Library

# Meta AI Prompt Library

Every prompt here is built for **meta.ai (Muse Spark)** and only uses **verified** capabilities:
reading public Instagram, Threads, and Facebook content, semantic search over posts since Jan 2025,
creator discovery, trend/comment reading, and brand-mention search.

## Template standard (every entry follows this)

Built on the prompting practices in `working-with-claude` (section 3), applied as model-general
engineering (Muse Spark's exact preferences are unverified, so the self-test is the backstop):

1. **Directive first line.** An action verb stating the task and hinting at the output. Biggest lever.
2. **Decomposed steps**, numbered, capability check always step 1.
3. **Interpolated inputs wrapped in a named tag** (`<brief>`, `<inputs>`) so meta.ai knows what is data.
4. **Explicit `<output_format>`**: exact columns, sort order, and how to mark uncertainty.
5. **A one-shot `<example>`** showing an ideal row (for table outputs).
6. **Honesty rules**: never fabricate a handle, number, or link; label estimates; cite real links.
7. **No em dashes**, ever (user rule).

Variables are `{{like_this}}`. The skill fills them and updates the example if a variable changes its shape.

---

## A1: Micro-influencer / ambassador finder
**Use when:** finding small creators to recruit for an ambassador program or partnership.
**Fills:** `{{niche}}`, `{{location}}` (optional), `{{follower_min}}`, `{{follower_max}}`, `{{vibe}}`, `{{count}}`