create-skill

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

• Third-party content exposure detected (high risk: 1.00). The workflow's Step 1 explicitly instructs the agent to "Use WebFetch" to fetch the latest spec from the public site https://agentskills.io so the agent will read and act on external, potentially untrusted web content when deciding frontmatter/schema and authoring behavior.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill's workflow explicitly instructs using WebFetch at runtime to fetch the spec from https://agentskills.io, and that remote spec directly controls frontmatter/schema and thus the agent's prompts/instructions, making it a required external runtime dependency.