living-text-style

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

• Potentially malicious external URL detected (high risk: 0.80). The skill contains runtime Playwright/WebFetch examples that fetch Reddit JSON (e.g., https://old.reddit.com/r/ClaudeAI/search.json and https://old.reddit.com/r/ClaudeAI/comments/{id}.json?sort=top&limit=30), which are invoked at runtime to ingest external page content into the agent context and thus can directly influence prompts/outputs.