agentleFS
oegeyilmaz9oegeyilmaz9/seo-aeo-geo-ultimate/skills/seo-technical/SKILL.md

seo-technical

Skill6 starsChanged 3 months agoApache-2.0View on GitHub

Use when investigating crawlability, indexability, canonicalization, rendering, performance, redirects, directives, sitemaps, crawler controls, site migrations, sudden traffic/index incidents, manual actions, hacked-site recovery, or when evaluating, generating, validating, publishing, and maintaining an evidence-scoped llms.txt publisher guide; return safe, testable implementation recommendations rather than a generic technical score.

39 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use seo-maps, seo-ecommerce and 4 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • timruebenachSkill

    seo-maps

    Maps intelligence for local SEO — geo-grid rank tracking, GBP profile auditing via API, review intelligence across Google/Tripadvisor/Trustpilot, cross-platform NAP verification (Google/Bing/Apple/OSM), competitor radius mapping, and LocalBusiness schema generation from API data. Three-tier capability: free (Overpass + Geoapify), DataForSEO (full intelligence), DataForSEO + Google (maximum coverage). Use when user says "maps", "geo-grid", "rank tracking", "GBP audit", "review velocity", "competitor radius", "maps analysis", "local rank tracking", "Share of Local Voice", or "SoLV".

    Also in 17 projects that use this

  • timruebenachSkill

    seo-ecommerce

    E-commerce SEO analysis: Google Shopping visibility, Amazon marketplace intelligence, product schema validation, competitor pricing analysis, and marketplace keyword gaps. Combines on-page product SEO with marketplace data from DataForSEO Merchant API. Use when user says "ecommerce SEO", "product SEO", "Google Shopping", "marketplace SEO", "product schema", "Amazon SEO", "product listings", "shopping ads", or "merchant SEO".

    Also in 17 projects that use this

  • mangollcSkill

    seo-backlinks

    Backlink analysis and link building strategy: competitor backlink profiling, outreach plan generation, monthly SEO reporting framework, toxic link identification, link opportunity scoring, and unlinked mention discovery.

    Also in 18 projects that use this

  • +3 more from timruebenach

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 5 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 36 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read references/technical-evidence-protocol.md before a formal technical audit.

    line 21
  2. Distinctive:

    Read external-claim-adjudication.md when evaluating third-party audit or scanner claims.

    line 22
  3. Distinctive:

    Read live-verification-matrix.md before closing a released technical, template, redirect, sitemap, canonical, hreflang, or indexability …

    line 23
  4. Distinctive:

    Read favicon-search-protocol.md when a site lacks a search-result icon, has the wrong icon, or a broad site optimization includes hostname …

    line 24
  5. Distinctive:

    If evidence is unavailable, give a safe collection plan—not a guessed fix.

    line 35
  6. Very specific:

    Find the smallest cause: Test conflict pairs such as noindex vs canonical, raw vs JavaScript-injected metadata, redirect target vs sitemap …

    line 41
  7. Distinctive:

    Classify the change or external claim: Use confirmed defect, layer mismatch, supported opportunity, false positive, or unverified.

    line 42

“Common” is measured: kept by at least 6 in 10 of the 36 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Technical SEO

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. Include concise robots/XML-sitemap discovery comments and llms.txt navigation sentences pointing to actual Markdown answers; preserve the native protocols and verify delivery as described in the local contract.

Lane outputs: technical-findings.md; markdown-delivery.md; live-verification.md (combine small related steps with stable section links). Own robots.txt and llms.txt delivery, MIME/UTF-8, explicit Markdown routing, access, redirects, cache invalidation and a documented HTTP canonical/indexing policy. Verify the full declared resource set after authorized release; no SPA HTML fallback may pass.

Purpose

Diagnose concrete technical behavior from headers, raw and rendered pages, robots files, sitemap inventories, application configuration, field data, logs, and platform documentation. Create scoped implementation recommendations; do not deploy code, alter production directives, or claim an indexing/ranking outcome.

For a standalone llms.txt request, create a maintained publisher guide—not a crawler directive or promised visibility control. Recommend it as a reasonable low-cost future-readiness layer when it can stay aligned with truthful public sources and has a maintenance owner; a documented consumer strengthens the case but is not mandatory.

