Skip to main content
skill CX Codex

blog-draft · luongnv89 — Claude Code skill: what it does, how to install

Guides article creation from reference notes and URLs to versioned drafts with mandatory source citations and git commits.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add luongnv89/claude-howto --skill blog-draft -a claude-code
Open on GitHub

What blog-draft does

The skill organizes article drafting into a sequential workflow. It sets up a dedicated directory under blog-posts/YYYY-MM-DD-slug/ and stores summaries of links, search hits, and local files in resources/. The agent then clarifies target angles, audience, and word count: short (500–800 words), medium (1000–1500 words), or long form over 2000 words.

Once you approve the structure in OUTLINE.md, the agent writes draft-v0.1.md. Every factual claim or data point requires an inline citation linked to the references list. When running in a git repository, it commits each major step with a docs: message and pushes changes to the remote branch.

Because the skill halts for user approval before moving between steps, it never drafts blindly in the background. Edits get saved into numbered iterations such as draft-v0.2.md until finalized as final.md. Updating an existing article bypasses research altogether and jumps directly to revision.

What it helps with

  • compile source summaries from URLs and local files into structured notes
  • generate and validate a post structure in OUTLINE.md
  • draft content with inline citations for data and statistics
  • track text revisions across draft-v0.1.md and subsequent version files
  • commit drafts and outlines directly to a git repository

Example prompts for the agent

  • Draft a technical post about Go performance using these benchmark links
  • Create a 1200-word post outline about database indexing based on local notes
  • Update the draft in blog-posts/2025-05-10-caching/ using the latest review comments

How to install blog-draft

1Skills CLI (Vercel)

npx skills add luongnv89/claude-howto --skill blog-draft -a claude-code

2Manually: put the files in place

~/.claude/skills/blog-draft/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/blog-draft/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about blog-draft

How do I install blog-draft in Claude Code?

Run npx skills add luongnv89/claude-howto --skill blog-draft -a claude-code. The skill configuration is stored at ~/.claude/skills/blog-draft/SKILL.md.

What does the blog-draft skill do?

It walks you through research, outlining, and drafting a blog post while enforcing source citations. It stores all materials in structured directories and tracks progress across incremental draft files.

Does blog-draft support editing existing articles?

Yes. When asked to update an existing post, it skips initial research steps and starts directly with the existing draft to produce the next iteration.

Source text

SKILL.md · MIT
---
name: blog-draft
description: Draft a blog post from ideas and resources. Use when users want to write a blog post, create content from research, or draft articles. Guides through research, brainstorming, outlining, and iterative drafting with version control.
---

## User Input

```text
$ARGUMENTS
```

You **MUST** consider the user input before proceeding. User should provide:
- **Idea/Topic**: The main concept or theme for the blog post
- **Resources**: URLs, files, or references to research (optional but recommended)
- **Target audience**: Who the blog post is for (optional)
- **Tone/Style**: Formal, casual, technical, etc. (optional)

**IMPORTANT**: If the user is requesting updates to an **existing blog post**, skip steps 0-8 and start directly at **Step 9**. Read the existing draft file(s) first, then proceed with the iteration process.

## Execution Flow

Follow these steps sequentially. **Do not skip steps or proceed without user approval where indicated.**

### Step 0: Create Project Folder

1. Generate a folder name using format: `YYYY-MM-DD-short-topic-name`
   - Use today's date
   - Create a short, URL-friendly slug from the topic (lowercase, hyphens, max 5 words)

2. Create the folder structure:
   ```
   blog-posts/
   └── YYYY-MM-DD-short-topic-name/
       └── resources/
   ```

3. Confirm folder creation with user before proceeding.

### Step 1: Research & Resource Collection

1. Create `resources/` subfolder in the blog post directory

2. For each provided resource:
   - **URLs**: Fetch and save key information to `resources/` as markdown files
   - **Files**: Read and summarize in `resources/`
   - **Topics**: Use web search to gather up-to-date information

3. For each resource, create a summary file in `resources/`:
   - `resources/source-1-[short-name].md`
   - `resources/source-2-[short-name].md`
   - etc.

