openwork-core

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

• Third-party content exposure detected (high risk: 0.80). This skill explicitly instructs cloning and pulling a public GitHub repository (https://github.com/anomalyco/opencode via git clone and git -C vendor/opencode pull) and tells the agent to use vendor/opencode for self-reference, meaning it will fetch and read untrusted third-party content that could contain injected instructions.