build-mcp-server

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

• Third-party content exposure detected (high risk: 0.70). This skill's SKILL.md explicitly directs the agent to "fetch https://claude.com/docs/llms-full.txt" ("Load Claude-specific context first"), a public third-party web document that the agent must read and use to shape its guidance and decisions, so external content can materially influence tool selection and behavior.

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

• Potentially malicious external URL detected (high risk: 1.00). This skill explicitly requires fetching external guidance at runtime — "Before answering... fetch https://claude.com/docs/llms-full.txt" — and uses that content to control the assistant's instructions/guidance, so the URL is a runtime dependency that directly influences prompts.