dx-roaster

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

• Third-party content exposure detected (high risk: 0.85). Outsider free text can enter the LLM context when repo is a GitHub URL: the skill clones the repo and reads outsider-authored files like README.md and other repo text (e.g., issues/templates/docs) into the agent for scoring and roast generation.