yy-openclaw-skills-sync-private

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

• Third-party content exposure detected (high risk: 1.00). The skill's SKILL.md explicitly derives and installs GitHub URLs (https://github.com/yanyang1116/skills/tree/main/skills/openclaw/) by fetching remote repositories, meaning the agent will ingest and install user-generated GitHub content that can change its tools/behavior.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill constructs and uses GitHub URLs at runtime (https://github.com/yanyang1116/skills/tree/main/skills/openclaw/) to fetch and reinstall remote agent skills whose code/instructions will execute and control the agent, so this is a high-risk external dependency.