Add category-top-100-change-tracker skill

[norightsforlis]

Summary

Adds the category-top-100-change-tracker skill: Weekly change tracker for a category's top ~100 sellers, adding a per-seller week-over-week movement column and the biggest movers.

Lives in skills/category-top-100-change-tracker/SKILL.md and is registered in the README skills table. Public behavior and user-visible workflow only — no internal service names, schema, or implementation details.

Public-Safety Checklist

• No credentials, tokens, keys, or secrets.
• No internal filesystem paths.
• No private Slack, Jira, Notion, admin, or service URLs.
• No private service names or implementation details.
• No customer data or private product data.
• Skill directory name matches SKILL.md frontmatter name.

🤖 Generated with Claude Code

[profunctor]

KISS review: two behavior issues before merge.

1. The skill promises weekly run-to-run tracking, but it is read-only and defines no durable baseline/input contract. A fresh scheduled/session run cannot read “the previous run’s recorded table”, so it will restart or infer movement without data.
2. It delegates core schema/routing to top-sellers and seller-overview, but those skills are not on main or in this PR. Please inline the needed columns/routing or include the dependency.

[profunctor]

Review — category-top-100-change-tracker

Summary: A weekly change-tracker over the top ~100 sellers in one Mercado Livre (Brasil) category, ranked by estimated average revenue. Each run diffs against the previous run's recorded table (rank rose/fell/unchanged/new), flags the most notable metric change, lists biggest risers/fallers, and re-emits the full table as the next baseline. Read-only, pt-BR default.

Findings:

• minor/major — Cross-references depend on unmerged sibling PRs. The skill points to "the top-sellers skill" (Add top-sellers-in-category skill #16) and "the seller-overview tracker" (Add seller-overview-tracker skill #5), neither of which exists on main. Confirm merge order and reference them by their canonical final slugs rather than prose nicknames.
• minor — Baseline persistence is underspecified. The value prop is run-to-run diffing, but it only says "read the previous run's recorded table" / "each run is its own session" — no mechanism for a fresh session to obtain the prior table. In practice it may always behave as a first run.
• nit — Comparable metrics ("sales trend flipped sign", "cancellation rate moved a point") are used in the workflow but not defined in Scope; they're deferred to "same fields as the top-sellers skill," coupling correctness to Add top-sellers-in-category skill #16.
• nit — The delta-header rule ("Variação", never a bare Δ) is stated twice; consider consolidating.

Structure, frontmatter, EN+pt-BR triggers, and README row all match conventions. No secrets or unsafe operations.

[norightsforlis]

Closing in favor of #16 (top-sellers-in-category). We're consolidating the top-sellers leaderboard into a single skill rather than shipping a separate top-100 change-tracker. Thanks for the review, @profunctor.