creator-pack
A Claude Code pack for Instagram & Facebook content creators.
Most "AI for creators" resources are scattered prompts. This one is enforced: the guarantees come from hooks (deterministic - they run every time), with skills, loops, and review agents on top. It's built around the things that actually get creators in trouble or throttle their reach: FTC/Meta disclosure, copyright/music, brand safety, and fabricated claims - while keeping the content human.
Every item serves a goal:
| Icon | Goal |
|---|---|
| 🪙 | Fewer tokens |
| 🎯 | Accuracy |
| ✅ | Truth-telling (no fake stats, real disclosure) |
| ⚡ | Speed |
| 💎 | Quality / craft / compliance |
⚠️ Read first
- Not legal advice. A drafting + guardrail tool. You're responsible for what you publish.
- US-informed. FTC disclosure rules are US; Meta's "Paid partnership" label and Community Standards apply broadly. Check your local rules.
- Hooks are guardrails, not guarantees.
disclosure-guardcatches common patterns - not every case. Run thecompliance-reviewerand read every post before publishing.
What's inside
| Folder | Contents | Layer |
|---|---|---|
hooks/ |
6 deterministic guardrails (disclosure, claims, copyright, brand-safety, secret-scan, review) + settings.json |
Enforced |
skills/ |
15 creator skills (Reels, carousels, Stories, captions, repurpose, sponsorship, analytics...) | On-demand |
loops/ |
6 workflow slash commands | Workflow |
commands/ |
one slash command per skill + agent | Shortcuts |
agents/ |
3 review subagents (content / compliance / engagement) | Review |
rules/ |
8 advisory CLAUDE.md fragments |
Advisory |
mcp/ |
optional connectors (browser / Meta / Canva) | Tooling |
templates/ |
content brief, reel script, calendar, media kit | Docs |
tests/ |
test_hooks.py - runs every hook against sample inputs |
Proof |
See INVENTORY.md for the full table.
Install
./install.ps1 -ProjectPath "C:\path\to\your\workspace" # Windows
./install.sh /path/to/your/workspace # macOS / Linux
Copies hooks -> .claude/hooks/, skills -> .claude/skills/, loops + commands -> .claude/commands/,
agents -> .claude/agents/, appends rules to CLAUDE.md, merges settings.json. Requires Python 3.8+.
Verify: python tests/test_hooks.py (15 cases / 6 hooks). Uninstall: ./uninstall.ps1 -ProjectPath ....
Slash commands
After install, type / in Claude Code:
- Workflows:
/make-reel,/carousel,/story,/repurpose,/calendar,/sponsor. - Skills:
/hook-writer,/reel-script,/carousel-post,/caption-hashtags,/meta-discovery,/content-ideation,/analytics-review,/sponsorship-pitch, ... (one per skill). - Agents:
/content-reviewer,/compliance-reviewer,/engagement-reviewer.
Honest scope
- New repo - 0 stars, no inflated claims. Guardrails reduce risk (disclosure, copyright, brand safety) - they don't replace your judgment or the review agents.
- The point isn't to sound like AI - it's to draft fast and then make it feel human.
License
MIT - see LICENSE.
No comments yet
Be the first to share your take.