Document and expose benchmark CLI helpers

[chubes4]

Problem

wordpress.bench is registered and smoke-tested, but the generic benchmarking contract is not documented as a first-class WP Codebox capability. Benchmark consumers currently need to infer behavior from command definitions, examples, and smoke tests.

A reusable primitive needs clear docs and CLI helpers around definition, execution, extraction, and comparison while keeping product-specific scoring outside WP Codebox.

Desired shape

Add docs and helper commands such as:

• benchmark contract documentation
• workload authoring examples
• metric and sample contract documentation
• artifact/provenance documentation
• wp-codebox bench run or equivalent recipe wrapper
• wp-codebox bench summarize
• wp-codebox bench compare
• wp-codebox artifacts bench-results
• human table output for scenarios/metrics
• JSON output for automation

Acceptance criteria

• Documentation explains what WP Codebox owns and what callers own.
• Documentation covers PHP, ability, WP-CLI, browser, and future workload types as they land.
• CLI helpers can summarize benchmark results from recipe-run output or artifact bundles.
• CLI helpers expose JSON and human-readable output.
• No product-specific benchmark suites, rewards, graders, or scoring policy are documented as WP Codebox responsibilities.

AI assistance

• AI assistance: Yes
• Tool(s): OpenCode (GPT-5.5)
• Used for: Drafted tracker from repository inspection and benchmark capability gap analysis; Chris remains responsible for prioritization and review.