Run one-shot prompts against Gemini models from the command line.
npx skills add https://github.com/srgaba/open-claw --skill gemini-srgaba
Please help me install this Agent Skill. Skill: gemini Source: https://github.com/srgaba/open-claw/tree/main/project/skills/gemini Command: npx skills add https://github.com/srgaba/open-claw --skill gemini-srgaba
--model and request structured output with --output-format json. - Extension Management: List and manage CLI extensions via gemini --list-extensions and gemini extensions. - Use Case: Pipe a log file summary request like gemini "Summarize the errors in this log: ..." into a shell script to automate routine analysis. ## Quick Start Ask the assistant to run gemini with a one-shot prompt such as "Answer this question: what is a reverse proxy" and return the response.High-intent search queries and answers about installing and using this skill.
Pass the prompt as a positional argument, for example gemini "Answer this question...". This runs a single non-interactive request and prints the response to the terminal.
Add the --output-format json flag before your prompt, such as gemini --output-format json "Return JSON". The response is then emitted as structured JSON suitable for scripting.
Use the --model flag followed by the model name, for example gemini --model <name> "Prompt...". This overrides the default model for that single invocation.
The CLI requires a login flow on first use. Run gemini once interactively and follow the prompts to authenticate, after which one-shot commands work normally.
The --yolo flag bypasses safety confirmations and is discouraged. Prefer default behavior so potentially destructive actions require explicit approval.
trpc-groupRuns Gemini CLI in one-shot mode for Q&A, summaries, and text generation.
Parallel-7Run headless gemini-cli prompts with JSON or text outputs.
AradotsoRoute prompts across OpenAI and Anthropic models to minimize cost.
namojoBridge Gemini CLI and API workflows into LangGraph nodes with safety checks.