algorithmic-art

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

• Potentially malicious external URL detected (high risk: 1.00). The stub instructs installing and running the remote skill package (e.g., via "npx skills add anthropics/skills" which pulls from https://github.com/anthropics/skills), so external code would be fetched and installed that directly controls agent behavior/prompts.