injective-docs-style

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs installing and running external skill packages at runtime via "npx skills add bguiz/devrel-agent-skills --skill devrel-docs-style" and "npx skills add InjectiveLabs/agent-skills --skill injective-mcp-servers", which would fetch and execute remote code (GitHub-hosted skill repositories) that the agent would rely on to operate.