Content Research Writer

---
name: content-research-writer
description: "Content research and SEO writing methodology. Guides the agent through topic research, keyword identification, competitive analysis, and writing SEO-optimized content that ranks well and provides genuine value to readers."
category: ai-agents-and-harness
source_repo: google/adk-samples
source_path: "python/agents/agent-skills-tutorial/app/skills/content-research-writer/SKILL.md"
source_url: https://github.com/google/adk-samples/blob/HEAD/python/agents/agent-skills-tutorial/app/skills/content-research-writer/SKILL.md
---


# Content Research & Writer Instructions

When asked to research a topic and write content, follow this methodology:

## Phase 1: Topic Research
1. Identify the **primary topic** and break it into subtopics
2. List 3-5 **key questions** the target audience would ask
3. Identify the **target keyword** and 3-5 related keywords
4. Note what makes this content **unique** — what angle hasn't been covered?

## Phase 2: Content Planning
1. Define the **content type**: tutorial, deep-dive, comparison, or opinion
2. Set the **target word count**: 1500-2500 words for tutorials, 800-1500 for opinion pieces
3. Plan **code examples** if technical (at least 2-3 per tutorial)
4. Identify **internal links** to related content and **external links** to authoritative sources

## Phase 3: SEO Optimization
Read `references/seo-guidelines.md` for detailed SEO rules using `load_skill_resource`.

Apply these during writing:
- Primary keyword in title, first paragraph, and 2-3 subheadings
- Natural keyword density (1-2%, never forced)
- Meta description draft (150-160 chars)
- URL slug suggestion (short, keyword-rich)

## Phase 4: Writing
1. Write the **hook** first — make the reader care in 2 sentences
2. Deliver on the **promise** of the title in every section
3. Use the **inverted pyramid**: most important info first in each section
4. End with **actionable takeaways** the reader can apply immediately

## Phase 5: Quality Check
- Does every section add unique value?
- Are all claims supported by evidence or code?
- Would you share this with a colleague? If not, what's missing?

---

**Source:** [`google/adk-samples`](https://github.com/google/adk-samples) → `python/agents/agent-skills-tutorial/app/skills/content-research-writer/SKILL.md`