make-game

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

• Third-party content exposure detected (high risk: 0.80). The skill explicitly fetches and ingests public tweets (via the tweet-pipeline.md flow that uses the fetch-tweet skill / WebFetch to pull tweet text from fxtwitter/twitter URLs) and uses that untrusted, user-generated text to "creatively abstract a game concept" that drives subsequent decisions and tool use, so it clearly exposes the agent to untrusted third‑party content that could enable indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches tweet text at runtime via the fetch-tweet flow (converted to URLs like https://api.fxtwitter.com//status/) and then uses that raw tweet content as the creative input injected into subagent prompts to generate game concepts, which means external content can directly control agent prompts.