skill-factory

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

• Third-party content exposure detected (high risk: 0.90). The skill's workflow explicitly accepts and fetches public website URLs (see "模式 A：智能网站分析" in SKILL.md and references/web-analysis-flow.md which call WebFetch/Read and search) and the example script scripts/website-analyzer.py.analyze_website(url) reads and uses that third-party web content to drive analysis and generate skill-planning decisions, so external page content can materially influence the agent's actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly uses WebFetch/Read at runtime to fetch and inject content from user-provided websites (e.g., https://vuejs.org) into its analysis and skill-generation prompts, so external URLs like https://vuejs.org can directly influence the agent's instructions.