gemini-review

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

• Third-party content exposure detected (high risk: 0.85). The required workflow includes a GitHub Action that reads diff.txt generated from git diff origin/${{ github.base_ref }}...HEAD (i.e., PR/branch code authored by outsiders) and injects it into the Gemini prompt via $(cat diff.txt) at runtime.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs installing and using the code-review extension at runtime via "gemini extensions install https://github.com/gemini-cli-extensions/code-review", which fetches and installs remote code that can change the agent's /code-review behavior.