# Omnishun Agent Skills

This file gives task-level guidance to desktop agents, answer engines, research assistants, and automation agents working with Omnishun.

## Research The Site
- Start with /llms.txt for orientation.
- Use /sitemap-index.xml to find static, blog, and CMS page sitemaps.
- Use /rss.xml for recent published posts.
- Prefer published URLs over snippets or cached copies.

## Analyze Content
- Extract the page purpose, target query/topic, entity coverage, evidence, missing sections, and internal linking opportunities.
- Check whether the page provides information gain beyond generic SEO advice.
- Flag thin claims, missing examples, missing visuals, missing data, and weak actionability.
- Recommend content updates as specific edits, sections, charts, tables, examples, or source requests.

## Create Content
- Follow the brand system at /brand-system.md.
- Write in a clear technical-editorial voice.
- Include scannable structure, semantic headings, definitions, examples, and evidence where appropriate.
- Support future visual, video, graphic, and interactive components when the topic benefits from them.
- Keep HTML semantic and readable without JavaScript.

## Analyze Data And Act
- Treat analytics, search visibility, SERP data, competitive data, and API outputs as inputs to a decision layer.
- Recommend next actions, not just observations.
- When a rule is conditional, state it as: signal, threshold, action, owner, and review cadence.
- Preserve human review for publishing, brand, legal, and high-impact changes.

## Useful URLs
- LLM guide: https://omnishun.com/llms.txt
- Agent guide: https://omnishun.com/agents.md
- Brand system: https://omnishun.com/brand-system.md
- Contact: https://omnishun.com/contact