| name | soc-media-writer |
| description | Write high-performing social media content engineered around named copywriting frameworks, scroll-stopping hooks, and retention mechanics. For video it produces three coordinated outputs — the spoken DIALOGUE (voiceover), the VIDEO SCENARIO (timed two-column shot list with visual cues), and the EXPLANATION (notes on the frameworks, hook, and retention beats used) — plus standalone written posts for LinkedIn, X, Instagram, and YouTube. Use whenever the user wants a social post, a Reel/Short/TikTok script, a video scenario or shot list, a hook or hook bank, a content matrix, a Hub-and-Spoke repurposing plan, a VSL, or any short- or long-form distribution content. Trigger even on casual asks like "write me a Reel about X," "give me hooks for Y," "script a 30-second video," "make this go viral," "write a LinkedIn post," or "turn this into TikToks." Also trigger when the user wants retention engineered in (e.g. "keep them watching," "re-hook every N seconds," "high retention script"). |
Social Media Writer
This skill turns a topic into engineered, distribution-ready content. It is built on a closed corpus of named, attributable frameworks (AIDA, PAS, Olson's ABT, McDonald's Pixar Story Spine, Welsh's Content Matrix, Hormozi's Value Equation, Galloway's Click & Watch) and the real platform numbers that govern reach (word budgets, retention thresholds, safe zones, algorithm weights). Nothing here is generic "engagement" advice — every choice maps to a framework and a metric.
House rules (always apply)
These mirror the HasBrainsAI / Total 66 content standard. Do not violate them:
- No filler. Every line earns its place. Cut throat-clearing ("In today's video..."), hedges, and restated obviousness.
- Name the framework with attribution. When you use a structure, the EXPLANATION names it and its originator (e.g. "ABT — Randy Olson," "Value Equation — Alex Hormozi"). Never present a borrowed framework as anonymous.
- Use real benchmarks. Quote the actual numbers from the reference files (140–160 wpm, ≥65% retention under 30s, 1.7s hook window) rather than vague claims like "keep it short."
- Match register to platform. A LinkedIn post is not a TikTok script. Pull the platform's rules before writing.
- Never fabricate proof. If a script needs a stat, credential, or testimonial the user hasn't given, leave a
[INSERT: ...] placeholder. Do not invent results.
Step 1 — Read the brief, then pull only what you need
Before writing, establish four things (ask only if the user hasn't already implied them; infer where you safely can):
| Variable | What it decides |
|---|
| Format | Written post (LinkedIn/X) vs. video (Reel/Short/TikTok) vs. long-form vs. VSL |
| Platform | Which algorithm rules, word budget, and safe zone apply |
| Funnel stage | TOFU / MOFU / BOFU → which copywriting framework fits |
| Duration (video only) | Sets the word budget and visual cut cadence |
Then load only the relevant reference files (progressive disclosure — don't read all five every time):
references/hooks.md — almost always. The hook is the single highest-leverage element. Read it for any video or any post that opens with a line.references/copywriting-frameworks.md — for written posts, VSLs, ad copy, and the body structure of any piece. Contains the framework-by-funnel-stage selection table.references/video-scripting.md — for any video. Word-budget math, the two-column scenario format, narrative frameworks (HHVCTA, ABT/ABYT, Pixar Story Spine), and the retention beat engine.references/platform-rules.md — for any video or platform-specific post. Algorithm weights, retention targets, and the pixel-exact safe zones.references/creator-systems.md — when the user wants a content system, a repurposing plan, a content matrix, value positioning, or thumbnail/title packaging.
Step 2 — Choose the framework (don't default to one)
The funnel stage drives the choice. Full selection logic is in references/copywriting-frameworks.md; the short version:
- TOFU / discovery → AIDA (curiosity + aspiration) or a Quip-style single insight.
- MOFU / problem-aware → PAS (Problem–Agitate–Solution) or BAB (Before–After–Bridge).
- BOFU / ready to act → FAB, 4Ps, or for high-ticket, the long-form AICPBSAWN / PASTOR.
- Video narrative spine → ABT for a single tension arc, HHVCTA for a how-to, Pixar Story Spine for a transformation story or case study, Story Circle (Harmon) for episode-length or series content.
- Thematic / manifesto content → Moral Premise (Stan Williams) as invisible spine — never stated directly, demonstrated through beats.
State your framework choice in the EXPLANATION so the user can override it.
Step 3 — Build the hook
Read references/hooks.md and select a hook category that matches the niche and avatar:
- Curiosity Gap — open a loop the brain wants to close.
- Loss Aversion — surface a mistake or wasted effort.
- Contrarian — challenge accepted advice to spark debate.
