Gemini 3 Pro Text & Vision

Usage

Run from the user's current working directory (do NOT cd into the skill directory):

Text generation:

uv run scripts/query_gemini.py --prompt "your question" [--system-prompt "instructions"] [--output "response.md"] [--api-key KEY]

Image analysis (vision):

uv run scripts/query_gemini.py --prompt "describe this image" --image "path/to/image.png" [--output "analysis.md"] [--api-key KEY]

Parameters

• --prompt / -p (required): Question, task, or instructions

• --image / -i : Path to image for vision analysis (jpg, png, webp, gif)

• --system-prompt / -s : System instructions to set context/persona

• --output / -o : Save response to file (always prints to stdout too)

• --api-key / -k : API key (overrides GRSAI_API_KEY env var)

API Key

Provide via --api-key argument or set GRSAI_API_KEY environment variable.

Troubleshooting

Symptom Resolution

Error: No API key provided.

Set GRSAI_API_KEY env var or pass --api-key

Error: Image not found:

Wrong path or unreadable file; verify the path exists

HTTP 401 Wrong or expired API key

Unexpected response format

API error; check prompt for policy violations

uv: command not found

Install: curl -LsSf https://astral.sh/uv/install.sh | sh , then restart terminal

For transient errors (HTTP 429, network timeouts), the script retries automatically up to 3 times. If retries are exhausted, surface the error to the user.

Output

Print the response to stdout. Optionally save to file with --output . Do not pipe output unless the user asks.

Examples

uv run scripts/query_gemini.py --prompt "Explain the difference between TCP and UDP in simple terms"

uv run scripts/query_gemini.py --prompt "What objects are in this image?" --image "photo.jpg" --output "analysis.md"