Influencer Complaint Rescue

SKILL: Influencer Complaint Rescue
A negative post from a 50k-follower account does more damage than a hundred from nobodies. This skill finds angry posts about your brand across Instagram, X, and TikTok, ranks the authors by reach, and pulls their public contact (email, bio link, or website) so support can resolve it privately before it goes viral.

You are running this skill on API Direct via its MCP tools. Execute the steps below yourself by calling the named tools in order — values in <angle brackets> come from a previous step. Then deliver the result described at the end.

INPUTS:
  - brand (required): <brand — ASK THE USER>
      Brand name to monitor for influential negative posts (no @ or # prefix)
  - platforms (optional): (optional — e.g. twitter, tiktok)
      Comma-separated list of platforms to run — options: instagram, twitter, tiktok. Omit to run all of them; name specific platforms to limit the run.
  - min_followers (optional): (optional — e.g. 10000)
      Optional minimum follower count. Complainers below this reach are dropped so the team focuses on accounts that can actually do damage.
  - region (optional): (optional — e.g. US)
      Optional TikTok region code to localize the complaint search to a single market. Applies only to the TikTok step.
  - depth (optional): 2 (default — override if you like)
      Optional number of result pages to pull per search platform (Instagram, X, and TikTok). Raise to widen the net on a noisy brand.

PLATFORM SELECTION: some steps are gated with "Only run if {platforms} includes X". If the user supplied a `platforms` value, run only the steps whose platform is listed. If the user did NOT supply `platforms`, run ALL steps — every platform the skill supports.

STEPS:
  1. Tool `search_instagram` — search_instagram(query='{brand} disappointed OR refund OR scam OR broken OR "worst" OR "never again"', get_sentiment=true, pages=2)
     Core complaint sweep. One combined query covers both mild and serious complaint language. Keep only posts with polarity==negative or dominant_emotion in anger/sadness. 2 defaults to 1 page if not set.
  2. Tool `instagram_user_profile` — instagram_user_profile(username=<top_negative_ig_author>)
     Fetch followers, category, public_email and external_url to rank complainers by reach and find a private contact channel. If {min_followers} is set, keep only authors at or above that follower count.
  3. Tool `search_twitter` — search_twitter(query='{brand} disappointed OR refund OR scam OR "worst" OR "never again"', get_sentiment=true, sort_by='most_recent', pages=2)
     Only run if {platforms} includes twitter: catch fresh angry tweets about the brand. Keep only polarity==negative posts. 2 defaults to 1 page if not set.
  4. Tool `twitter_user_profile` — twitter_user_profile(username=<top_negative_tweet_author>)
     Only run if {platforms} includes twitter: pull follower count and bio/website link to rank the loudest complainer by reach and find a contact channel. Apply the {min_followers} threshold if set.
  5. Tool `search_tiktok` — search_tiktok(query='{brand} disappointed OR refund OR scam OR "worst"', get_sentiment=true, sort_by='most_recent', publish_time=30, pages=2)
     Only run if {platforms} includes tiktok: surface complaint videos from the last 30 days. Keep only polarity==negative. Pass region={region} only if set, otherwise omit for a global search. 2 defaults to 1 page if not set.
  6. Tool `search_tiktok_users` — search_tiktok_users(query=<top_negative_tiktok_author>)
     Only run if {platforms} includes tiktok: look up the complainer's TikTok profile for follower count and bio contact. Apply the {min_followers} threshold if set.

DELIVER: A cross-platform triage list of negative posts (Instagram by default, plus optional X and TikTok) ranked by author reach, each paired with the complainer's public email or bio contact channel so support can reach out and resolve privately before it spreads.

Note: each underlying tool call is billed at its normal endpoint price; get_sentiment adds a small per-page surcharge. Page through results as needed but stop once you have enough to deliver the outcome.