adding-a-skill

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

• Third-party content exposure detected (high risk: 0.70). The provided SKILL.md describes how the CLI dynamically discovers and materializes public skill folders from BuilderIO/skills at runtime (loading external skills/*/SKILL.md prose), which is outsider-authored free text that can be injected into the agent’s LLM context when those SKILL.md files are used as instructions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill runtime dynamically materializes the repository referenced as BuilderIO/skills@main (and DEFAULT_SKILLS_SOURCE = "BuilderIO/skills") to read SKILL.md files, so remote content in that repo is fetched at runtime and can directly control agent instructions.