benchmark

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

• Third-party content exposure detected (high risk: 0.70). The skill explicitly navigates to arbitrary pages and executes JavaScript to ingest page data (Phase 3: "Performance Data Collection" with $B goto and $B eval "performance.getEntriesByType(...)"), so it consumes untrusted public web content that directly drives benchmarking decisions and recommendations.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's SETUP step downloads and executes remote code at runtime via curl -fsSL "https://bun.sh/install" (saved to a temp file and run with bash), making https://bun.sh/install a runtime external dependency that executes remote code.