Trend Scout

Scan current industry trends and convert them into brand-aligned content angles. Uses Perplexity for trend discovery. Output feeds post-writer, series-planner, and image-creator.

Purpose

Trend Scout is the research bridge between foundation skills (visual-extractor, framework-builder) and content skills (post-writer, series-planner). It answers:

What topics are trending in our category RIGHT NOW?
How can our brand authentically connect to each trend?
Which trends are urgent vs. evergreen?
What platform is best for each angle?

The output — creative-memory/trend-angles.md — becomes the content direction layer that post-writer and series-planner read to choose what to write about.

Modes

Mode	When to Use	Behavior
Full Scan	No `trend-angles.md` exists, or it's empty/scaffold only	Run all 5 steps from scratch
Refresh	`trend-angles.md` already has data	Check `creative-log.md` for last scan date → update stale trends, preserve evergreen angles

Auto-Load Protocol

On every invocation, BEFORE any research:

Check creative-memory/ directory → If missing, create it
Read brand-memory/voice-profile.md (read-only) → Brand tone for angle framing
Read brand-memory/positioning.md (read-only) → Core positioning = the lens for trend-to-angle conversion
Read ALL .md files in creative-memory/ (except README.md) → Existing guidelines, frameworks, log
Optional: Read research-memory/market-landscape.md (read-only) → Market trends, seasonality context
If trend-angles.md has data AND creative-log.md shows a previous trend-scout run → suggest Refresh mode

Input Gathering

Collect from the user conversationally. Do NOT dump a form — ask naturally.

Field	Required	Description
Brand category / Industry	YES	What space does the brand operate in?
Target audience	YES	Who are we creating content for?
Focus keywords	Optional	Specific topics or angles to explore
Platform priority	Optional	Which platforms matter most?
Language	Optional	결과물 작성 언어 (default: English)

If brand-memory/ exists, pre-fill category and audience from positioning.md — ask user to confirm or correct.

If this is a Refresh, show the current trend-angles.md summary and ask: "Any specific direction you want to explore, or should I scan broadly for what's new?"

Process

Step 1: Context Load + Input Collection

Goal: Establish brand context and determine scan scope.

Execute Auto-Load Protocol (above)
Extract from positioning.md: brand category, core positioning, target audience, key differentiators
If positioning.md missing → ask user directly for category + positioning
Determine mode: Full Scan (no existing data) or Refresh (existing data found)
For Refresh: parse existing trend-angles.md to identify current trends and their dates

Output: Clear brand context + mode decision + scope of search.

Step 2: Trend Discovery (Perplexity)

Goal: Identify 3-5 current trends relevant to the brand's category and audience.

Tool: perplexity_ask

Query pattern (Full Scan):

What are the top 3-5 trending topics in the [category] space right now that brands can leverage for social media content?

For each trend:
- Name the trend in 3-5 words
- Explain why it's trending NOW (what triggered it)
- Describe who cares about this (audience overlap with [target audience])
- Estimate how long this trend will remain relevant (days, weeks, months)
- Name 1-2 platforms where this trend is most active

Focus on trends that a [brand type] brand could authentically participate in.

Query pattern (Refresh):

Here are the trends we identified previously for [category]:
[list existing trends from trend-angles.md]

Which of these are still relevant? What new trends have emerged since [last scan date]?
For any new trends, provide: name, trigger, audience relevance, estimated lifespan, best platforms.

Parameters:

search_recency_filter: "week" (default) or "month" (if fewer than 3 results)
search_context_size: "medium"

Source verification: After getting trends, use perplexity_search to find 1-2 credible source URLs per trend (industry publications, platform data, research reports).

Fallback: If fewer than 3 trends found with "week" filter, retry with "month".

Output: 3-5 validated trends with sources.

Step 3: Brand Angle Conversion

Goal: Transform each raw trend into a brand-specific content angle.

For each trend from Step 2, apply the positioning lens:

Connection point: How does this trend relate to our brand's core positioning?
Audience hook: What aspect of this trend would our target audience care about most?
Brand voice fit: How would we talk about this in our brand's tone? (reference voice-profile.md)
Angle statement: Write a one-sentence content direction that a copywriter could run with

Urgency tagging — assign one label per trend:

Tag	Meaning	Action Window
🔴 High	Time-sensitive, ride it NOW	Within 1 week
🟡 Medium	Relevant window open	2-4 weeks
🟢 Low	Slow-burn, long shelf life	1-2 months

Platform mapping — for each angle, identify the 1-2 best platforms based on:

Where the trend is most active
Where the brand's audience is most engaged
Which format best suits the angle (long-form → LinkedIn, visual → Instagram, etc.)

Output: 3-5 trend angles with urgency tags and platform recommendations.

Step 4: Evergreen Angles

Goal: Add 2-3 timeless content angles derived from brand positioning.

