swift-review

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

• Third-party content exposure detected (high risk: 0.75). The skill ingests the user-selected PR diff / file contents at runtime (e.g., via gh pr diff, git diff, or reading a file), and those diffs/files are commonly authored by outsiders, so their free-text code/comments can be included in the LLM context.