adverse-selection-prior

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow (see "Step 4: Score proposer info asymmetry" in SKILL.md and numerous lines in resources/methodology.md) explicitly instructs the agent to check and ingest public news and public signals (e.g., recent news feeds, beat-reporters, roster/public signals), which are untrusted third-party sources and are used to adjust the prior and recommended action, creating a clear avenue for indirect prompt injection.