skill-data-tracker · Social Data Tracking & Trend Analysis
A social data-tracking skill with three modes: A — recording daily snapshots of follower counts and engagement; B — analyzing follower growth trends, acceleration shifts, and milestone forecasts; C — tracking a single post's data from publish to decay, classifying it as flash-hit / steady-growth / long-tail.
It's the authoritative base for follower and time-series snapshots — snapshots only, never publish events. Growth rates, moving averages, and milestone projections are all computed deterministically by scripts, never by hand.
Example invocation: "Log today's Xiaohongshu follower count and show me the growth trend."
Full brief
Positioning
skill-data-tracker is the follower/time-series snapshot base of the attribution chain: it records social metric snapshots, analyzes follower growth trends, and tracks content lifecycles — driving operations decisions with time-series data. Time-series snapshots only; publish events belong to skill-publish-log.
Core capabilities
- Mode A — snapshot recording: logs daily followers, total likes, total posts, and per-post data; one snapshot per day per platform (same-day re-records overwrite); auto-computes delta vs. the last snapshot.
- Mode B — growth trends: per-point daily growth/rate, 7-day moving average, trend direction (accelerating / stable / decelerating), milestone forecast (at this pace, Y followers in X days), plus trend insights and recommendations.
- Mode C — content lifecycle: rebuilds a post's time series across snapshots (Day 0/1/3/7/14/30), returns daily increments, peak day, half-life, and a flash-hit / steady-growth / long-tail classification with follow-up strategy implications.
- Growth-attribution export:
export-followersaggregates the follower ledger for skill-publish-analytics — no hand-maintained second ledger. - Deterministic computation: snapshot I/O, growth rates, moving averages, milestone projection, and lifecycle classification all run in scripts; the LLM only fills parameters, interprets results, and writes growth advice.
- Hard rules: never modify or delete existing snapshots (same-day overwrite excepted); one snapshot per day; growth rates need ≥ 3 data points (fewer → warning only); milestone forecasts ≤ 30 days out; data sources transparent — no fabricated metrics.
Workflow
- Gather metrics (from the profile or user input; ask once for anything missing)
- A: script
snapshot→ show delta and save path - B: script
trend→ LLM writes trend insights and audience-relevant advice - C: script
lifecyclerebuilds across snapshots → LLM writes follow-up content strategy per type
Inputs & outputs
| Input | Notes |
|---|---|
| mode | Required — record / growth / lifecycle |
| platform | Required for A — platform name |
| followers | Required for A — current follower count |
| total_likes / total_posts | Optional for A |
| post_snapshots | Optional for A — per-post recent data |
| post_title | Required for C — the post to track |
| time_range | Optional for B — analysis window (default last 30 days) |
Output: data snapshot report / growth trend analysis (with milestone forecast) / content lifecycle analysis (with type classification).
Boundaries with adjacent skills
- skill-publish-log: the publish-event base (title / link / type / source skill); this base stores follower/engagement time-series snapshots — no double-recording.
- skill-publish-analytics / skill-social-performance-review: consume these snapshots for growth attribution and period comparisons (read-only).
Fit
- Daily follower/engagement check-ins
- Judging whether growth is accelerating or decelerating
- Forecasting when the next follower milestone lands
- Classifying a post as flash-hit vs. long-tail and planning what follows
Before you start
- Metrics come from user input or snapshot files — report them honestly, never fabricate.
- Trend analysis needs at least 3 data points; fewer yields a warning only.
- Max one snapshot per platform per day; existing snapshots are read-only.
- Snapshots are isolated by profile/platform; without a profile, the
defaultdirectory is used.
skill-data-tracker is part of the Aiglade Skill library. Invoke it from the Aiglade chat box in plain language.