hola-poke-yeoksam

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

• Third-party content exposure detected (medium risk: 0.30). The skill connects to a remote MCP server (https://hola-poke-yeoksam-skill.onrender.com/mcp) to fetch menu info, shop info, and enter events, allowing the server operator or anyone who can influence the remote server's responses to inject arbitrary free text into the agent's workflow via fields like message and next_action.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs the agent to execute an external CLI installer from an unverified third-party repository (@nomadamas/k-skill) via npx, establishing an uncontrolled runtime dependency.