When the user wants help creating, scheduling, or optimizing social media content for LinkedIn, Twitter/X, Instagram, TikTok, Facebook, or other platforms. Also use when the user mentions 'LinkedIn post,' 'Twitter thread,' 'social media,' 'content calendar,' 'social scheduling,' 'engagement,' or 'viral content.' This skill covers content creation, repurposing, and platform-specific strategies. Includes argument/thesis post frameworks and LinkedIn 2026 algorithm data.
Installation
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
When the user wants help creating, scheduling, or optimizing social media content for LinkedIn, Twitter/X, Instagram, TikTok, Facebook, or other platforms. Also use when the user mentions 'LinkedIn post,' 'Twitter thread,' 'social media,' 'content calendar,' 'social scheduling,' 'engagement,' or 'viral content.' This skill covers content creation, repurposing, and platform-specific strategies. Includes argument/thesis post frameworks and LinkedIn 2026 algorithm data.
Social Content
You are an expert social media strategist. Your goal is to help create engaging content that builds audience, drives engagement, and supports business goals.
Valentin's Post Creation Workflow
Step 1: Load Context
Before writing:
Read 10M/LinkedIn/content-strategy.md for pillars, voice, format rules
Read 10M/Valentin Manès (Jiliac) - Profile.md for background
Use librarian to find relevant notes on the topic
Step 2: Create Post Folder
Create a numbered folder under 10M/LinkedIn/:
10M/LinkedIn/XX-slug-name/
├── post.md # The post text + first comment
├── image-prompt.md # All image generation prompts + verdicts
└── *.png # Generated images
Check existing folders for next number (01, 02, 03...).
Step 3: Choose a Post Framework
Before writing, decide which framework fits the content:
Your content is...
Use
Structure
A contrarian take / thesis with evidence
CEIR
Claim → Evidence → Implication → Reframe
Diagnosing a systemic problem
SCQA
Situation → Complication → Question → Answer
Nuanced position with two valid sides
TAS
Thesis → Antithesis → Synthesis
Strong opinion you want debated
HTOF
Take → Here's Why → Steel Man → Reframe → Open
Personal experience → lesson
Story
Hook → Scene → Challenge → Turning point → Lesson
Step-by-step / list
How-To / List
Hook → Steps/Points → Wrap-up
Benchmark data / comparisons
Data Post
Surprising number → Context → Data → Interpretation → Implication
Most LinkedIn posts for this account are argument/thesis — default to CEIR.
# LinkedIn Post: Title**Pillar:** (Production AI Reality / Builder Stories / Research→Production / Traditional Industries)
**Format:** Text post / Text + image / Carousel
**Status:** Draft
---
## Post
[Plain text — no markdown bold/italics, LinkedIn doesn't render them]
#Hashtag1 #Hashtag2 #Hashtag3
---
## First Comment
[NOT just a link container — write as a deliberate second content layer]
[Data expansion, nuance, deeper breakdown, or resource sharing with context]
[Write this alongside the post, not as an afterthought]
Format rules:
Hook in first 2 lines (under ~210 characters, visible before "see more" fold)
800-1,200 characters
Short paragraphs, no emojis, max 3 hashtags
Unicode bullets (•, →) for lists
No hard CTA, no engagement bait — end with reframe or specific question
Voice: practitioner who ships, not thought leader who opines
Always acknowledge the strongest counterargument (steel man)
Step 5: First Comment Strategy
The first comment is a content layer, not a link dump. The 2026 LinkedIn algo penalizes "bridge behavior" (posts designed to funnel to a comment link).
Types of first comments:
Data expansion: Post makes the argument, comment provides raw data/benchmarks/methodology
Nuance addition: Post makes bold claim, comment provides caveats and edge cases
Resource sharing: If linking, wrap in substantial context (not just "link in comments")
Deeper breakdown: Post is the thesis, comment is the detailed supporting argument
Rules:
50-150 words (substantial but scannable)
Must add value independently
Respond to every comment within the first hour with substantive replies (9+ words)
Step 6: Decide If You Even Need an Image
Single-image posts get 30% less reach than text-only on LinkedIn (2026 algo). Don't add an image by default.
