Agents and skills for content production

The problem

Producing content across multiple formats required consistent templates, reliable source material, specialist review, and a way to turn documentation and support insights into content priorities.

My contribution

Mansa built a content production workflow using Claude Code and Cursor, with reusable skills, configured specialist agents, and documentation templates for technical guides, blog posts, social media posts, and internal documentation. He connected planning, source research, drafting, diagram creation, documentation QA, metrics analysis, and human review.

Approach

• Define templates and writing standards for each content format, then map source material into structured drafts without inventing product facts or changing code examples.
• Configure specialist agents for project planning, technical research, documentation, and diagrams, with skills coordinating handoffs from source mapping through draft and review.
• Separate static documentation and blog QA from developer-journey walkthroughs. Check source fidelity and structure, record success, warning, or blocker findings, and flag steps that cannot be verified.
• Build a ReadMe metrics collector and analyst workflow for Median’s documentation: page views, page quality, and search gaps inform prioritized recommendations. ReadMe is the documentation platform in this workflow.
• Create support-to-knowledge synthesis and review-agenda skills, internal updates, and sprint-planning workflows, with redaction rules and human approval gates.

Artifacts

• Specialist agent definitions and reusable workflow skills
• Technical, blog, social, and internal documentation templates and content structures
• Documentation and blog QA helper scripts
• ReadMe metrics collection script and analysis/reporting agent definitions
• Draft and review artifacts, support-to-knowledge agendas, and a team workflow guide

Supported outcomes

• Established reusable content production steps across documentation and marketing formats, with specialist handoffs and explicit human review.
• Created source-fidelity and developer-journey QA processes that surface documentation gaps and verification limits.
• Connected documentation metrics and support synthesis to a structured content-prioritization workflow; measured improvements are not verified.