social-trend-monitor

name	social-trend-monitor
description	Produces dated trend briefs with source URLs, confidence and recency notes, audience fit scores, and risk flags—feeds social-media-manager and tiktok-video-ads-creator. No posting; research-only artifact under workspace/drafts/social/trends/.
metadata	{"clawdbot":{"emoji":"📡"},"openclaw":{"emoji":"📡"}}

social-trend-monitor

Trend intelligence as a standalone dated report, parallel to how qf-course-researcher produces a dated research page—each run is self-contained and citable.

Scope boundary (do not duplicate course research)

• qf-course-researcher / QuizFactor course work = category, curriculum, competitor courses, evidence for product decisions.
• social-trend-monitor = what is surfacing on social platforms now — trending hashtags, topics, formats, memes, breakout sounds (short-video), and conversation spikes on X, Instagram, TikTok, LinkedIn, etc., scoped to the human’s niche + geo.
• Goal: give social-media-manager and creatives tags and angles to ride the feed, not to redo deep course-market research.

Prerequisites

• Web search / news tools when available; else INFERENCE-ONLY banner at top.
• Human: niche, geo, languages, competitors to watch, topics to avoid.
• Output:

  workspace/drafts/social/trends/<YYYY-MM-DD>-<niche-slug>.md
  

Credentials & API (qf-style)

• Web search / browse: Same expectation as qf-course-researcher — gateway must expose search tools, or mark output INFERENCE-ONLY.
• Optional paid SERP: If configured, document env vars in workspace/INTEGRATIONS.md (see auto-research-agent).

High-level Workflow

1. Query plan (queries.md)

   • List 5–10 search queries you will run (or would run); date of run.

2. Raw hits table (raw-hits.md)

   • Columns: Topic | Source URL | Publisher | Published date | Snippet | Angle.

3. Scoring (scored-topics.md)

   • For each consolidated topic (5–15 rows):
     • Recency (hours/days since peak).
     • Relevance to niche (0–10).
     • Brand fit (0–10) per USER.md.
     • Risk (misinfo, controversy) — low/med/high.

4. Angles (angles.md)

   • 3 content angles per high-fit topic: hook idea, format suggestion, avoid notes.

5. Declined topics (declined.md)

   • Trends rejected and why (off-brand, high risk, stale).

6. Handoff

   • “Consume by: social-media-manager (calendar), tiktok-video-ads-creator (hooks).”
   • Before hype-engine publish: social-media-manager may copy 0–2 vetted hashtags/topics from this file into X and LinkedIn feed payloads only (see that skill’s Trending hashtags section).

7. Scheduling

   • Daily slug daily-<niche>; do not overwrite—new date file each run.

Outputs (required)

• trends-<YYYY-MM-DD>-<niche-slug>.md as master (can embed or link child sections).

Agent Checklist

• INFERENCE-ONLY if no live search.
• Published dates on raw hits where visible; unknown marked unknown.
• No “viral” claim without describing evidence (velocity, volume).
• Risk column populated; high-risk not recommended without human OK.
• Declined list non-empty when many hits were filtered (shows rigor).
• File path returned to user and parent skills.