SXO Search Experience Optimization
Reverse-engineer Google SERPs to find page-type and search-intent mismatches behind weak rankings, then turn the findings into actionable UX improvements.
The skill requires SSRF-safe rendering, user confirmation before DataForSEO calls, and explicit limitation reporting, showing some awareness of permissions and external effects. It does not fully disclose page-data flows, third-party search-data handling, credential isolation, rollback, or source attribution within the skill itself, so points are deducted.
The pipeline, classification rules, error handling, and quality checklist are fairly detailed, and the skill references supporting scripts and documents. However, this is a static review with no execution of key paths; script parameters, WebSearch availability, and revision-level reproducibility of all referenced files cannot be confirmed, so the score stays below the static ceiling.
Trigger phrases, required URL/keyword inputs, command variants, and output structure are clear across SERP analysis, user stories, personas, and wireframes. Non-fit boundaries, Chinese-language support, and reachability of the Google/WebSearch-dependent core for mainland-China users are not adequately specified, so points are deducted.
The skill has front matter, MIT licensing, a version, layered reference documents, command documentation, output examples, error handling, and limitations guidance. It lacks skill-specific installation notes, FAQs, changelog details, explicit maintenance responsibility, and an update path; some governance evidence exists only at repository level, so points are deducted.
The intended output converts live SERP signals into page-type judgments, gap analysis, persona scoring, and concrete wireframes, which could directly support SEO decisions. The core results depend on live fetching and search, while the supplied materials provide no verifiable key-path output evidence for this skill, so the static ceiling requires a conservative deduction.
The taxonomy, scoring method, and signal sources are documented and auditable, and the README claims tests and primary-source alignment. The supplied evidence does not show seo-sxo-specific test results, independent review, or reproducible key-path examples; SERP observations also cannot be independently verified from static files, so points are deducted.
- Core analysis depends on Google/WebSearch, target-URL fetching, and optionally DataForSEO; network reachability, anti-bot controls, and regional differences may produce incomplete results.
- The skill requires at least five SERP results but does not define result snapshots, locale, language, device, or timestamp, limiting reproducibility.
- The skill does not clearly specify retention, redaction, or lifecycle handling for sensitive page content, queries, or third-party API data.
- README-level testing and security claims are repository context and do not establish that seo-sxo key outputs were tested.
What does this skill do, and when should you use it?
This is the individual SXO skill inside the claude-seo repository, focused on Search Experience Optimization rather than a complete technical SEO audit. It fetches and parses a target page, analyzes Google results for a keyword, checks whether the page type matches the SERP consensus, and evaluates the page through user-story and persona lenses. Its output covers the SERP landscape, page-type alignment, derived user stories, a seven-dimension gap score, persona scores, and prioritized actions. It is a good fit for teams whose pages are technically sound but still fail to rank.
It fetches the target URL with scripts/render_page.py in automatic mode and uses scripts/parse_html.py to extract the title, H1, meta description, heading hierarchy, word count, structured data, CTAs, and media elements. If no keyword is supplied, it derives a primary keyword from title/H1 overlap. It then uses WebSearch to analyze up to 10 organic results and SERP features, including page type, content format, authority tier, schema, media signals, People Also Ask, ads, related searches, local packs, shopping results, and AI Overviews. Using page-type classification rules, it detects mismatches, derives 3-5 user stories from SERP signals, and calculates a 0-100 SXO Gap Score across page type, content depth, UX, schema, media richness, authority signals, and freshness. It also derives 4-7 personas and scores relevance, clarity, trust, and action; the wireframe mode produces IST/SOLL semantic HTML outlines with concrete placeholders.
- An SEO consultant investigates a page with strong technical metrics but persistently weak rankings and tests whether the page type conflicts with the SERP.
- A content team is restructuring a page for a target keyword and uses PAA, ad copy, and related searches to define user stories and content requirements.
- A product or growth team is improving a landing page and uses persona scoring to identify concrete clarity, trust, or CTA gaps.
- An SEO and UX team needs an implementation-oriented redesign brief and uses wireframe mode to produce current-state and target-state page outlines.
What are this skill's strengths and limitations?
- Targets page-type and search-intent alignment issues that ordinary technical SEO checks can miss.
- Turns PAA questions, ad themes, related searches, and AI Overview signals into user stories and recommendations.
- Provides a distinct SXO Gap Score and explicitly separates it from an SEO Health Score.
- Offers persona-level relevance, clarity, trust, and action scoring with concrete improvement suggestions.
- It depends on an accessible target URL, page fetching, and WebSearch; failed fetching or a small SERP sample limits the analysis.
- Pages with scroll-triggered hydration, interaction-loaded content, or race-prone third-party widgets may produce noisy findings.
- DataForSEO, Firecrawl, and Banana are optional extensions of the broader repository and should not be attributed to this skill as default capabilities.
- The SKILL.md uses Claude Code-style /seo commands, so its trigger and tool integration require adaptation in other clients.
How do you install this skill?
The skill is located at skills/seo-sxo/SKILL.md in the repository. The README documents collection-level installation in Claude Code: run /plugin marketplace add AgriciDaniel/claude-seo, then /plugin install claude-seo@agricidaniel-claude-seo; alternatively, run git clone --depth 1 https://github.com/AgriciDaniel/claude-seo.git followed by bash claude-seo/install.sh. The README does not document a separate installation procedure for only the seo-sxo directory. The repository states Python 3.10+ is required, and Playwright Chromium can be used for SPA rendering.
How do you use this skill?
In Claude Code, run /seo sxo <url> for a full SXO analysis, or /seo sxo <url> <keyword> with a specific keyword. Use /seo sxo wireframe <url> for an IST/SOLL wireframe, or /seo sxo personas <url> for persona-only scoring. If no keyword is supplied and none can be extracted, the skill asks for a target keyword. DataForSEO is optional: when its MCP tools are available, the skill requires a cost estimate and user confirmation before API calls; otherwise it uses WebSearch and notes the reduced precision.
How does this skill compare with similar options?
Unlike a technical SEO audit that primarily checks crawlability, performance, and structured data, this skill focuses on whether a page matches the dominant SERP page type, content format, and user needs for a keyword. It is also narrower than tools such as Screaming Frog or Ahrefs Site Audit: its scope is single-page SERP reverse analysis, user-story derivation, persona scoring, and wireframe guidance.