The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/audit.py script uses subprocess.run to call the curl utility for fetching website headers and HTML content. The command is constructed as a list, which prevents shell command injection. This is a functional requirement for site auditing.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external websites, creating an indirect prompt injection surface.
Ingestion points: Web content (HTML, robots.txt, XML sitemaps) fetched by the audit script from remote URLs.
Boundary markers: The resulting audit reports use markdown backticks to encapsulate raw evidence, separating it from the agent's instructions.
Capability inventory: The skill has access to file read/write and script execution via Bash.
Sanitization: The script utilizes standard parsing libraries and truncates evidence to a maximum of 1000 characters to prevent processing large or malicious payloads.
[DYNAMIC_EXECUTION]: The scripts/test_audit.py file employs importlib.util to dynamically load the audit script for unit testing. This is standard testing practice and is restricted to the local filesystem.

seo-ops