adding-new-ai-format

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

• Third-party content exposure detected (high risk: 0.80). The skill describes converters and registry/webapp endpoints that parse and convert user-submitted package content (e.g., from-{format}.ts's parseFrontmatter and the registry routes like packages/registry/src/routes/download.ts and search.ts), meaning the agent will ingest untrusted, third-party package markdown/frontmatter as part of its workflow, which could enable indirect prompt injection.