The Agent Skills Directory

[PROMPT_INJECTION]: The skill provides instructions for design workflows and does not contain directives to ignore system rules, switch to restricted modes, or bypass safety filters.
[DATA_EXFILTRATION]: There are no network operations (like curl or fetch) or patterns indicating the exfiltration of sensitive data.
[EXTERNAL_DOWNLOADS]: No external scripts, packages, or binary dependencies are requested or downloaded.
[REMOTE_CODE_EXECUTION]: The skill is purely instructional markdown. It does not invoke subprocesses, use eval-like functions, or generate dynamic code for execution.
[CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or references to sensitive local credential files (e.g., .ssh, .aws) were found.
[NO_CODE]: The skill consists entirely of markdown instructions and does not include any scripts or executable logic, significantly reducing the attack surface.

ux-flows