The Agent Skills Directory

[SAFE]: The skill is designed to analyze web content from user-provided URLs to identify SEO decay. This involves processing external data, but the instructions focus on scoring metrics and generating updated prose without executing untrusted code.
[COMMAND_EXECUTION]: The skill instructions involve writing audit summaries and state updates to a local memory/ directory. This is a standard practice for this vendor's skill set to maintain persistence and record-keeping across sessions within the agent's environment.
[DATA_EXPOSURE]: The skill accesses project-specific files such as CLAUDE.md and CONNECTORS.md to gather context for its SEO analysis. These operations are limited to the project's own documentation and do not target sensitive system credentials or private user data.

content-refresher