Weekly Content Marketing Engine

What this agent does

Keeps a website's content pipeline full automatically. This agent runs a three-stage pipeline that turns a website URL into a full SEO blog post and a set of social drafts.

1. stage-1-gap-scan — crawls a site's most recent/prominent pages (via sitemap or homepage/blog navigation), then uses web search/research to find content gaps and trending questions in the site's niche, heuristically scored for SEO potential.
2. stage-2-blog-draft — turns the top-ranked topic into a full, publication-ready SEO blog post (title, meta description, H2/H3 body) matched to the site's own voice, backed by real research (no fabricated stats).
3. stage-3-social-repurpose — repurposes the finished blog post into a LinkedIn post, an X/Twitter thread, and a carousel slide outline, all as draft files.

Requirements

• Runs entirely on free, built-in web tools (web_search, web_research, x_search, read_webpage, browser tools) — no paid crawling, SEO, or content-generation API required.
• Works on Zo Computer natively and on Claude AI (claude.ai, Claude Code, or the Anthropic API).
• Never publishes anything on its own — every output is a draft file for manual review.

Setup

1. Unzip the package into your workspace, keeping the folder structure intact.
2. Run a starter prompt with your website URL to see the full pipeline end to end.
3. (Optional) Pass a focus_topic to skip auto topic-selection and write about a specific angle instead.

What you get

• Content gaps and trending questions in your niche discovered automatically from your own site and the wider web
• A full, publication-ready SEO blog post matched to your site's voice, backed by real research with no fabricated stats
• A LinkedIn post, an X/Twitter thread, and a carousel slide outline repurposed from the same post
• Every output as a review-ready draft file — nothing is ever published automatically