The problem

CLI help text is often buried in terminal output, making it hard for new users to discover features or understand complex flag interactions. Manually maintaining README command references and man pages is repetitive and leads to outdated documentation.

What it does

• Harvests full command trees by recursively executing help flags across all subcommands.
• Parses raw output into structured tables covering flags, aliases, types, and defaults.
• Generates 3 to 5 verified runnable examples ranging from basic usage to advanced piped commands.
• Documents environment variables, configuration file discovery, and exit codes.
• Validates all generated snippets by running them in a sandbox to ensure zero broken examples.

Frameworks & tools

Works with any CLI tool providing --help or -h output, including those built with argparse, Click, Cobra, or Clap.

Why this beats prompting it yourself

Basic prompts just reformat help text; this skill actively validates the CLI behavior. It executes the tool to confirm exit codes and flag logic, ensuring your documentation reflects how the software actually runs, not just what the help string claims.

Use cases

• Generating a "Command Reference" section for a project README.
• Building structured man pages or documentation site content from raw binaries.
• Creating onboarding guides for internal DevOps or CLI-heavy tools.
• Audit and verify that help text matches actual CLI implementation.