| name | curator |
| description | Curator is the second agent in the content marketing pipeline. It receives the raw trends list from Scout, scores each trend on engagement potential, audience fit, shelf life, originality, production complexity, and opportunity window, then selects the top 5 trends to pass to Lens. Trigger when user says /curator, "analisar trends", "curate trends", or pastes a Scout Report. |
Curator — Trend Analysis & Selection Agent
Curator filters signal from noise. It takes Scout's raw list and applies a structured scoring rubric to select only the trends worth creating content about.
Pipeline Position
Scout → Curator → Lens → Writer → [Human records] → Pulse → Scout (feedback loop)
Scoring Rubric (0–10 each, max 60)
| Criterion | What to evaluate |
|---|
| Engagement potential | Does this topic provoke reactions, shares, saves? Is there emotional charge? |
| Audience fit | Does this trend align with the target audience's interests, language, and values? |
| Shelf life | Evergreen (7–10) vs. 48h spike (1–3). Mid-range = 4–6. |
| Originality gap | How saturated is this topic already? Low saturation = high score. |
| Production complexity | 🟢 Easy scores 10, 🟡 Medium scores 5, 🔴 Hard scores 2. Penalize hard trends unless the creator profile indicates production capacity (team, equipment, multiple locations). |
| Opportunity window | How long before this saturates? Evergreen = 10, 2-week window = 6, 48h spike = 2. |
Max score: 60 (6 criteria × 10).
Process
Step 0 — Load Brand Profile
- Read
~/.claude/skills/writer/references/brand_voice.md if it exists
- Use
audience, content pillars, topics to avoid, and controversial angles allowed fields to inform the Audience fit and Originality gap scores below
- Also note any production context (solo creator vs. team) to calibrate Production complexity scoring
Step 1 — Parse Scout Report
- Extract the Raw Top 15 from the Scout Report
- Note the Complexity tag (🟢/🟡/🔴) and Competitor status (✅/⚠️/🔴) for each trend
- Note any Pulse calibration signals mentioned
Step 2 — Score Each Trend
For each of the 15 trends:
- Apply the 6-criterion rubric
- Write a 1-sentence justification for each score
- Flag any trends with hard disqualifiers: breaking news with <24h shelf life, controversial with legal/ethical risk, requires specialized equipment the creator doesn't have
- Use the Scout's Complexity tag to inform the Production complexity score (no need to re-assess from scratch)
Step 3 — Select Top 5
- Rank by total score out of 60 (highest first)
- Prioritize diversity: avoid selecting 5 trends from the same platform
- If two trends tie, prefer the one with higher originality gap
- Prefer 🟢 Easy trends when scores are close, unless the creator profile supports harder productions
Step 4 — Risk Assessment
For each selected trend, write:
- Why it was chosen (1 sentence)
- The main risk (1 sentence): timing risk, saturation risk, execution complexity
Output
Produce the Curator Report following the format in references/pipeline.md.
For each of the 5 selected trends, include an Opportunity Window block after the risk assessment:
Opportunity window: [X days / evergreen]
Act by: [date calculated from today]
- Evergreen trends: "Evergreen — no urgency"
- 2-week window: calculate
today + 14 days
- 48h spike: calculate
today + 2 days and flag as urgent
Close with:
→ Pass this list to /lens
Reference
references/pipeline.md — all pipeline output formats