skills-management

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

• Third-party content exposure detected (high risk: 0.90). SKILL.md explicitly instructs the agent to fetch and evaluate public skills.sh and GitHub pages (e.g., "use WebFetch on https://skills.sh///" in the "Review & Compare Results" section and the remote-skill-assessment reference), so the agent will read untrusted, user-generated web content and use it to drive installation/recommendation actions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs the agent to WebFetch and parse external skill pages at runtime (e.g., https://skills.sh///), so remote content would be fetched during execution and can be injected into the model context to directly influence prompts/behavior.