remove-ai-patterns

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow reads arbitrary user-provided file text at scripts/detect.js → fs.readFileSync(file, "utf8") and then runs the detector on that content via upstream/detector/patterns.js, creating an indirect prompt-injection surface if outsiders can supply/queue those files.