agent-skill-maker

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly requires using the WebFetch tool at runtime to retrieve and read these external documents—https://agentskills.io/specification.md, https://code.claude.com/docs/en/skills.md, and https://platform.claude.com/docs/en/agents-and-tools/agent-skills/best-practices.md—which are treated as required, authoritative references that directly control the agent's prompts and behavior.