Security audit
Security checks for vulnerabilities and agentic risk
This Markdown-only skill explains how to use Gemini CLI for delegated analysis, with some credential and trigger-scope cautions but no hidden execution or persistence.
Install only if you intend to use Gemini CLI for delegated code or document analysis. Be deliberate about which files you include with @path or recursive globs, confirm the Gemini account/model and cost before running commands, and handle GEMINI_API_KEY through safer local secret practices rather than exposing it in shared terminals or persistent shell history.
Skill selects an external model or provider that may use a different account or billing plan than the operator expects. Undisclosed model switches can cause unexpected cost or quota consumption.
gemini -p "@src/**/*.py Summarize these files"
# With specific model
gemini --model gemini-2.5-pro-exp -p "..."
# JSON output
gemini --output-format json -p "..."
The trigger list contains generic terms like "gemini", "cli", "delegation", and "google" without narrowing context or exclusion conditions. These words are broad enough to overlap with ordinary discussion, making activation boundaries unclear.
The documentation instructs users to export a GEMINI_API_KEY directly in the shell without warning about shell history, process environment exposure, or safer secret-management practices. In agent and CLI-heavy environments, this can lead to accidental credential leakage to logs, subprocesses, shared shells, or screenshots, making the issue contextually more dangerous.
No suspicious patterns detected.