Firecrawl SEO Site Analyzer
Crawl, map, scrape, and search entire websites through the Firecrawl MCP server.
The skill requires installation and availability checks for Firecrawl MCP and documents credit awareness, errors, and fallback options. It does not clearly require confirmation before large crawls or browser actions, fully disclose data flows, address sensitive scraped content, explain external effects, or provide rollback, so points are deducted.
Commands, parameters, dependency checks, common errors, and fallback behavior are documented. However, the core path depends on an external Firecrawl MCP server, skill-specific reproduction and test coverage are absent, abnormal-input diagnostics are limited, and claims such as anti-bot handling are not verified by the supplied files; the score therefore stays within the static-review cap.
Trigger phrases, crawl/map/scrape/search scenarios, and several parameter boundaries are clear. Non-fit cases, privacy constraints, scale limits, and Chinese-language guidance are weak, and the core dependency may be difficult to reach from mainland-China networks, so points are deducted.
The skill has front matter, MIT licensing, a version, installation guidance, quick reference, parameter tables, examples, error handling, and cross-skill integration notes. It lacks a skill-specific changelog, explicit maintenance responsibility and update path, FAQ coverage, and a more complete output contract, so it is not full marks.
The documented workflow can organize site discovery, crawling, JavaScript rendering, and site-scoped search into directly usable SEO operations. Results still depend on the external service, downstream skills, and user filtering; the supplied examples are descriptive rather than reproducibly verified, so the static cap and uncertainty require deductions.
The skill includes parameter definitions, an error table, and procedural guidance, while repository context claims CI and tests. The supplied tests do not cover this Firecrawl skill's MCP calls, output correctness, or cost controls, and README claims and examples are not independent reproduction evidence, so only limited credit is justified.
- Large crawls may consume substantial Firecrawl credits; the skill asks the user to be informed but does not require confirmation or enforce a budget cap.
- Crawling arbitrary user-supplied sites creates external network activity; confirm targets, scope, privacy, authorization, and robots/access constraints before execution.
- Firecrawl MCP is a core dependency, but mainland-China reachability, data residency, and third-party data handling are not documented.
- No dedicated tests for this skill's MCP calls or output correctness are supplied; SEO conclusions require human review.
What does this skill do, and when should you use it?
This is the Firecrawl extension skill bundled in the claude-seo repository, focused on site-wide crawling, URL discovery, single-page scraping, and site-scoped search. It uses Firecrawl MCP to render JavaScript pages and return content, metadata, links, HTML, or screenshots. It fits users who need site-level SEO collection and structural analysis, but it requires an installed and connected Firecrawl MCP server. Large crawls consume Firecrawl credits.
It provides /seo firecrawl crawl <url> for page content, metadata, and link extraction; map for URL discovery without fetching page content; scrape for JavaScript-rendered single-page extraction with optional screenshots; and search for searching within a specified website. Crawl and scrape options include page limits, link depth, include or exclude paths, output formats, wait conditions, timeouts, and browser actions. The collected data can support content inventories, broken-link checks, URL-pattern analysis, orphan-page discovery, and site-structure analysis.
- An SEO consultant auditing a large site can map URLs first, then crawl only priority sections or pages.
- A team working with React, Vue, or Angular pages can use scrape to capture content that static retrieval may miss.
- A technical SEO specialist checking internal links can crawl link outputs and inspect targets for broken URLs.
- A content team building a page inventory can extract titles, meta descriptions, headings, metadata, and page content at scale.
- An analyst investigating URL patterns, parameter bloat, duplicates, or orphan pages can map the site before targeted crawling.
What are this skill's strengths and limitations?
- Covers full-site crawling, URL mapping, JavaScript-rendered scraping, and site-scoped search.
- Supports path inclusion and exclusion for focused audits.
- Can return Markdown, HTML, links, and screenshots.
- Is specifically suited to dynamic sites and single-page applications.
- Requires the Firecrawl MCP server and cannot operate independently of that service.
- Crawling and mapping consume Firecrawl credits and may be constrained by quotas, rate limits, or plan limits.
- Timeouts, 403 responses, robots.txt restrictions, or anti-crawling policies can produce incomplete results.
- The provided material does not document tests or cross-platform validation for this extension itself.
How do you install this skill?
From the claude-seo repository directory, run ./extensions/firecrawl/install.sh. The skill requires the Firecrawl MCP extension to be installed and connected; if firecrawl_scrape or another Firecrawl tool is unavailable, installation and configuration must be completed first. The source does not document the full Firecrawl API-key setup procedure.
How do you use this skill?
In a Claude Code environment that supports the skill, run /seo firecrawl crawl https://example.com, /seo firecrawl map https://example.com, /seo firecrawl scrape https://example.com/page, or /seo firecrawl search "keyword" https://example.com. Tell the user the estimated credit usage before a large crawl; the documented defaults are up to 100 crawled pages and link depth 3, with adjustable parameters.
How does this skill compare with similar options?
Compared with fetch_page.py, Firecrawl scrape is better suited to JavaScript-rendered pages, clean Markdown extraction, and screenshots; fetch_page.py is positioned for static pages and response-header access. When Firecrawl is unavailable, the source also suggests WebFetch for basic HTML retrieval or fetch_page.py for single-page analysis.