| name | caption-tone |
| description | Write platform-ready GenX captions and descriptions with brand guardrails. Use for caption, description, hook, LinkedIn post, Instagram caption, YouTube Shorts description. |
| category | genx-marketing |
| version | 2.1.0 |
Caption Tone
Goal: produce concise, brand-safe captions. Do not invent facts or sources.
Inputs
Ask only if missing:
- audience: B2C, B2B, or Both
- platform: Instagram, LinkedIn, YouTube Shorts, email, other
- tone: professional, casual, edgy, warm, wry, reflective, analytical/data-driven, bold, platform-native
- source proof: URL, quote, result, or user-provided context
Default: Both / Faceless Instagram / professional.
Brand Rules
Load:
memory/brand-rules.mdconfig/brand-voice.yaml
Avoid blacklisted phrases in GenX-written copy. If a quoted source contains them, flag the quote and write around it.
B2C: real science + real soul, transformation through proof, not woo.
B2B: results, delivery, operators, no advice-merchant language.
Flow
input -> audience/platform/tone -> draft variants -> brand_gate -> skills/security-rail/SKILL.md -> templates/caption-pack.md
Output
Use templates/caption-pack.md for multi-caption work.
For quick caption requests, return:
## Decision
Recommended caption: ...
Variants:
1. ...
2. ...
3. ...
Brand Gate: pass/revise/shelve
Security Rail: APPROVE/REVISE/SHELVE/NEEDS HUMAN REVIEW
Next step: ...
Platform Defaults
- Instagram: short hook, 1 idea, readable line breaks, light CTA.
- LinkedIn: operator context, proof, practical takeaway.
- YouTube Shorts: direct description, source/context, no fluff.
- Email: use
skills/email-hook/SKILL.md instead.
Gate
Run goblin_recon.tools.brand_gate when feasible. Run skills/security-rail/SKILL.md before final delivery. If factual, competitor, performance, health, legal, or sensitive claims lack proof, mark NEEDS HUMAN REVIEW or REVISE.
