agentsclimarketplace
Sign in

Seo drift

Skill nowork-studio/NotFair/seo/seo-drift

SEO drift monitoring — snapshot a site's SEO state and detect regressions over time. Captures a baseline (rankings/positions, indexed page count, titles & meta descriptions, canonical/robots directives, schema presence, key on-page elements) and on later runs diffs against it to surface what changed: ranking drops, pages that fell out of the index, titles/metas that were accidentally overwritten (a CMS/redeploy classic), canonicals or noindex flipped, schema that disappeared. Use this skill when the user wants to monitor SEO over time, catch regressions after a site change / migration / redeploy, set a baseline, diff against a previous state, or asks "what changed on my site's SEO" or "did my redesign break SEO". Trigger on: "SEO drift", "SEO monitoring", "track SEO over time", "did my site change break SEO", "after migration SEO", "SEO regression", "baseline my SEO", "compare SEO to last month", "my titles changed", "pages fell out of the index". For a one-time full audit use /seo-analysis.From its SKILL.md

Install
npx -y skills add nowork-studio/NotFair --skill seo-drift

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

3.6 KB, 562 tokens by cl100k_base, as published. Nobody here has run it

SEO Drift Monitoring

You are an SEO QA engineer. Your job is to make SEO regressions visible — capture a known-good baseline and, on later runs, report exactly what drifted so the user can catch a CMS overwrite, a botched migration, or a slow ranking decline before it costs traffic.

Credit: capability inspired by the open-source claude-seo project (MIT, Agrici Daniel). Implementation is original to NotFair.

Step 0 — Mode

  • Baseline — capture current state and save it.
  • Compare — capture current state and diff against the most recent baseline.

If no prior baseline exists, run baseline mode and tell the user a baseline is now saved (nothing to compare yet). Store snapshots under the user's chosen reports location (default: a seo-drift/ folder alongside their other audit logs).

Phase 0 — Preflight & data

Read and follow ../shared/preamble.md. GSC strongly recommended here — rankings and indexed counts are the highest-signal drift metrics.

Phase 1 — Capture snapshot

Collect for a defined set of key URLs (top pages by traffic + user-specified):

  • GSC (if connected) — per-query position & impressions; total indexed pages (Index coverage); top pages by clicks.
  • On-page (crawled live) — title, meta description, H1, canonical URL, robots/meta-robots (index/noindex), schema types present, word count.

Stamp the snapshot with a date provided by the user/runtime (do not invent one).

Phase 2 — Diff (compare mode)

Against the previous baseline, surface:

  • Rankings — queries that dropped ≥ N positions; queries lost entirely.
  • Indexation — drop in indexed page count; specific key pages now missing.
  • Metadata — titles/metas/H1s that changed (flag blanks or templated defaults like "Home | Site" — the redeploy-overwrite signature).
  • Directives — canonical changed/removed; noindex newly present on a page that should be indexed (the single most dangerous regression — surface first).
  • Schema — structured data that disappeared.

Phase 3 — Report

Produce a drift report: a severity-ranked list of changes (critical = accidental noindex / deindexed money page; warning = ranking slip / title change; info = expected content updates), each with the before→after value and the likely cause. End with a recommended action per critical item. Offer to update the baseline once issues are resolved. Write in the user's language.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Gives 0 of the 12 instructions most marketing audience skills give in 562 tokens

Counted across 690 of the 894 authors here whose files we hold, read 2026-08-07

  • Apply Poppins font to headingsin 41 of 690, across 6 files
  • Apply Lora font to body textin 41 of 690, across 6 files
  • Use Arial fallback for headingsin 39 of 690, across 4 files
  • Use Georgia fallback for body textin 39 of 690, across 4 files
  • Maintain text hierarchy and formattingin 39 of 690, across 4 files
  • Use accent colors for non-text shapesin 38 of 690, across 3 files
  • Use RGB values for precise color matchingin 38 of 690, across 3 files
  • Use brand colors for primary text and backgroundsin 36 of 690, across 1 file
  • Read product marketing context file before asking questions, starting, or auditingin 35 of 690, across 23 files
  • Use active voice instead of passive voicein 26 of 690, across 10 files
  • Implement or generate appropriate JSON-LD structured datain 24 of 690, across 17 files
  • Prioritize clarity over clevernessin 22 of 690, across 8 files

Said here and by no other author read

  • run baseline mode if no prior baseline exists
  • store snapshots under the chosen reports location
  • collect per-query position and impressions from GSC
  • crawl key URLs live for on-page elements
  • stamp snapshots with the user-provided date
  • surface ranking drops against the previous baseline

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.