Skip to content
Back to skills

Blog Generate

ASecurity

This skill should be used when the user asks to "turn this lab into a blog post", "write a blog post from this module", "generate a blog from my Showroom content", "convert my workshop to a blog", "write a Red Hat Developer blog post", or "create a blog post from my demo".

  • 22 stars
  • 0 votes
  • 0 copies
  • 2 views
  • Added June 2, 2026
documentationnode

Security analysis

A100/100

Pro scans all 2 files and shows the line behind each finding

Scanned June 2, 2026

npx -y skills add rhpds/rhdp-skills-marketplace --skill blog-generate --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Blog Generate?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Blog Generate
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/rhpds-blog-generate/badge)](https://www.skillsdirectory.com/skills/rhpds-blog-generate)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: showroom:blog-generate
description: This skill should be used when the user asks to "turn this lab into a blog post", "write a blog post from this module", "generate a blog from my Showroom content", "convert my workshop to a blog", "write a Red Hat Developer blog post", or "create a blog post from my demo".
---

---
context: main
model: claude-sonnet-4-6
---

# Blog Generator

Orchestrates agents to transform completed Red Hat Showroom lab or demo content into blog posts for Red Hat Developer, internal blogs, or marketing platforms.

## Architecture

This skill is an orchestrator. Transformation and review are delegated to agents:
- `showroom:file-generator` (Sonnet) — reads source .adoc files, transforms to blog Markdown
- `showroom:module-reviewer` (Sonnet) — quality check on generated blog post

See `@showroom/docs/SKILL-COMMON-RULES.md` for source traceability and attribution rules.

---

## Phase 1 — Planning Form (ALL questions at once)

```
Let's set up your blog post. Answer what you know:

Source modules (file paths to completed .adoc modules — can be multiple):

Target platform:
  1. Red Hat Developer (developers.redhat.com)
  2. Internal Red Hat (Source, Memo, The Stack)
  3. Medium / dev.to / Hashnode
  4. Marketing / announcement

Blog type:
  1. Technical tutorial ("How to...")
  2. Product announcement ("Introducing...")
  3. Thought leadership ("Why...")
  4. Case study / success story
  5. Quick start guide

Technical depth:
  1. Highly technical (code-heavy, for developers)
  2. Moderately technical (balanced)
  3. Marketing-focused (business benefits, light on code)

Target word count: 500-800 / 1000-1500 / 2000+ (or leave blank for auto)

Showroom link for "Try it yourself" CTA (optional):

Writing style (optional — skip for standard Red Hat technical blog style):
  Describe: "storytelling-first, relatable analogies, humor where appropriate"
  OR paste 1-3 paragraphs from a blog post you wrote
  OR give a URL to a published post you authored
  Saved profile? Point to: ~/.claude/context/my-writing-style.md
```

Confirm in one line:
```
📋 Blog plan: [type] for [platform] — [depth] — [word count]
Source: [N modules]

Generating blog post...
```

---

## Phase 2 — Spawn File Generator Agent

```
Task tool:
  subagent_type: showroom:file-generator
  prompt: |
    TARGET_FILE: <output path, e.g. blog-post.md>
    FILE_TYPE: blog
    CONTENT_TYPE: <workshop|demo>
    LAB_TYPE: <ocp|rhel|vm|ai|unknown>
    REPO_PATH: <repo root if known, else none>
    FULL_SPEC:
      source_files: [<list of .adoc paths>]
      blog_type: <tutorial|announcement|thought-leadership|case-study|quick-start>
      platform: <redhat-developer|internal|medium|marketing>
      technical_depth: <highly-technical|moderately-technical|marketing-focused>
      word_count: <500-800|1000-1500|2000+>
      showroom_link: <URL or null>
      lab_name: <auto-detected from source>
```

---

## Phase 3 — Inline Quality Check

The file-generator agent applies blog-specific quality checks internally. After it returns, verify:

- Tone is narrative and conversational — not a numbered step list
- Word count matches requested target (±20%)
- At least one code block is present for technical posts
- CTA appears at the end (if showroom_link was provided)
- All external links are valid references from source material
- Source attribution present (one of):
  - Red Hat Developer: "This post is based on the workshop [Lab Name](showroom_link)"
  - Internal: "Source: [original module path]"
  - Marketing: "Based on customer use cases from [reference]"

Fix any issues inline before delivering.

---

## Phase 4 — Deliver

```
✅ Blog post generated: <filename> (<word_count> words)

Quality: 0 Critical, 0 High, N Warnings
  [list warnings if any]

Platform guidance:
  Red Hat Developer: Submit at developers.redhat.com/write-for-us
  Internal: Post to Source or Memo
```

---

## Blog types reference

| Source | Blog type | What changes |
|---|---|---|
| Workshop | Tutorial | Exercises → narrative how-to, keep code |
| Workshop | Quick start | First module only, condensed |
| Demo Know | Announcement | Business value extraction |
| Demo Show | Case study | Capability walkthrough |
| Any | Thought leadership | Patterns + trends, minimal code |

---

## Related Skills

- `/showroom:create-lab` — create workshop content
- `/showroom:create-demo` — create demo content
- `/showroom:verify-content` — verify content quality

Files in this skill

  • SKILL.md4.4 KB
  • workflow.svg3.7 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.