skill-creator

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow creates tasks and runs skill evals (via scripts/prepare_eval.py and executor/grader agents) which consume eval prompts and test files from evals.json. Since external users or untrusted sources could author these eval inputs/prompts, this ingestion presents an indirect prompt injection risk.