4. Each summary should include:
   ```markdown
   # Source: [Title/URL]

   ## Key Points
   - Point 1
   - Point 2

   ## Relevant Quotes/Data
   - Quote or statistic 1
   - Quote or statistic 2

   ## How This Relates to Topic
   Brief explanation of relevance
   ```

5. Present research summary to user.

### Step 2: Brainstorm & Clarify

1. Based on the idea and researched resources, present:
   - **Main themes** identified from research
   - **Potential angles** for the blog post
   - **Key points** that should be covered
   - **Gaps** in information that need clarification

2. Ask clarifying questions:
   - What is the main takeaway you want readers to have?
   - Are there specific points from the research you want to emphasize?
   - What's the target length? (short: 500-800 words, medium: 1000-1500, long: 2000+)
   - Any points you want to exclude?

3. **Wait for user responses before proceeding.**

### Step 3: Propose Outline

1. Create a structured outline including:

   ```markdown
   # Blog Post Outline: [Title]

   ## Meta Information
   - **Target Audience**: [who]
   - **Tone**: [style]
   - **Target Length**: [word count]
   - **Main Takeaway**: [key message]

   ## Proposed Structure

   ### Hook/Introduction
   - Opening hook idea
   - Context setting
   - Thesis statement

   ### Section 1: [Title]
   - Key point A
   - Key point B
   - Supporting evidence from [source]

   ### Section 2: [Title]
   - Key point A
   - Key point B

   [Continue for all sections...]

   ### Conclusion
   - Summary of key points
   - Call to action or final thought

   ## Sources to Cite
   - Source 1
   - Source 2
   ```

2. Present outline to user and **ask for approval or modifications**.

### Step 4: Save Approved Outline

1. Once user approves the outline, save it to `OUTLINE.md` in the blog post folder.

2. Confirm the outline has been saved.

### Step 5: Commit Outline (if in git repo)

1. Check if current directory is a git repository.

2. If yes:
   - Stage the new files: blog post folder, resources, and OUTLINE.md
   - Create commit with message: `docs: Add outline for blog post - [topic-name]`
   - Push to remote

3. If not a git repo, skip this step and inform user.

### Step 6: Write Draft

1. Based on the approved outline, write the full blog post draft.

2. Follow the structure from OUTLINE.md exactly.

3. Include:
   - Engaging introduction with hook
   - Clear section headers
   - Supporting evidence and examples from research
   - Smooth transitions between sections
   - Strong conclusion with takeaway
   - **Citations**: All comparisons, statistics, data points, and factual claims MUST cite the original source

4. Save the draft as `draft-v0.1.md` in the blog post folder.

5. Format:
   ```markdown
   # [Blog Post Title]

   *[Optional: subtitle or tagline]*

   [Full content with inline citations...]

   ---

   ## References
   - [1] Source 1 Title - URL or Citation
   - [2] Source 2 Title - URL or Citation
   - [3] Source 3 Title - URL or Citation
   ```

6. **Citation Requirements**:
   - Every data point, statistic, or comparison MUST have an inline citation
   - Use numbered references [1], [2], etc., or named citations [Source Name]
   - Link citations to the References section at the end
   - Example: "Studies show that 65% of developers prefer TypeScript [1]"
   - Example: "React outperforms Vue in rendering speed by 20% [React Benchmarks 2024]"

### Step 7: Commit Draft (if in git repo)

1. Check if in git repository.

2. If yes:
   - Stage the draft file
   - Create commit with message: `docs: Add draft v0.1 for blog post - [topic-name]`
   - Push to remote

3. If not a git repo, skip and inform user.

### Step 8: Present Draft for Review

1. Present the draft content to user.

2. Ask for feedback:
   - Overall impression?
   - Sections that need expansion or reduction?
   - Tone adjustments needed?
   - Missing information?
   - Specific edits or rewrites?

