architect

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

• Third-party content exposure detected (low risk: 0.10). GENERATION reads only first-party skill-controlled procedure text from ${CLAUDE_PLUGIN_ROOT}/questions/phase-4-generate.md and writes/validates the resulting blueprint, while any browsing/external reading (via agent-browser) occurs only when the user explicitly shares a reference site to read during ARCHITECTURE.