paid-ads

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

• Third-party content exposure detected (medium risk: 0.30). The required search-terms pass ingests Google Ads search query text that outsiders can generate by typing queries into Google, with no agent selection step, so this is medium rather than low.

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

• Potentially malicious external URL detected (medium risk: 0.30). The only operational inbound path is the instruction at SKILL.md line 12 to read linked skill files from the linked source (github.com/Quegenx/agent-skills) when the installed copy is unavailable, and the agent would adopt that content as instructions; Quegenx is not an established first-party vendor, so this is unofficial, stable-looking third-party content and the tier is medium.