seo-performance
Skill6 starsChanged 3 months agoApache-2.0View on GitHub
Use when establishing or comparing conventional SEO performance from authorized Google Search Console, Bing Webmaster Tools, analytics, server-log, indexation, or Core Web Vitals/RUM evidence, or when designing and evaluating a controlled SEO experiment; preserve dimensions and data-quality limits, keep metric families separate, and never turn observational movement into a causal or composite SEO score.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- 01clauding/claude-seo-skilltheir own version, first seen Mar 13, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- 01clauding/claude-seo-skill★ 4same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository27 other files it gives its agents.
- llms.txtllms.txt
- ai-search-researchSkill
- ai-visibility-monitorSkill
- optimise-seoSkill
- seo-action-planSkill
- seo-aeoSkill
- seo-agenticSkill
- seo-architectureSkill
- seo-auditSkill
- seo-authoritySkill
- seo-commerceSkill
- seo-competitor-pagesSkill
- seo-contentSkill
- seo-geoSkill
- seo-hreflangSkill
- seo-imagesSkill
- seo-localSkill
- seo-news-discoverSkill
- seo-pageSkill
- seo-planSkill
- seo-programmaticSkill
- seo-researchSkill
- seo-schemaSkill
- seo-sitemapSkill
- seoSkill
- seo-technicalSkill
- seo-videoSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 10
Read markdown-delivery-contract.md for every task.
- Distinctive: line 17
Read performance-measurement-protocol.md before normalizing, comparing, or interpreting data.
- Very specific: line 32
When the input is a supported Google Search Console, Bing Webmaster Tools, or organic GA4 CSV, use python …
- Very specific: line 38
Create seo-performance-run.json.
- Distinctive: line 51
Search Console and Bing metrics are platform-defined observations.
- Distinctive: line 53
Do not invent missing exports, reconstruct hidden queries, bypass property access, or claim ranking, traffic, revenue, or conversion …
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
SEO Performance
Markdown delivery contract
Read markdown-delivery-contract.md for every task. Write a substantive .md document for every material work step and register its direct destination in the work index.md. For public-site work, apply the owner-required Markdown discovery profile across the full eligible inventory; preserve native standards and immutable evidence formats.
Lane outputs: performance-baseline.md; performance-comparison.md; provider-receipt.md (combine small related steps with stable section links). Keep query-to-canonical ownership, conventional search metrics and public Markdown delivery metrics distinct. Render validated receipts/runs as readable Markdown while retaining immutable originals and privacy. Provider acceptance is not indexing or AI consumption.
Build a hash-pinned seo-performance-run.json from authorized first-party exports. Measure conventional search behavior, indexation, user experience, and organic outcomes without blending them into AI visibility metrics or implying that an intervention caused movement.
Read performance-measurement-protocol.md before normalizing, comparing, or interpreting data. For a controlled SEO experiment, also read controlled-seo-experiment-protocol.md before assigning cohorts or declaring an outcome. For an explicitly authorized Search Console, Bing Webmaster Tools, or IndexNow operation, read provider-operations-protocol.md immediately before touching the property.
Required inputs
- Require the exact property, source, time window, timezone, dimensions, export method, and collection time.
- Preserve the provider export, then create the versioned normalized JSON source envelope used by the run and bind that envelope with SHA-256. The run's rows, indexation observations, and Web Vitals must match it exactly.
- Record query anonymization, row limits, sampling, canonical aggregation, and missing dimensions explicitly.
- Require the prior hash-pinned run for comparisons. Never compare unlike properties, sources, dimensions, or window durations without marking them non-comparable.
Procedure
- Define the decision and the smallest useful metric family: search performance, indexation, field experience, organic sessions, conversion, or a bounded combination. For broad organic growth, preserve separate cohorts for nearer-term opportunities and the approved high-volume head-term leadership families; a currently absent head term remains a measured zero/unknown visibility target rather than disappearing from the baseline.
