fdk-uat

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

• Third-party content exposure detected (medium risk: 0.30). The skill workflow involves pulling code and bootstrap scripts from arbitrary user-specified or repository-defined GitHub raw URLs at runtime during the UAT process.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill executes curl against https://raw.githubusercontent.com/<owner>/<repo>/... to fetch and pipe setup/bootstrap scripts into a shell, but the repository path contains user-controlled placeholders (<owner>/<repo>), which represents an arbitrary/unknown third-party or dynamic remote source.