Read references/technical-evidence-protocol.md before a formal technical audit. Read external-claim-adjudication.md when evaluating third-party audit or scanner claims. Read live-verification-matrix.md before closing a released technical, template, redirect, sitemap, canonical, hreflang, or indexability change. Read favicon-search-protocol.md when a site lacks a search-result icon, has the wrong icon, or a broad site optimization includes hostname identity. For feature lifecycle, crawler, IndexNow/change notification, or protocol-capability work, also read platform-control-protocol.md and validate the current registry before recommending a control. For llms.txt suitability, generation, validation, or publishing, read llms-txt-protocol.md and use its bundled template and validator. For a migration/replatforming, sudden traffic or indexation incident, manual action, or hacked-site recovery, read migration-and-recovery-protocol.md before recommending or approving a response.

Evidence before recommendations

Collect the affected URLs, expected state, locale/device where relevant, timestamp, HTTP response/headers, raw HTML, rendered output when JavaScript is involved, canonicals/robots directives, sitemap membership, redirects, and any authorized first-party reporting. For performance, distinguish lab diagnostics from field/user data. For crawler behavior, distinguish a request in logs from a search/index/AI-answer outcome.

Supported crawler CSV and Apache/Nginx access logs can be normalized with python "<suite-root>/scripts/import_seo_exports.py" crawler-csv ... or server-log ...; follow <suite-root>/docs/DATA-IMPORT-ADAPTERS.md. These adapters preserve provenance and privacy boundaries but deliberately do not fabricate HTML captures, a Site Graph, indexation, crawler identity, or performance outcomes.

If evidence is unavailable, give a safe collection plan—not a guessed fix. Never bypass authentication, rate limits, paywalls, WAF controls, or robots rules to create evidence.

Workflow

  1. State the expected behavior. Is the URL intended to be indexable, canonical, localized, discoverable, rendered, fast enough for users, or excluded? Identify the owner/system that can change it.
  2. Capture the actual behavior. Record raw response, directives, canonical, rendered content, redirect chain, linked/sitemap evidence, hosting/CDN behavior, and relevant device/locale. Keep transient tool errors separate from confirmed defects and identify the layer each claim measures.
  3. Find the smallest cause. Test conflict pairs such as noindex vs canonical, raw vs JavaScript-injected metadata, redirect target vs sitemap URL, or locale target vs hreflang cluster. Do not infer a cause from a single score or generic audit rule.
  4. Classify the change or external claim. Use confirmed defect, layer mismatch, supported opportunity, false positive, or unverified. Only a confirmed defect is automatically required. Link platform-specific controls to current primary documentation.
  5. Plan the implementation. For every recommendation state the affected URLs/templates, owner, precondition, exact desired state, verification capture, risk, approval, and rollback. Route cross-team work through seo-action-plan.
  6. Verify after change. Match the verification universe to the change universe: one page for a page-only edit, the affected route class for a shared template/resolver, and the full corpus or justified closed population for sitemap/index contracts. Re-capture raw, hydrated, hosting/CDN, and provider layers in scope. Verify redirects by status, Location, sitemap exclusion, and target state rather than by document-only metadata. Discovery, crawling, indexing, performance, and AI visibility are separate outcomes.

For migrations and incidents, preserve a last-known-good state, exact change timeline, affected cohorts, launch/rollback criteria, and unresolved competing hypotheses. For a compromise, security containment and root-cause remediation own the response; SEO verification follows rather than replacing them.

Technical domains

  • Crawl/discovery: robots handling, internal discovery, server health, sitemap inclusion, redirects, and crawl-access failures.
  • Index/canonical: canonical/noindex conflicts, duplicate URL handling, status codes, redirects, and content availability.
  • Rendering: raw versus rendered title, robots, canonical, meaningful content, and structured data; use actual captures rather than framework assumptions.
  • Search-result identity: hostname home-page favicon declaration, icon response/format/dimensions, stable URL, crawl access, and representative brand asset. Do not mark favicon work complete from a file existing somewhere in the repository.
  • Experience: responsive layout, interaction and rendering diagnostics, and field data where available. Core Web Vitals and tests guide improvement; they do not guarantee a ranking result.
  • International: send locale clusters and annotations to seo-hreflang; do not repair language targeting from a single URL.
  • Structured data: send truthful markup changes to seo-schema; validate the visible page and eligible documentation first.
  • Optional machine-readable guide: for llms.txt, assess public source-of-truth coverage, inclusion/exclusion policy, freshness, owner, and release path. When upkeep is low-cost and reliable, offer and generate it as future-readiness even without a confirmed consumer. Keep it separate from robots, access controls, and promises of ranking, retrieval, or citation.

llms.txt publication workflow

Let <suite-root> mean ${CLAUDE_PLUGIN_ROOT} in Claude Code. In Codex, read .seo-suite-runtime.json beside this SKILL.md when present and use its suite_root value; otherwise use the absolute repository checkout.

