The Agent Skills Directory

[SAFE]: The skill requires API keys for the Orthogonal and Anthropic services. These are correctly listed as prerequisites for the user to configure, with no hardcoded secrets or credentials present in the code.
[COMMAND_EXECUTION]: The skill uses specific command-line tools such as orth, exa-search, and perplexity for SEO auditing and data retrieval. These commands use template placeholders for user-provided data and are consistent with the skill's stated purpose.
[PROMPT_INJECTION]: The workflow in Step 5 involves processing external content from URLs and search results and interpolating it into a prompt for content refreshing. This represents an indirect prompt injection surface; however, the risk is minimal as the resulting output is intended for content generation and the skill lacks high-privilege capabilities that could be exploited via this vector.

evergreen-content-refresher