Skip to content
SkillModel-led1,328 tokens in use

Outlier Post Finder

Finds posts, reels, shorts and tweets that beat an account's own median performance, explains why they worked and builds a hook swipe file.

6,856 stars on the repository MIT LicenseScrapeCreators/social-media-research-skills · skills/outlier-post-finder
npx skills add ScrapeCreators/social-media-research-skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing Outlier Post Finder costs

60tokens idle
1,328tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

shelldeclarednetworkdeclaredwrites filesdeclared

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Excellent
  • Frontmatter15/15
  • Structure25/25
  • Examples9/15
  • Output10/15
  • Size15/15
  • Resources10/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Needs attention
  • Development activity24
  • Maintenance8
  • Community40
  • Maturity24
  • Momentum40

Outlier Post Finder pulls recent public posts from a creator or brand, normalizes their metrics, computes a median baseline for each account and format, and labels posts as huge, strong or mild outliers by lift over that baseline.

It then fetches details, transcripts and optionally comments for the winners and reports repeatable patterns, hooks to reuse and content ideas, with source URLs and a confidence level.

What You Get

  • An outlier report with sample size, window, baseline and a confidence level
  • A ranked table of the biggest outliers with metric, lift and why each likely worked
  • Repeatable patterns with evidence, and hooks pulled verbatim from captions or transcripts
  • Content ideas based on the outliers, with source URLs

Common Use Cases

  • Finding a competitor’s best-performing Shorts or Reels
  • Building a swipe file of hooks from a creator’s standout posts
  • Checking which formats lift performance on a single account
  • Finding what to learn from a creator before planning content

What Makes It Unique

Lift against each account’s own median replaces a ranking by raw views, so a small account’s breakout post is not buried under a large brand’s average one. Platforms are scored separately, formats are split where possible, and the skill calls confidence low under 10 posts.

Under The Hood

The skill is one SKILL.md of instructions with no scripts. The agent calls the ScrapeCreators REST API with the SCRAPECREATORS_API_KEY environment variable, using feed endpoints for TikTok, Instagram, YouTube, Facebook, LinkedIn, X, Threads and Bluesky, then detail and transcript endpoints for winners. It is told to read the per-endpoint API docs when a parameter is unclear, which means web fetches. Its frontmatter allows the shell, file reading and writing, and web fetches, so the agent can run commands, call the API and save the report to a file.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers