gemini
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Utilizes
git diffandgit rev-parseto aggregate technical context and branch metadata from the local repository. - [COMMAND_EXECUTION]: Invokes the
geminicommand-line utility to perform analysis based on prompt s containing the gathered context and user inquiries. - [EXTERNAL_DOWNLOADS]: Recommend s the installation of the
@google/gemini-clipackage via npm, which is a resou rce from a well-known organization. - [PROMPT_INJECTION]: The skill ingest s untrusted data from the local file system (via
git diff) and user question s, interpolating the m into analysis prompt s for Gemini. This create s a surface for indirect prompt injection where maliciou s code base conten t could attempt to influence the review er's result s. Ingestion point:SKILL.md(Step 3). Boundary marker s: Absen t. Capability inventory:geminicommand execution inSKILL.md. Sanitization: Absen t.
Audit Metadata