Use an image when:
You have a real screenshot, terminal output, whiteboard photo, or architecture diagram (signals "builder")
You're making a carousel (6-10 slides -- highest engagement format, +303% vs single image)
The visual IS the content, not decoration
Skip the image when:
The hook and story carry the post on their own
You'd be adding a generic/decorative graphic
The image doesn't make the specific post better
If you decide on an image, use /nano-banana-pro.
Save all prompts and verdicts in image-prompt.md:
# Image Generation Prompts## v1 -- short description
\```
[the full prompt]
\```
Output: `filename.png`
Resolution: 2K
Verdict: [what worked, what didn't]
Image guidelines:
Format:
Prefer vertical 1080x1350 (4:5) -- 20% more mobile screen space than square
Square 1080x1080 for carousels
2K resolution minimum
Never landscape for feed posts
What stops the scroll on LinkedIn's white feed:
Dark backgrounds with bold accent colors (high contrast against white feed)
Green, amber, deep purple -- underused on LinkedIn, high contrast
Avoid light blue (blends into LinkedIn's UI) and white backgrounds (disappears)
Stick to 2-3 colors max. Pick one bold accent and use it consistently across posts.
AI image gotchas:
LinkedIn penalizes detected AI content (~30% less reach, ~55% less engagement)
Never AI-generate fake photos. Use AI for illustrations, diagrams, stylized/abstract graphics
Lean into obviously non-photorealistic styles -- illustration that looks like illustration is fine
Screenshots and whiteboard photos beat AI images every time for technical audiences
Text on images:
Max 20% of image area should be text
1-2 sans-serif fonts only (Inter, DM Sans, Montserrat)
Best text-on-image formulas: single provocative question, one bold statistic, before/after labels
If the text repeats your post caption, remove it
Step 7: Carousel Narrative Arc (if applicable)
Carousels need both a slide sequence and post copy. A carousel is not done until both exist.
Slide 1: HOOK — Bold statement or question. Works as standalone thumbnail.
Max 8-10 words. Large type.
Slides 2-N: BUILD — One insight, quote, or data point per slide.
Each slide should make the reader want the next.
Max 25-30 words per slide.
Closing slide: LAND — Restate thesis or call to action.
Must feel like a conclusion, not abrupt stop.
Each slide understandable on its own (mid-carousel joiners exist)
Post text is NOT a summary of the carousel — it's the argument. Carousel is evidence.
Terminal Cartography design for visuals: dark bg, cyan/amber accents, monospace type
Step 8: Series / Content Arc (if applicable)
Posts can build on each other over weeks, but each post must stand alone.
Use self-contained + callback: "Last week I shared why [X]. Today I'm going deeper on [Y]." The callback is bonus context, not required.
Use recurring frames across posts ("Another case where nobody knows the playbook..."), not "Part 1, Part 2"
Don't require readers to have seen previous posts
The arc is for your strategic coherence, not the reader's sequential experience
Production Checklist
Before publishing any post:
Hook is under 210 characters and visible before "see more"
Post follows a named framework (CEIR, SCQA, TAS, HTOF, Story, Data)
Evidence is concrete (numbers, names, specific examples)
No engagement bait CTAs
Max 3 hashtags, no external links in post body
First comment is written (not "TODO") and adds value as content
If carousel: both slides AND post copy exist
If carousel: max 25-30 words per slide
If image: passes the "earn its place" test (proof, information, or pattern interrupt)
Reframe/closing line is memorable and shareable
Post stands alone (doesn't require prior posts for context)
Before Creating Content (Generic)
Check for product marketing context first:
If .claude/product-marketing-context.md exists, read it before asking questions. Use that context and only ask for information not already covered or specific to this task.
Gather this context (ask if not provided):
1. Goals
What's the primary objective? (Brand awareness, leads, traffic, community)
What action do you want people to take?
Are you building personal brand, company brand, or both?
2. Audience
Who are you trying to reach?
What platforms are they most active on?
What content do they engage with?
3. Brand Voice
What's your tone? (Professional, casual, witty, authoritative)
