establish-project

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

• Third-party content exposure detected (medium risk: 0.30). In SKILL.md step 1 (create-mode) it dispatches an establish-project brownfield scan that reads repository files as “untrusted input” and writes the resulting .skills/<slug>-scan.md digest containing repo text/citations into the agent’s runtime context for the subsequent grilling interview.