Social Content Writer

Social Content Writer

Generates ready-to-publish social media posts tailored to each platform's native format, algorithm preferences, and audience expectations. Every post is built from brand voice in agency.config.json, not generic marketing copy.

Prerequisites

• agency.config.json populated (agency info, case studies, services, outreach tone)
• Topic or theme to write about
• Optional: research output from blog-researcher or linkedin-researcher

Phase 0: Intake

Read agency.config.json:

• agency.name, agency.founder, agency.tagline -- brand identity
• outreach.tone -- voice/style baseline (adapt per platform)
• outreach.banned_phrases -- never use these
• case_studies[] -- proof points for case study posts
• services[] -- service areas for thought leadership angles

Accept parameters:

• platform -- (required) one of: linkedin, twitter, instagram, reddit
• post_type -- (required) one of: thought-leadership, case-study-snippet, hot-take, engagement-question, carousel-outline, thread, story-caption
• topic -- (required) subject or theme
• research -- (optional) output from researcher skills for data-backed posts
• case_study -- (optional) specific case study slug from config
• cta_type -- (optional) comment, dm, link, follow, none. Default: comment
• count -- (optional) number of variants to generate. Default: 2

Phase 1: Platform Voice Adaptation

Adapt the base outreach.tone for each platform:

Phase 2: Post Type Generation

Thought Leadership

• Open with a contrarian or surprising insight from the founder's perspective
• Support with data, observation, or experience
• Close with a question or takeaway
• LinkedIn: 150-300 words, line breaks every 1-2 sentences
• Twitter: Single tweet (280 chars) or thread (3-7 tweets)
• Instagram: 100-200 word caption paired with a quote card concept
• Reddit: 150-400 words matching the subreddit's expertise level

Case Study Snippet

• Lead with the result (number, metric, transformation)
• Brief context (who, what problem)
• What was done (2-3 bullet points max)
• Result reinforced
• Soft CTA
• LinkedIn: 150-250 words, use bullet points for the "what we did"
• Twitter: Single tweet with key metric + thread for breakdown
• Instagram: Stat-forward caption with before/after framing

Hot Take

• Bold, opinionated statement on an industry trend
• Back it up with one proof point
• Invite disagreement
• LinkedIn: 50-150 words, conversational
• Twitter: Single tweet, max edge without being hostile
• Instagram: Quote-style caption

Engagement Question

• Ask a genuine question the audience cares about
• Provide 2-3 options or your own take first
• Designed to drive comments
• LinkedIn: Question + context (50-100 words)
• Twitter: Question only (under 200 chars)
• Instagram: Question in caption, use poll sticker concept for stories

Carousel Outline

• LinkedIn/Instagram only
• 5-10 slides with headline + supporting text per slide
• Slide 1: Hook (scroll-stopping statement)
• Slides 2-8: Core content (one idea per slide)
• Slide 9: Summary or key takeaway
• Slide 10: CTA slide (follow, DM, comment)
• Output as structured slide-by-slide text

Thread

• Twitter/LinkedIn only
• 3-10 posts in sequence
• Post 1: Hook that promises value ("Here's how we 3x'd conversion rate for a sports brand")
• Posts 2-8: One insight per post, each standalone-readable
• Final post: Summary + CTA
• Number each: 1/, 2/, etc.

Story Caption

• Instagram only
• Ultra-short (1-2 lines)
• Designed to pair with a visual
• Include engagement prompt (poll, question, emoji slider)

Phase 3: Quality Check

Before returning, verify every post:

1. Banned phrase scan: Check against outreach.banned_phrases. Rewrite if any match.
2. Character limits: Twitter posts under 280 chars. LinkedIn under 3000 chars. Instagram under 2200 chars.
3. Tone check: Re-read against platform voice. Rewrite if it sounds corporate, generic, or salesy.
4. Hashtag rules: LinkedIn max 5, Instagram 5-15, Twitter max 3, Reddit zero.
5. CTA count: One CTA per post (or none for pure value posts).
6. Self-promo check for Reddit: No direct links or pitches. Help first, credibility naturally.
7. No emojis overload: Max 2-3 per LinkedIn post, 5-8 per Instagram caption, 0-1 per tweet.

Phase 4: Output

Return structured JSON:

{
  "platform": "linkedin",
  "post_type": "thought-leadership",
  "variants": [
    {
      "text": "Full post text with line breaks...",
      "hashtags": ["#Shopify", "#D2C", "#Ecommerce"],
      "character_count": 487,
      "word_count": 82,
      "cta": "What's your take? Drop a comment.",
      "visual_suggestion": "Screenshot of a Shopify analytics dashboard showing the metric",
      "best_posting_time": "Tuesday 9-11 AM IST / Tuesday 7-9 AM GMT"
    },
    {
      "text": "Variant 2...",
      "hashtags": ["#ShopifyPartner", "#EcommerceTips"],
      "character_count": 412,
      "word_count": 71,
      "cta": "DM me if you want the full breakdown.",
      "visual_suggestion": "Carousel with 5 key points from the post",
      "best_posting_time": "Wednesday 12-2 PM IST"
    }
  ],
  "topic": "Why most D2C brands underinvest in product page CRO",
  "personalization_source": "Kibi Sports case study",
  "generated_at": "2026-03-13T10:00:00Z"
}

Phase 5: Scheduling Recommendation

For each post, suggest optimal posting time based on platform best practices:

For India-focused audience, adjust to IST prime hours. For US/UK audience, adjust to their local times.

Example Usage

Trigger phrases:

• "Write a LinkedIn post about CRO for D2C brands"
• "Create a Twitter thread about our Kibi Sports results"
• "Write an Instagram caption for a case study post"
• "Generate a hot take about Shopify themes for LinkedIn"
• "Write a Reddit-friendly comment about Shopify development"
• "Create a LinkedIn carousel outline about ecommerce mistakes"
• "Generate 3 engagement questions for LinkedIn this week"