3. **Wait for user response.**

### Step 9: Iterate or Finalize

**If user requests changes:**
1. Note all requested modifications
2. Return to Step 6 with the following adjustments:
   - Increment version number (v0.2, v0.3, etc.)
   - Incorporate all feedback
   - Save as `draft-v[X.Y].md`
   - Repeat Steps 7-8

**If user approves:**
1. Confirm the final draft version
2. Optionally rename to `final.md` if user requests
3. Summarize the blog post creation process:
   - Total versions created
   - Key changes between versions
   - Final word count
   - Files created

## Version Tracking

All drafts are preserved with incremental versioning:
- `draft-v0.1.md` - Initial draft
- `draft-v0.2.md` - After first round of feedback
- `draft-v0.3.md` - After second round of feedback
- etc.

This allows tracking the evolution of the blog post and reverting if needed.

## Output Files Structure

```
blog-posts/
└── YYYY-MM-DD-topic-name/
    ├── resources/
    │   ├── source-1-name.md
    │   ├── source-2-name.md
    │   └── ...
    ├── OUTLINE.md
    ├── draft-v0.1.md
    ├── draft-v0.2.md (if iterations)
    └── draft-v0.3.md (if more iterations)
```

## Tips for Quality

- **Hook**: Start with a question, surprising fact, or relatable scenario
- **Flow**: Each paragraph should connect to the next
- **Evidence**: Support claims with data from research
- **Citations**: ALWAYS cite sources for:
  - All statistics and data points (e.g., "According to [Source], 75% of...")
  - Comparisons between products, services, or approaches (e.g., "X performs 2x faster than Y [Source]")
  - Factual claims about market trends, research findings, or benchmarks
  - Use inline citations with format: [Source Name] or [Author, Year]
- **Voice**: Maintain consistent tone throughout
- **Length**: Respect the target word count
- **Readability**: Use short paragraphs, bullet points where appropriate
- **CTA**: End with a clear call-to-action or thought-provoking question

## Notes

- Always wait for user approval at outlined checkpoints
- Preserve all draft versions for history
- Use web search for up-to-date information when URLs are provided
- If resources are insufficient, ask user for more or suggest additional research
- Adapt tone based on target audience (technical, general, business, etc.)

---

**Last Updated**: August 4, 2026
**Claude Code Version**: 2.1.220
**Sources**:
- https://code.claude.com/docs/en/skills
**Compatible Models**: Claude Fable 5, Claude Opus 5, Claude Sonnet 5, Claude Sonnet 4.6, Claude Opus 4.8, Claude Haiku 4.5

© luongnv89 · MIT · luongnv89/claude-howto

Similar skills

All skills for Writing and docs →

blog-draft by other authors comes first.

From the same set13

code-review-specialist Audits pull requests and scripts for security bugs, performance bottlenecks, and maintainability. Generates a 1 to 5 rating with concrete code fixes. CX Codex +1 ★ 45,376repo stars claude-md Creates, updates or audits CLAUDE.md following agent-onboarding practice. Handy once the file grows long and the agent stops following it. CX Codex +1 ★ 45,376repo stars blog-draft Guides you through researching, outlining, and drafting blog articles with inline citations and version tracking in Git. CX Codex +1 ★ 45,376repo stars refactor Walks the agent through six refactoring phases: code analysis, test check, code smell hunt, plan, incremental edits and review. CX Codex +1 ★ 45,376repo stars doc-generator Reads your source and writes endpoint docs: OpenAPI specs, request examples, error codes. Useful when the API grew faster than the README. CX Codex +1 ★ 45,376repo stars code-review-specialist Evaluates code for security flaws, performance bottlenecks, and design issues. Useful when reviewing pull requests or planning refactoring. CX Codex +1 ★ 45,376repo stars claude-md Creates, updates and audits CLAUDE.md against a set of rules that keep the file short and useful for an agent. Handy when your project instructions have grown out of hand. CX Codex +1 ★ 45,376repo stars brand-voice A skill that holds marketing copy, customer emails and public pages to one tone of voice. Handy when several people write for the same brand. CX Codex +1 ★ 45,376repo stars

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor