find-skills

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to search and read skills from the public skills.sh leaderboard and to fetch/install packages from GitHub or other sources via npx skills find / npx skills add, meaning the agent will ingest untrusted, user-generated third-party content (skills.sh and GitHub repositories) and use that content to decide which skills to recommend or install.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs running runtime install commands (e.g., "npx skills add owner/repo@skill") which fetch and install remote skill code from sources like https://skills.sh/ and GitHub (for example vercel-labs/agent-skills), meaning external content is downloaded at runtime and can execute code or supply prompts controlling the agent.