better-planning-build

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow includes reviewing a “loop PR” (e.g., PR #88) and running a Skeptic pass over that loop-produced code, which necessarily ingests outsider-authored free text such as the PR’s description/comments and issue/PR body content into the agent’s LLM context via the PR review/retrieval path.