quick-game

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt explicitly tells the agent to check for and, if missing, ask the user for a MESHY_API_KEY and "store for model generation," which requires the agent to receive and handle a secret (and likely include it in API calls), creating a high exfiltration risk.

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

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow can fetch outsider-authored free text from a user-supplied tweet URL (x.com/twitter.com/fxtwitter.com) via the fetch-tweet skill, and then the agent uses that tweet text to generate the game concept—so the LLM ingests outsider tweet body text as prompt context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill will fetch tweet content at runtime when given a tweet URL (e.g., https://x.com/*/status/* or https://twitter.com/*/status/* or https://fxtwitter.com/*/status/* via the fetch-tweet skill), and that fetched content is used to generate the game concept and subagent prompts, so external content directly controls agent instructions.