The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes standard shell commands including curl, grep, sed, diff, and comm to perform technical SEO checks. These commands are clearly defined in the references/checks.md file and are scoped to the primary purpose of auditing category and product pages.
[EXTERNAL_DOWNLOADS]: Downloads public HTML content, sitemap.xml, and robots.txt from user-specified target sites using curl to evaluate crawlability and indexing status.
[DYNAMIC_EXECUTION]: Employs inline Python scripts via python3 -c to process HTML text, calculate word counts, and join URLs. These scripts use standard libraries and follow static templates provided within the skill's reference files.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted HTML data from external websites, creating a potential attack surface.
Ingestion points: External website content is fetched via curl in references/checks.md and piped into subsequent commands.
Boundary markers: The skill does not use explicit delimiters to separate external data from agent instructions, though the agent's actions are constrained by the provided workflow.
Capability inventory: The skill can perform network reads (curl), write to temporary files in /tmp/, and process data using Python.
Sanitization: The skill applies regex-based filtering (re.sub) to strip potentially dangerous or irrelevant HTML tags like <script>, <style>, and navigation headers before calculating word counts, which reduces the risk of processing malicious instructions embedded in the site content.

ecommerce-technical-seo-audit