deploy-skills-in-antigravity

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

• Third-party content exposure detected (high risk: 0.85). The required workflow ingests outsider-authored free text when it runs npx skills add <repository URL> at runtime, which downloads a third-party repository’s files (including that repo’s SKILL.md/docs) and then the verification step reads the relocated SKILL.md contents into the agent context.

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

• Potentially malicious external URL detected (high risk: 0.90). This skill explicitly runs installation commands at runtime using URLs such as "npx skills add https://github.com/username/repo-name" (invoked via the run_command tool), which fetches and installs remote repository code that will be executed/loaded and thus can directly control agent behavior and is a required dependency.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly instructs the agent to run installation commands, remove and move directories (including rm -rf), relocate files into a global/shared location, and to request elevated permissions if needed—actions that modify and can destructively change the host machine's state.