Evergreen angles are NOT trend-dependent. They come from:

The brand's core value proposition
Recurring audience pain points (from positioning.md or research-memory/)
Category fundamentals that never go out of style

For each evergreen angle, define:

Theme: The broad topic area
Brand angle: The brand's unique perspective on this theme
Best platforms: Where this type of content performs best
Reuse cadence: How often this angle can be recycled (weekly, monthly, quarterly)

For Refresh mode: Keep existing evergreen angles intact. Only add new ones if a clear gap is identified. Never delete existing evergreen angles — they are stable by design.

Output: 2-3 evergreen angles appended after current trends.

Step 5: Save & Log

Goal: Write all findings to creative-memory/trend-angles.md and log the execution.

5a. Write trend-angles.md

Use the exact schema below:

# Trend Angles
> Last updated: [YYYY-MM-DD]
> Source skill: trend-scout
> Mode: [Full Scan / Refresh]

## Current Trends
| # | Trend | Brand Angle | Urgency | Best Platforms |
|---|-------|-------------|---------|----------------|
| 1 | [trend name] | [brand-specific angle statement] | 🔴/🟡/🟢 | [platform list] |
| 2 | ... | ... | ... | ... |

## Evergreen Angles
| # | Theme | Brand Angle | Best Platforms | Reuse Cadence |
|---|-------|-------------|----------------|---------------|
| 1 | [theme] | [brand-specific angle] | [platforms] | [weekly/monthly/quarterly] |

## Sources
| Trend | Source URL | Retrieved |
|-------|-----------|-----------|
| [trend name] | [URL] | [YYYY-MM-DD] |

For Refresh mode: Do NOT overwrite the entire file. Update the Current Trends table (replace stale trends, add new ones). Keep Evergreen Angles section intact unless explicitly adding new ones. Append > Refreshed: [date] below the header.

Language rule: 섹션 헤더와 테이블 컬럼명은 영어로 유지합니다. 본문, 셀 값, 설명, 분석 텍스트는 사용자가 지정한 언어로 작성합니다. 언어가 지정되지 않으면 English로 작성합니다.

5b. Update creative-log.md

Append one row:

| [YYYY-MM-DD] | trend-scout | Full Scan / Refresh | [brief summary: e.g., "Found 4 trends: AI content tools, micro-communities, ..."] | Perplexity |

Perplexity MCP Tool Guide

Tool	When to Use	This Skill
`perplexity_ask`	Fast Q&A, current trends, factual	Step 2 main discovery
`perplexity_search`	Find specific URLs, verify sources	Step 2 source verification

Common parameters:

search_recency_filter: "week" for trends (default), "month" if results are sparse
search_context_size: "medium" — trends need current data but not deep analysis

Query best practices:

Always include the exact brand category in every query
Specify "for social media content" to get actionable trends (not just news)
Ask for estimated trend lifespan — crucial for urgency tagging
One discovery query + one verification query per run (minimize API calls)
For Refresh: provide existing trend list so Perplexity can identify what's new vs. stale
Language: 사용자가 English 외 언어를 지정한 경우, 모든 query 끝에 "Respond in [language]."를 추가

Memory Access Rules

Memory	Permission	Purpose
`brand-memory/positioning.md`	Read-only	Core lens for trend → angle conversion
`brand-memory/voice-profile.md`	Read-only	Tone consistency for angle framing
`research-memory/market-landscape.md`	Read-only	Market context, seasonality awareness
`creative-memory/trend-angles.md`	Read & Write	Primary output file
`creative-memory/creative-log.md`	Write (append)	Execution history

NEVER write to brand-memory/ or research-memory/. These are owned by vibe-mkt and mkt-study respectively.

Error Handling

Scenario	Response
Perplexity MCP unavailable	Inform user → generate evergreen angles only from positioning.md
`positioning.md` not found	Ask user directly for brand category + positioning → proceed
Fewer than 3 trends found	Expand: widen recency filter from `"week"` → `"month"`, broaden category keywords
Existing `trend-angles.md` unparseable	Suggest Full Scan mode to rebuild from scratch
Source verification returns no URLs	Keep the trend but mark source as "Perplexity synthesis (no direct URL)"

Quality Checklist

Before saving, verify:

At least 3 current trends identified with brand angles
Every trend has an urgency tag (🔴/🟡/🟢) and platform recommendation
At least 2 evergreen angles included
All angles are filtered through brand positioning (not generic trend descriptions)
Sources section has at least 1 URL per trend (or explicit "no direct URL" note)
creative-log.md updated with execution record
brand-memory/ was read-only (no writes)
For Refresh: existing evergreen angles preserved, only stale trends replaced

What This Skill Does NOT Do

Write social posts → Use post-writer (consumes this output)
Plan campaigns → Use series-planner (consumes this output)
Generate images → Use image-creator (downstream)
Deep market research → Use market-scanner from mkt-study plugin
Competitor analysis → Use competitor-analyzer from mkt-study plugin

Trend Scout stays focused on discovering trends and converting them into brand content angles — nothing more.