scrape-codegen

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via the runtime-loaded spec snapshots: /scrape-codegen-analyze {spec_path}/spec.json {spec_path}/values/{page_id}.json and the corresponding {spec_path}/pages/{page_id}/{html_variant}.html are read from {spec_path} (which is not guaranteed to be user-authored) and then provided to the analysis agent, which uses them to generate code—this includes arbitrary page/values text from outside the operating user.