superskills-upgrade

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill explicitly fetches and pulls code from a public GitHub repository (uses git fetch, git show origin/main:VERSION, and git pull origin main) and then runs the repository's setup script, so untrusted third-party content from that remote could alter behavior or inject instructions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill runs git fetch/git pull against the repository remote (suggested as https://github.com/ariadoss/superskills.git) during runtime and then executes ./setup from the checked-out code, so remote content from that URL can execute code and directly change agent behavior.