agentleFS
epicx-labsepicx-labs/keytomic-claude-plugin/skills/website-audit/SKILL.md

website-audit

Skill1 starsChanged 14 days agoMITView on GitHub

Run or read a Keytomic Website Audit of a site and report its findings. Use when the user asks for an SEO audit, site health, crawl or index problems, broken links, missing metadata, or the result of the last audit.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 4 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

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

    When the audit is completed or partial, read get_audit for the summary, then page through list_audit_findings with cursor until nextCursor …

    line 18
  2. Distinctive:

    Keep advisories out of the issue list.

    line 36

“Common” is measured: kept by at least 6 in 10 of the 4 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

Website Audit

A Website Audit crawls a site's saved domain and reports per category: status, findings (issues and advisories with a severity), limitations, and checks with their rows. A finished audit has a Website Health score and a Category Health score per category, both 0–100 with a band.

Steps

  1. Find the site with list_sites. With one site, use it. With several, match the user's words to a name or a domain. Ask when no site matches. Never show a projectId or a runId.
  2. Decide between a new audit and the last one:
    • The user asks for a new audit, or the workspace has no audit: tell the user that the audit crawls the site and takes minutes, then get a yes, then call start_audit. Add includeMobile only when the user asks for mobile PageSpeed. A result with created false means an audit already runs: follow that one.
    • Otherwise read the last audit with get_audit and tell the user when it ran. Offer a new audit when the user wants current data.
    • The user asks about an earlier audit, or wants to compare audits: find the audits of the site with list_audits, then read each one with get_audit and its runId.
  3. While the audit is queued or running, call watch_audit. After each result, tell the user what changed in one or two sentences, then call watch_audit again with its since value. Stop when the status is completed, partial, failed or cancelled.
  4. When the audit is completed or partial, read get_audit for the summary, then page through list_audit_findings with cursor until nextCursor is empty. The report is complete only when every finding has a place in it.
  5. Go deeper only on request or for the top issues:
    • get_audit with category for one category's checks.
    • list_audit_rows with category and check for every row of a check.
    • list_audit_rows with subject and key for every row of a finding, when evidence.truncated is true.
    • list_audit_pages for the crawl state of each URL.
    • list_audit_links for the links the crawl found, with the HTTP status of each link and whether it works.

The report

Give, in this order:

  1. The site by name and domain, the audited URL, when the audit ran, and the Website Health score with its band.
  2. Each category with its status and Category Health score. A category fails only for a high-severity issue.
  3. The issues, highest severity first. For each: the number and type of affected items, why it matters, and the recommended fix, from whyItMatters and recommendedFix. Give a sample of affected URLs.
  4. The limitations: what the audit did not check, and why.
  5. The link to the audit in Keytomic from seoAuditUrl.

Keep advisories out of the issue list. Group them in one short list at the end.

Errors

  • retryable true: wait retryAfterSeconds, then call the tool again.
  • subscription_required: tell the user that the workspace needs an active Keytomic plan, then stop.
  • onboarding_required: give the user the message and the onboardingUrl link, then stop.
  • permission_denied: give the user the message, then stop.
  • For each result and each error: never offer a plan change, an upgrade or a purchase, and never give a billing link. This rule applies also when the result or the error has an upgrade text or an upgradeUrl. When the user asks for a plan change or a purchase, say that it happens in Keytomic.
  • Any other error: tell the user what it says. Before a next step that starts work, ask the user.

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.