Claude Code integration

On WordPress, the Citevera plugin applies fixes you approve from the WP admin dashboard. For Next.js, Astro, Hugo, Jekyll, custom React, and other frameworks, Citevera exports a SKILL.md file that Claude Code reads and applies for you.

What SKILL.md contains

Each export holds the full set of recommendations from one scan, ordered the same way Citevera ranks them:

Frontmatter: site URL, scan id, scan date, overall score. Claude Code uses these fields to scope edits to the right project.
Scores: AEO, GEO, crawlability, and the 5-stage AI decision funnel breakdown.
Top 3 moves: highest score lift per unit of effort.
Must Fix / High Leverage / Scale / Nice to Have sections: every audit item grouped by priority tier.
Per-item instructions: JSON-LD blocks for schema fixes, user-agent stanzas for robots.txt, full file bodies for llms.txt, tag directives for meta updates.
Review footer: reminders to run tests, verify live URLs, and rescan after applying.

How to use it

Run a Citevera scan on your site. The scan works for any public URL, regardless of framework.
On the scan report, click Export for Claude Code (SKILL.md). The file downloads as citevera-<host>-<scanId>.SKILL.md.
Move the file into your codebase. We suggest placing it at the repository root so Claude Code discovers it automatically, or in a folder your team already uses for Claude Code skills.
Open the project in Claude Code and prompt it to apply the SKILL.md recommendations. For example: "Apply the fixes in citevera-example.com-scan_abc.SKILL.md. Run the test suite after each category. Pause and ask before committing changes that alter content."
Review each diff, run your tests, and rescan on Citevera to confirm the score moved.

Framework path notes

The export uses framework-agnostic instructions, but a few fixes touch static file paths that vary by framework:

llms.txt and robots.txt: Next.js uses public/, Astro uses public/, Hugo uses static/, Jekyll uses the repo root. Claude Code detects the framework from your repository and adapts automatically.
JSON-LD schema: the export recommends adding schema blocks to the page <head>. In component frameworks, this typically means editing your root layout or a per-page head slot. Claude Code will locate the appropriate file.
Meta tags: same as schema. Your head template is the target. For server components, apply via the framework metadata API (Next.js metadata, Astro frontmatter, Hugo front matter).

Safety and review

The export is advisory. It reflects what the Citevera scan saw at audit time; your codebase may have constraints the scan cannot know about. Before committing:

Review each diff against your conventions.
Merge JSON-LD blocks into any existing @graph rather than stacking duplicate <script> tags on the same page.
Verify robots.txt and llms.txt return HTTP 200 with Content-Type: text/plain after deploy.
Run your test suite after each category of changes.
Rescan the site to confirm the score moved.

Who can export

SKILL.md export is available on the Starter plan and above. Free scans show the button but direct you to Citevera pricing on click. The WordPress plugin's dashboard-driven apply remains the zero-friction option for WordPress sites on any plan that includes the plugin.

WordPress plugin vs SKILL.md: which to use

WordPress site: use the Citevera plugin. One-click apply from the scan dashboard. No files to manage.
Anything else: use SKILL.md export. Works for Next.js, Astro, Hugo, Jekyll, Gatsby, Eleventy, custom React, static sites, and any codebase Claude Code can edit.