The Agent Skills Directory

[NO_CODE] (SAFE): The skill consists entirely of Markdown-based instructions and scoring rubrics. There are no scripts, binaries, or shell commands included in the provided content.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill is designed to process untrusted data (user-provided web pages or marketing copy). While it does not include explicit boundary markers or sanitization logic, it possesses no exploitable capabilities such as file writing, network requests, or command execution, limiting the impact to potential reasoning bias within the session.
[DATA_EXPOSURE] (SAFE): The skill requests performance data (conversion rates, traffic sources) from the user to aid analysis. However, it lacks automated mechanisms to exfiltrate this data or access sensitive environment variables or local credential files.

page-cro