loop/ralph-status— check pipeline progress/ralph-cancel— stop the loop
What It Ships
The default pipeline covers the full SaaS content lifecycle: product launch blogs, thought leadership pieces, data-driven articles, case studies, social media posts, newsletters, and metrics reporting—12 stories out of the box, fully customizable through prd.json and seed scripts.
Repository
Gallery