extension-builder

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

• Third-party content exposure detected (high risk: 0.75). The required workflow explicitly instructs treating Pi source text as untrusted input and then extracting/reading referenced Pi pi.extensions/pi.skills files (including SKILL.md), which are outsider-authored free-form text that can be ingested into the agent’s LLM context during adaptation.