WP-CLI Reference
Gratis AI Agent ships a wp gratis-ai-agent command family for benchmarking the agent, managing abilities, and querying agent status from the command line. All commands require WP-CLI 2.0 or higher.
Installation
The WP-CLI commands are registered automatically when the plugin is active. Verify with:
wp gratis-ai-agent --info
wp gratis-ai-agent benchmark
Runs the Agent Capabilities benchmark suite — a set of complex, multi-step prompts that exercise the full ability surface. Use this to evaluate model performance, compare AI providers, or validate ability packs before deploying to production.
Synopsis
wp gratis-ai-agent benchmark [--question=<id>] [--provider=<provider>] [--model=<model>] [--output=<format>] [--save]
Options
| Option | Description |
|---|---|
--question=<id> | Run a single benchmark question by ID instead of the full suite |
--provider=<provider> | Override the configured AI provider for this run (e.g. anthropic, openai) |
--model=<model> | Override the configured model for this run (e.g. claude-opus-4-6, gpt-4o) |
--output=<format> | Output format: table (default), json, csv |
--save | Save the benchmark results to the database for historical comparison |
Examples
Run the full benchmark suite with the current provider and model:
wp gratis-ai-agent benchmark
Run a single question (q-restaurant-website) and output as JSON:
wp gratis-ai-agent benchmark --question=q-restaurant-website --output=json
Run against a specific model and save results:
wp gratis-ai-agent benchmark --model=claude-opus-4-6 --save