agentic-awesome-skills-installer

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

• Third-party content exposure detected (high risk: 0.75). The required workflow is installing a skill library from a (potentially user-specified) GitHub repo/URL at runtime, and the installed SKILL.md playbooks (outsider-authored repository content) are then loaded into the agent/tool context as readable text, creating an indirect prompt-injection surface.

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

• Potentially malicious external URL detected (high risk: 0.90). The doc instructs running remote installers that fetch and execute code and skill playbooks at runtime (e.g., npx agentic-awesome-skills and explicit git/curl fetches such as https://github.com/sickn33/agentic-awesome-skills and https://github.com/sickn33/agentic-awesome-skills/archive/refs/tags/v13.13.0.tar.gz), which will download SKILL.md files that are injected into agent context and thus directly control prompts.