agent-rules

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

• Third-party content exposure detected (high risk: 0.75). At runtime, promote loads and ingests the repo’s docs/reflection-log/[0-9]*.md entries (free-form “What happened/What to do differently” text authored by whoever recorded the failures), which are outsider-authored content relative to the operating user’s chosen inputs.