specification-website

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md describes runtime fetching of outsider-authored spec content over HTTP (e.g., per-page Markdown via https://specification.website/spec/<category>/<slug>.md, or large loads via llms-full.txt / llms.txt) that would be ingested into the agent’s LLM context as free text.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs agents to fetch and use external machine-readable endpoints at runtime (e.g., https://mcp.specification.website/mcp and https://specification.website/llms-full.txt) which can supply prompts or one-shot context to the agent, so these are runtime dependencies that directly control agent behavior.