create-parser-package

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

• Third-party content exposure detected (high risk: 0.70). The SKILL.md workflow's Step 1 explicitly instructs the agent to "Search the web for additional samples or documentation when the format is unfamiliar," meaning the agent will fetch and read open/public web content (which may be untrusted/user-generated) and use that information to shape parsing decisions and package code.