draft-creator

Create a content draft from a topic idea or content requirements.

0. Load Project Context

Look for ./seo-memory.md (the default seo-memory location).

• File exists and readable — load it. Use this context throughout the draft to keep product references, terminology, and positioning accurate.
• File doesn't exist or inaccessible — ask the user: "I don't see a project memory file (seo-memory.md). Would you like to run /seo-memory first to capture your project context? This helps me weave accurate product references into the blog. If not, I'll skip this step."
  • If yes — run the seo-memory skill, then continue.
  • If no — proceed without project context.

1. Gather Requirements

Ask the user what they want to write about if not provided.

If the input is a vague idea (not full requirements), research reader intent:

1. Try a Google search query likely to find this idea in SERP
2. If SERP is highly related → run serp-gap-analysis skill
3. If not related → try another query
4. After 3 failed attempts → stop and report that this may not be a good content idea (no matching searches)

Skip SERP analysis if the input already contains these requirements:

• SERP intent summary
• Word count
• Content format
• Must-have topics
• Competitor gaps

Required Fields (resolved from input or research)

From SERP analysis:

• SERP intent summary
• Word count
• Content format
• Must-have topics
• Competitor gaps

Important: SERP analysis may return multiple options for some fields (e.g., word count ranges like 1. 1500-2500, 2. 2500-3500). Make sure user confirms them before proceeding to the next step.

Propose these based on the above:

• Target reader
• Primary outcome (what the reader can do/understand after reading)

Setup

Create ./blogs/<topic>/ folder for all writing-related files.

2. Find Information Gain (subagent)

Run the content-researcher skill. This produces:

• ./blogs/<topic>/knowledge-base.md — collected knowledge from SERP pages
• ./blogs/<topic>/under-discussed.md — questions competitors don't answer well

Information gain = what the reader still cannot do, decide, understand, or trust after reading the top results. The under-discussed questions are your information gain.

3. Create Outline (subagent)

Use this structure, adapting to content format and word count:

1. Hook — grab attention, establish relevance
2. Problem — what's at stake, why it matters
3. Why existing advice is incomplete — set up your unique angle
4. Your framework/method — the core value
5. Examples/applications — make it concrete
6. Mistakes/caveats — build trust, add depth
7. Conclusion/next step — actionable takeaway

Adapt Based On

• Word count — expand or compress sections accordingly
• Must-have topics — verify knowledge base has content to cover each
• Under-discussed questions — integrate 2-3 as your information gain

Output Format

## {H2 Title}
{1-2 sentence description of what this section covers}
({word count} words)

Repeat for each section. Total word counts should match the target.