Reviews short, high-value strings where every character counts -- page titles, meta descriptions, schema description fields, FAQ answers, GitHub repo taglines, profile bios, and social-card copy. These strings are too short for Flesch scoring or paragraph-level readability analysis, so this skill applies the checks that actually matter at that scale.
- Front-loading -- the most specific, searchable term sits near the start of the string
- Concreteness -- names, numbers, and specific claims over vague language
- Filler and hedging -- flags wasted characters like "really", "just", "very", "basically"
- Active voice -- unless the object is genuinely the subject
- Title/description duplication -- descriptions that restate the title waste SERP space
- Word choice -- prefers common synonyms unless the domain term is what users search for
- Truncation fit -- checks against platform limits (SERP title ~60 chars, description ~160, GitHub ~100, Twitter bio 160)
- One idea per field -- titles and descriptions that try to promise two things land fuzzy
Trigger this skill when you want to review short copy. Example prompts:
- "Check these meta descriptions"
- "Review my page titles for SEO"
- "Is this tagline good?"
- "Audit my GitHub bio"
The skill outputs a per-string report with pass/warn/fail on each check and a concrete rewrite for any string that fails.
This skill is commonly chained into by other skills:
- astro-seo -- invokes metadata-check on generated SEO strings
- wp-readme-optimizer -- invokes it on plugin names, short descriptions, and FAQ answers
- github-repo -- invokes it on repo descriptions and README taglines
- github-profile -- invokes it on bios and pinned repo descriptions
For multi-paragraph prose, use readability-check instead.
npx skills add jdevalk/skills --skill metadata-check