/jab-hook

Inspiration accounts

Your personal list lives at ${MAKERSKILLS_CONFIG:-$HOME/.config/makerskills}/jab-hook/inspiration.local.md — not in this repo (gitignored, on disk only). The list of operators you're learning from is your taste — keep it private.

See inspiration.example.md in this directory for the schema + setup instructions.

How the skill loads it

1. Read ${MAKERSKILLS_CONFIG:-$HOME/.config/makerskills}/jab-hook/inspiration.local.md for the personal account list
2. If missing, skip the inspiration scan (skill still works, just without external pattern pull)

Per-platform organization (load-bearing)

The local file is organized by platform — ## LinkedIn, ## X (Twitter), ## Instagram, etc. — because what works on one platform usually doesn't work on another:

• LinkedIn: long-form professional, carousel/document posts, framework breakdowns, narrative + takeaway
• X: short (≤280 chars), thread mechanics, reply-as-content, punchier + snarkier
• Instagram: visual-first, reel-driven, lifestyle / behind-the-scenes
• TikTok: vertical video, hook-heavy, trend-aware short-form
• YouTube: long-form (more relevant to youtubeskills than jab-hook)

When the skill scans inspiration, it pulls from the section matching the TARGET PLATFORM of the draft. Cross-posted drafts (e.g., X + LinkedIn) scan both sections.

Same person, two platforms

When an operator appears on multiple platforms with strong presence on each, list them in both relevant sections — their craft per platform is usually different. Treat each listing as a separate inspiration source for that platform.

Pattern library — also local + platform-tagged

When the skill extracts hook patterns / structural patterns from inspiration-account posts, it writes them to:

${MAKERSKILLS_CONFIG:-$HOME/.config/makerskills}/jab-hook/patterns.local.md

Each pattern is tagged with the source platform (e.g., platform: linkedin) so future scans know which patterns transfer where. Some patterns are platform-portable (story-open hooks); others are platform-specific (LinkedIn carousel structure, X thread mechanics).

This file is gitignored. The pattern library grows with use and stays your private playbook.

Methodology (this stays public — recipe, not data)

When drafting a promo or educational post:

1. Determine the target platform(s) — X, LinkedIn, both, etc.
2. Pick 1–2 accounts from the matching platform section(s) whose audience overlaps the property being promoted
3. Pull their recent posts via social-fetch
   • LinkedIn: agent-browser with modal dismissal (no API key needed) or ScrapeCreators if SCRAPECREATORS_API_KEY set for full post data
   • X: ScrapeCreators recommended; free tier returns preview only
4. Extract: hook openers, structural patterns (numbered list / contrarian-take / story-open / thread mechanics), CTA styles, post length, line-break rhythm
5. Choose ONE pattern that fits the property + content type + platform conventions
6. Apply the pattern to YOUR voice — don't mimic their voice, only the structure

Notes

• LinkedIn often blocks unauthenticated scraping. If agent-browser walls, fall back to: pasting examples, pulling from their public newsletters, or scraping their X accounts.
• Never copy phrasing — only structure, format, hook archetype.
• Inspiration is a starting point. Your voice rules in voice.md (+ voice.local.md overlay) always override format choices.
• Some patterns transfer across platforms (story-open hooks); others don't (LinkedIn carousels won't work on X). The patterns.local.md platform tag helps the skill respect the boundary.