lab-retro

SUSPICIOUS. Most capabilities match a retrospective skill, but the final feedback step routes participant data through an unverifiable Vercel proxy rather than the official Baserow API. This is a data-flow integrity issue and makes the skill higher risk than its stated purpose warrants, though there is no clear evidence of credential theft or confirmed malware.