@skene/database-skills

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

• Potentially malicious external URL detected (high risk: 0.90). The README/skill docs explicitly instruct running runtime commands that fetch and execute remote code (e.g., "npx @skene/database-skills init" and "npx skills add https://github.com/SkeneTechnologies/skene/tree/main/skills"), which install/modify agent skills (SKILL.md) that an AI agent will read and thus can directly control prompts/behavior.