  1. Inventory the canonical public resources a client should understand; exclude private, authenticated, staging, personalized, ephemeral, and security-sensitive URLs.
  2. Name the maintenance owner and the source changes that must trigger regeneration or review.
  3. Copy <suite-root>/skills/seo-technical/assets/llms.txt.template, replace every placeholder from the public source of truth, and keep only useful absolute HTTPS links.
  4. Validate the candidate with python "<suite-root>/scripts/validate_llms_txt.py" validate-file <site-root>/llms.txt.
  5. Deliver or publish it at /llms.txt only within the authorized scope. Verify the live response and add the file to the ordinary documentation/content release path.
  6. Describe the file as an owner-required publisher guide within the Markdown discovery profile, or an optional guide outside that profile. Google documents that special AI text files are not required for AI Search eligibility; do not convert this file into a ranking, retrieval, citation, or traffic claim.

Crawler-control policy

Treat crawler controls as an explicit business, legal, privacy, and technical decision. Identify the exact current documented user-agent and its declared purpose before proposing a rule. Training, search indexing, user-initiated fetches, and product-specific controls can be different agents. The Robots Exclusion Protocol is crawler guidance, not authorization or proof of downstream behavior.

Do not copy a blanket AI-bot block/allow template, assume a bot token is stable, or claim that blocking/allowing a bot changes citations. Crawler-control, noindex, canonical, robots, and WAF changes are high-risk and require an approved action plan, test URL, rollout/rollback, and current vendor documentation.

Using the same <suite-root> definition, before a feature, crawler, change-notification, or protocol decision, run python "<suite-root>/scripts/validate_platform_controls.py" validate-registry "<suite-root>/manifests/platform-controls.json" --bundle "<suite-root>/manifests" --as-of <current-UTC-time>. Treat an expired registry or unknown control as a research gap. Copy a registry row into a site-specific decision only after adding the actual owner, evidence, scope, and approval; the shipped registry deliberately leaves business decisions undecided.

For IndexNow or another notification path, submit only added, updated, or deleted URLs within the documented scope. Verify key/endpoint ownership, response receipts, rate/error handling, logs, and rollback. Notification proves a request was sent or received, not crawling or indexing.

Formal evidence handoff

When this work needs a cross-team, approval-ready plan, package evidence-bound findings as an immutable seo-findings.json bundle using the checked-out suite contract. Keep every referenced capture/source below raw/, retain declined claims and limitations, and run python "<suite-root>/scripts/validate_seo_findings.py" validate-findings <bundle>/seo-findings.json --bundle <bundle>. Send only a passing bundle to seo-action-plan; otherwise label the handoff provisional.

Live release verification

After a separately authorized release, verify the exact candidate/deployment identity using live-verification-matrix.md. For the raw HTTP and sitemap layer, run python "<suite-root>/scripts/verify_live_release.py" verify --plan <release-plan.json> --output <live-release-report.json>. Bind approved per-page copy with expected_title, expected_description, and expected_h1 overrides so a stale, generic, or truncated live value cannot pass merely because the element exists. When hostname identity is in scope, declare site_identity in the plan so the verifier checks the home-page icon link, exact favicon response, content type, square dimensions, and configured quality floor. A passing report proves only the declared live delivery checks; it does not prove crawl, index, rank, retrieval, citation, referral, or conversion.

Output

For each finding provide: observed evidence and timestamp; expected state; scope; confidence/limitation; proposed owner/action; verification; risk; and rollback. When requested and supportable, also return the validated llms.txt, its included/excluded source map, owner, and refresh triggers. Group only by priority/owner, never a blended technical score. Route broader page/editorial questions to seo-page or seo-content.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.