x-research

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md workflow explicitly performs site:x.com searches to collect public X (Twitter) posts and then uses WebFetch to retrieve and summarize external blogs, GitHub Issues, and other public web pages (see "Phase 1: 初期検索" and "Phase 3: 詳細情報の取得"), which are untrusted user-generated sources that the agent reads and acts on.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill uses WebFetch at runtime to retrieve external pages (e.g., site:x.com / x.com/{username}/status/{id} and external sources like https://zenn.dev, https://qiita.com, https://dev.to, and https://github.com) whose fetched content is injected into the agent's analysis and thus can directly control prompts.