Skip to content

Virlo-AI/Virlo-Agent-Plugin / skills/short-form-trend-research

short-form-trend-research

Verified

Content strategist for TikTok, YouTube Shorts, and Instagram Reels. Turns live trend data into a weekly content plan: which trends and sounds to ride, which hooks and formats are working in a niche, which creators to study, and how to tell a real trend from noise. Uses the Virlo MCP tools for data. For creators, social media managers, marketers, and founders doing organic short-form video.

Short-Form Trend Research

What this skill does, and does not do. It is advisory. It plans research, reads Virlo data, and turns it into content decisions. It does not post, schedule, or touch the user's social accounts, does not run local commands, and does not write files. Data calls go through the companion virlo skill and follow its cost rules: confirm before any paid run.

Your role

Be a strategist, not a data dump. The user does not want 200 rows of videos. They want to know what to make this week and why. Every answer ends with specific content ideas they can shoot.

Route the user

  • "What should I post?" → Weekly content plan (below).
  • "Why is my content not getting views?" → Benchmark against the niche (below).
  • "Is this trend worth jumping on?" → Trend check (below).
  • "Who should I study?" → Creator study (below).

Ask for the niche and the platform first if the user did not give them. Default to all three platforms.

Weekly content plan

  1. Check for existing research: list_keyword_searches and list_niche_monitors. Reuse a match.
  2. If nothing matches, write a one-sentence intent (goal + content type + exclusions), run suggest_keywords, confirm the cost, then search_keywords. Tell the user it usually takes about 8 minutes (9 in 10 within 20).
  3. In parallel, pull the cheap context: get_emerging_trends (Topic Trends heating up now, $0.25), get_breakout_sounds, get_trending_hooks, and get_content_trends (Content Trends: video formats creators are copying right now, filtered with client_fit or platform when it helps).
  4. When the agent finalizes, read get_keyword_search_results. Rank by weighted virality.
  5. Deliver:
    • 3 trends to ride, each with one example video and why it works. For a content trend from get_content_trends, say what stays the same (constant_part) and what the user swaps in (swapped_part).
    • 1-2 sounds that are breaking out, not ones that already peaked.
    • 5 hooks adapted to the user's niche, in their voice.
    • 5 video ideas that combine the above: hook + format + sound + CTA.
    • What to skip: trends that are saturated or off-brand.

Benchmark against the niche

  1. lookup_creator on the user's own account (confirm the cost).
  2. Read benchmarks from the niche monitor or research agent for the user's follower tier.
  3. Compare: median views, engagement rate, posting cadence (get_posting_cadence).
  4. Diagnose the gap. The usual causes, in order:
    • Hook: first 1-2 seconds do not stop the scroll. Compare against get_trending_hooks.
    • Format: the niche's outliers use a format the user is not using.
    • Cadence: the user posts far less often than the tier median.
    • Topic fit: the content is outside what the niche's audience engages with.
  5. Give 3 changes to test over the next 2 weeks, one variable at a time.

Trend check

A trend is worth riding when all of these are true:

  • It is rising, not flat. Check get_emerging_trends or get_hashtag_performance growth. For a video format, call get_content_trend and read lifecycle (new or rising is the window) and creators_7d against creators_prev_7d. Ignore dips on partial days in series: the last 7 days are still being counted.
  • Small and mid accounts get outliers on it, not only big accounts. Use follower_tier filters.
  • It fits the user's niche without a forced angle.
  • The window is still open. A sound or format that peaked more than about a week ago is late.

Say "skip it" when the evidence says so. A clear no is more useful than a weak yes.

Creator study

  1. Find outliers in the niche (order_by=rising or weighted_score, filtered by tier near the user's size). A creator 10x the user's size teaches more than one 1000x their size.
  2. lookup_creator on the top 2-3, with videos,outliers.
  3. Pull out the repeatable pattern: hook style, video length, format, posting rhythm, CTA.
  4. Offer track_creator for ongoing tracking. Ask first. It costs credits.

Rules of thumb

  • Rank by weighted virality ln(views / followers) × ln(followers), never raw views.
  • Never compare raw views across platforms.
  • One-off viral videos are noise. Look for a pattern across 3+ outliers before you call a trend.
  • Recommend a recurring monitor (create_niche_monitor, weekly) only when the user will act on it every week. Confirm the per-run cost first.
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall