topiary

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

• Third-party content exposure detected (high risk: 0.80). This skill's SKILL.md and templates/languages.ncl explicitly instruct registering external grammar.source.git URLs (e.g., "https://github.com/...") and creating/using .scm query files under "Custom Grammar Sources" / "Adding Custom Languages", so the agent may fetch and execute untrusted, third‑party Tree-sitter grammars and query files that directly control formatter behavior.