collaborative-delivery

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

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via the adversarial_anchor runner path: scripts/panel_round.py runs .agents/skills/claude-runner/scripts/run_claude.py, captures its stdout/stderr, and the runner wrapper can feed that (outsider-authored model output) back into the host LLM context for subsequent phases; this is not user-chosen content.