blog-team

name	blog-team
description	Launch the blog post creation team — interviewer, researcher, voice reviewer, QA checker in parallel tmux panes
arguments	[{"name":"topic","description":"The blog post topic or a brief description","required":true},{"name":"links","description":"Comma-separated URLs to research (web pages, YouTube videos)","required":false}]

Orchestrate a blog post creation team through 4 phases for: $ARGUMENTS

Phase 1: Gather (Parallel)

Launch two Agent subagents in parallel:

1. blog-interviewer — spawned as Agent with subagent_type: "blog-interviewer". Interviews the user via SendMessage to gather scope, opinions, key points, and personal experiences.
2. blog-researcher — spawned as Agent with subagent_type: "blog-researcher". Researches the topic by scraping links, searching the web, and extracting YouTube transcripts.

Pass the topic and links to both agents. Wait for both to complete.

Phase 2: Draft

Combine interviewer notes + researcher findings. Write the blog post to packages/blog/content/2.blog/YYYYMMDD.slug-name.md with complete frontmatter.

Frontmatter Template

---
title: 'Post Title'
description: 'Brief summary for SEO'
date: 'YYYY-MM-DD'
image:
  src: '/images/blog/YYYYMMDD-HHMM-slug.png'
  alt: 'Description'
authors:
  - name: Chris Towles
    to: https://twitter.com/Chris_Towles
    avatar:
      src: /images/ctowles-profile-512x512.png
badge:
  label: 'Category'
---

Drafting Guidelines

• Use the brand-voice skill for voice/tone rules and the blog-content-architect skill for personality
• Use the interviewer's gathered answers as the backbone — these are the user's actual opinions and experiences
• Use the researcher's findings as supporting evidence and context

Phase 3: Review (Parallel)

Launch three tasks in parallel:

1. blog-voice-reviewer — spawned as Agent with subagent_type: "blog-voice-reviewer". Reviews draft for brand voice consistency.
2. blog-qa-checker — spawned as Agent with subagent_type: "blog-qa-checker". Validates frontmatter, links, lint, typecheck, tests.
3. blog-image — use the blog-image skill to generate the hero image.

Wait for all to complete.

Phase 4: Report

Apply any revisions from voice-reviewer and qa-checker. Report summary to user:

• Post path and title
• Image status
• QA results
• Any issues that need manual attention

Error Handling

• If an agent fails, note the failure and continue with remaining agents
• If image generation fails, leave the frontmatter image path as a placeholder and note it in the report
• Always complete Phase 4 even if other phases had failures