x-mastery-mentor

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md (Scene E: "账号诊断与数据采集" Step 2) explicitly instructs the agent to fetch and scrape public X pages (open https://x.com/{username} via computer-use / browser tools) and to read/interpret users' tweets (untrusted, user-generated content) as input for diagnoses and action plans, which can directly change tool use and recommendations.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly opens and scrapes https://x.com/{username} at runtime (computer-use / browser tools) to extract tweets which are injected into the agent's context and are a required dependency for the account-diagnosis flow, so external page content can directly control the agent's prompts/outputs (https://x.com/{username}).