← Back to list
linkedin-hook-extractor Checked
Reverse-engineer the hook formula from a viral LinkedIn post URL. Returns which of the 20 canonic…
9.1Overall rating
Installation
🤖 Install via AI
Copy the prompt below and send it to your AI assistant (e.g. Claude Code) — it will follow the instructions and install automatically.
Install the "linkedin-hook-extractor" skill by following the instructions at https://skill123.me/install/linkedin-hook-extractor.
⌨️ Command line install
Run in your terminal — downloads and installs to ~/.claude/skills/.
curl -fsSL https://skill123.me/install/linkedin-hook-extractor.sh | bash
📦 Download ZIP
Download the zip and extract it into your skills directory (e.g. ~/.claude/skills/), then restart your session.
⬇ Download v1.0.0 · 5 KBAbout this skill
Overview
Reverse-engineer the hook formula from a viral LinkedIn post URL. Returns which of the 20 canonical 2026 formulas it uses (anaphora, R.I.P., year-pivot, time-anchor, curiosity-gap, contrarian, comment-gate, emotional cold-open, named-gratitude, and 11 more), why it worked, and a blank template. Use to learn from a competitor's post, not to write your own (use linkedin-post-writer).
Source
- Repo: https://github.com/sergebulaev/linkedin-skills
- Path: skills/linkedin-hook-extractor
Score breakdown
Trigger
9.0
Clear purpose with explicit use-for-learning vs post-writer boundary; good trigger coverage.
Structure
9.3
Compact 80-line SKILL.md; classification rules and one worked example as references.
Workflow
8.4
Deterministic feature-to-formula mapping, dual-candidate confidence scoring, paste fallback, source audit step.
Content
10.0
Concrete feature lists and a real worked example; very tight writing.
Engineering
9.0
Correct frontmatter; references exist; bundle-root dependencies only.
Security
9.5
Public post data only, no credentials; fetched text treated as data with prompt-injection defenses.
Analytical skill that fetches a public post via no-cookie Apify (or pasted text) and classifies its hook against 20 documented formulas with confidence scores, then emits a blank template plus source-tell cautions. Feature-mapping rules are concrete; untrusted-content rules are inline.