- Capture the authorized provider export without changing its rows. Redact secrets, personal data, account identifiers, and sensitive query data where required; record the resulting limitation. Create a normalized JSON source envelope with source/property/window/dimensions plus the exact rows, indexation, and Web Vitals used by the run; hash-pin it and retain the provider export beside it when permitted.
When the input is a supported Google Search Console, Bing Webmaster Tools, or organic GA4 CSV, use
python "<suite-root>/scripts/import_seo_exports.py" <adapter> ...and follow<suite-root>/docs/DATA-IMPORT-ADAPTERS.md. Inspect the normalized envelope before using it; an imported envelope is evidence input, not a validated Performance Run. - Normalize only declared dimensions. Keep query, page, country, device, search appearance, and date distinct. Preserve platform aggregation rules.
- Recompute CTR from clicks and impressions. Keep clicks, impressions, CTR, average position, sessions, conversions, revenue, indexation, and Web Vitals as separate measures. When an eligible property exposes Google Search Console's dedicated generative-AI performance preview or Bing's AI Performance report, preserve its availability, preview status, platform definitions, and supported dimensions as a separate first-party metric view. Do not infer unavailable queries or merge provider-reported visibility with observed AI answers.
- For a baseline, record no prior run or drift. For a comparison, verify prior identity, hash, chronology, property, source, dimensions, and equal window duration before calculating descriptive deltas.
- Describe movement as observational association. Record confounders such as demand, seasonality, reporting changes, access changes, migrations, campaigns, or incomplete query data.
- Create
seo-performance-run.json. Let<suite-root>mean${CLAUDE_PLUGIN_ROOT}in Claude Code. In Codex, read.seo-suite-runtime.jsonbeside thisSKILL.mdwhen present and use itssuite_rootvalue; otherwise use the absolute repository checkout, then runpython "<suite-root>/scripts/validate_seo_performance.py" validate-run <bundle>/seo-performance-run.json --bundle <bundle>. - Route evidence-backed issues to the relevant audit owner and then
seo-action-plan. Route AI mention/citation observations toai-visibility-monitor; never merge the two reports into one score.
Provider operations mode
Treat provider mutation as a separate mode, not as measurement or implied closeout. Verify the exact property and current documented capability at mutation time; capture pre-state; restate the authorized action and target set; execute only that set; record the provider receipt/time/status, partial failures, and post-state; then validate a separate provider-operation-receipt.json. Search Console sitemap submission, URL Inspection, recrawl/validation UI flows, Google's restricted Indexing API, Bing submission, and IndexNow have different capabilities—never substitute one for another.
Let <suite-root> use the same resolution rule as above, then run python "<suite-root>/scripts/validate_provider_operation.py" validate-receipt <bundle>/provider-operation-receipt.json --bundle <bundle>. A receipt is not a Performance Run. Provider acceptance is not crawl, indexing, serving, citation, referral, traffic, or conversion; schedule those as separately observed follow-ups and preserve provider lag.
For an experiment, predeclare the hypothesis, assignment unit, eligible cohort, primary outcome, guardrails, duration/stopping rule, confounders, and rollback before implementation. Require verified exposure and a comparable holdout or label the result observational; do not retrofit causation after seeing a favorable delta.
Boundaries
Search Console and Bing metrics are platform-defined observations. Average position is not a fixed rank. A canonical URL row is not proof of every variant's behavior. A crawler request is not indexation, citation, referral, or conversion. An answer citation is not an organic visit.
Do not invent missing exports, reconstruct hidden queries, bypass property access, or claim ranking, traffic, revenue, or conversion guarantees.
Output
Return the validated run, source and window, measured dimensions, separate metric tables, data-quality limits, comparable drift when available, and the next evidence owner. When broad-growth tracks were declared, report nearer-term and head-term cohorts separately and retain missing/zero-visibility strategic queries as explicit limitations or research/measurement dependencies. In provider-operations mode return the separately validated receipt, exact property/targets, response and failures, pending outcome states, and follow-up owner. Keep conclusions no stronger than the supplied data.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

