dogfood

SUSPICIOUS: the skill's core purpose is coherent, but it has medium security risk because it chains into other skills/tools, can take external actions like PR commenting and browser interaction, and processes untrusted web content during exploratory QA. No strong signs of malware or credential theft are present.