ax-eval runs coding agents against reproducible CLI scenarios and writes evaluation profiles.

Use it to improve CLI help, docs, and AGENTS.md guidance by seeing whether agents complete the task, how many wrong turns they take, what they spend, and which artifacts changed.

Common commands:
  ax-eval scenarios
  ax-eval template scenario > ax-eval-fixtures/my_scenario.yaml
  ax-eval validate --scenario ax-eval-fixtures/my_scenario.yaml
  ax-eval guidance list
  ax-eval guidance start
  AX_EVAL_ENABLED=1 ax-eval discover mytool --tool opencode
  AX_EVAL_ENABLED=1 ax-eval run --scenario my_scenario --tool opencode
  ax-eval show <run-id>

Use `ax-eval template <kind>` for copyable schema examples.

Usage: ax-eval [OPTIONS] <COMMAND>

Commands:
  run        Run a test scenario
  discover   Discover how well a target CLI describes itself to LLM agents
  scenarios  List available scenarios
  show       Show details for a saved run
  clean      Clean cache and legacy transcript artifacts
  guidance   Show guidance for building LLM-usable tools and docs
  validate   Validate scenario YAML without running
  template   Print copyable scenario, config, rubric, and script templates
  help       Print this message or the help of the given subcommand(s)

Options:
  -v, --verbose
          Enable verbose output (or set RUST_LOG for fine-grained control)

  -h, --help
          Print help (see a summary with '-h')

  -V, --version
          Print version

Common commands:
  ax-eval scenarios
  ax-eval template scenario > ax-eval-fixtures/my_scenario.yaml
  ax-eval template config > ax-eval-config.toml
  ax-eval validate --scenario ax-eval-fixtures/my_scenario.yaml
  ax-eval guidance start
  AX_EVAL_ENABLED=1 ax-eval discover mytool --tool opencode
  AX_EVAL_ENABLED=1 ax-eval run --scenario my_scenario --tool opencode
  ax-eval show <run-id>