- "The Why" Hook — present a logical impossibility or surprising outcome; triggers active problem-solving.
- In Medias Res — drop into a scene already in motion; the viewer is inside the story before deciding whether to care.
- Contradicting Emotions — juxtapose conflicting feelings; signals transformation content and makes the viewer stay to resolve the contradiction.
Every specific claim or element named in the hook is a Chekhov's Gun — it must be paid off explicitly in the body. An unclosed hook loop is clickbait and tanks completion rate.
Step 4 — Engineer retention (video)
Retention is not "make it interesting" — it is a placed schedule of beats. The full retention beat engine is in references/video-scripting.md. Two layers, both required:
- Visual refresh — every 3–5 seconds (the document's standard): a cut, zoom, b-roll insert, graphic pop-up, or sound effect. Tighter for shorter videos (every 1–1.5s at 15s; every 3–4s at 60s).
- Retention checkpoint / re-hook — default every ~9 seconds (configurable; set by the user's request): a stronger pattern interrupt that re-earns attention — a new open loop, a payoff tease, a tension bump, or a "but here's the thing" turn. This sits on top of the visual cuts and is what prevents mid-video drop-off.
Mark every beat explicitly in the VIDEO SCENARIO with its timestamp and type. If the user names a different cadence (e.g. "re-hook every 6 seconds"), use theirs; 9s is only the default.
Map the whole script against a narrative spine (HHVCTA / ABT / Pixar / Story Circle) so value is distributed across the runtime — never back-loaded, since most viewers never reach the end. For transformation content, layer in the Scene & Sequel micro-engine from references/video-scripting.md to chain beats causally rather than episodically. For videos with no external antagonist, install a Ticking Clock to supply urgency.
Step 5 — Respect the constraints
- Word budget (video):
words ≤ duration_seconds × 2.5. A 30s video ≤ 75 words; 60s ≤ 150 words. Enforce this — an over-budget script reads rushed. See references/video-scripting.md.
- Safe zones: keep all text overlays, faces, logos, and CTAs inside the platform's safe container (e.g. Instagram Reels central 840×1350 box on a 1080×1920 canvas). Specify placement in the scenario. See
references/platform-rules.md.
- Retention target: state the metric the piece is built to hit (e.g. "engineered for ≥65% average view duration on Shorts under 30s").
Output formats
A) Video request → produce all THREE outputs, in this order
Use this exact structure (the full template with an example is in assets/video-script-template.md — read it before your first video output):
## 1. VIDEO DIALOGUE (the spoken voiceover)
[Clean spoken script only, within word budget. No stage directions.
Word count + duration stated at top.]
## 2. VIDEO SCENARIO (timed two-column shot list)
| Time | VISUAL (shot / b-roll / on-screen text / cut) | AUDIO (spoken line + SFX) |
| 0.0–1.0s | [pattern interrupt + opening text overlay] | [hook line] |
| ... | ... [mark every visual cut and every retention beat] | ... |
[Retention beats flagged inline: ⟳ visual refresh, ★ retention checkpoint/re-hook]
## 3. EXPLANATION (director's notes)
- Framework: [named + originator] and why it fits the funnel stage
- Hook: [category] and the behavioral trigger it uses
- Retention schedule: where the beats land and why
- Platform fit: word budget used, retention target, safe-zone notes
- Levers to pull if it underperforms
B) Written post request → produce the post + a short EXPLANATION
## POST
[The ready-to-publish copy, formatted for the platform — line breaks, hook line,
body via the chosen framework, single clear CTA.]
## EXPLANATION
- Framework + originator, funnel stage, hook category, CTA logic, platform fit.
C) System request (content matrix / Hub-and-Spoke / packaging) → follow references/creator-systems.md
Produce the actual filled artifact (a populated matrix, a staggered spoke calendar, title/thumbnail options), not a description of the method. Templates are in assets/.
Files in this skill
soc-media-writer/
├── SKILL.md ← you are here (orchestrator)
├── references/
│ ├── hooks.md hook categories, formulas, visual pattern interrupts
│ ├── copywriting-frameworks.md AIDA/PAS/BAB/FAB/4Ps/PASTOR/QUEST/ACCA/OATH + funnel table
│ ├── video-scripting.md word-budget math, two-column format, HHVCTA/ABT/Pixar, retention engine
│ ├── platform-rules.md YouTube/TikTok/Reels/X algorithm weights, retention targets, safe zones
│ └── creator-systems.md Welsh, Hormozi, Galloway systems + DM funnels
└── assets/
├── video-script-template.md the three-output template + a worked example
└── content-matrix-template.md blank Welsh Content Matrix + Hub-and-Spoke calendar
