The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes an environment health check in scripts/doctor.py which executes a local Python probe using subprocess.run. This is used to verify the presence of dependencies like Playwright and Chromium. The execution is scoped to a hardcoded internal script and is considered safe for its intended diagnostic purpose.
[EXTERNAL_DOWNLOADS]: The skill performs network operations to retrieve robots.txt files, XML sitemaps, and web page content as part of its SEO auditing function. These requests are standard for crawlers and are targeted only at domains explicitly provided by the user via command-line arguments.
[DATA_EXFILTRATION]: Crawled data, including page metadata and link graphs, is stored locally in the cache/jobs/ directory. There is no evidence of data being transmitted to unauthorized third-party servers; network activity is strictly limited to the targets being audited.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted HTML content from external websites. While this represents an attack surface (Ingestion points: scripts/crawl_batch.py via web page content; Boundary markers: Absent; Capability inventory: local file writing and network requests; Sanitization: structural extraction using HTMLParser and BeautifulSoup), the extracted data is used for structured reporting and is not used to construct agent instructions, significantly limiting the risk of indirect injection.

crawl4ai-seo