aviz-skills-installer

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

• Third-party content exposure detected (high risk: 0.80). The installer explicitly clones and fetches public content from the AVIZ Skills Library (https://github.com/aviz85/claude-skills-library and https://aviz.github.io/claude-skills-library/skills/{skill-name}.html), which are open, third-party user-maintained pages the agent is instructed to retrieve and read for skill lists and setup, creating exposure to indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). At runtime the installer/git-clone pulls code and skill documentation from https://github.com/aviz/claude-skills-library.git (and the GH Pages at https://aviz.github.io/claude-skills-library/skills/{skill-name}.html), which supplies SKILL.md and scripts that are copied/installed and can inject prompts or execute remote code—so this is a runtime external source that directly controls agent behavior.