The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple internal Python scripts (e.g., sitemap_discovery.py, agent_ux_check.py, render_page.py) using the claude-seo run command. These scripts accept user-provided URLs as positional arguments, creating a surface for potential command injection if the underlying runner does not properly sanitize inputs.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to fetch and analyze external websites, robots.txt files, and XML sitemaps. This behavior is necessary for its primary purpose as an SEO auditor.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests and processes untrusted data from external web pages.
Ingestion points: Data enters through the analysis of HTML content, robots.txt, and sitemaps fetched by scripts like sitemap_discovery.py and render_page.py in the SKILL.md and references/agent-friendly-pages.md files.
Boundary markers: No explicit delimiters or instructions to ignore embedded malicious content are present in the prompts.
Capability inventory: The agent has the ability to execute subprocesses (via claude-seo run) and perform network operations across all analyzed scripts.
Sanitization: No sanitization or validation of the fetched external content is described before it is processed by the agent.

seo-technical