Get the output of a job step

## Usage

`circleci job output get <job-id> [flags]`

## Arguments

`<job-id>` is the UUID of the job whose step output to fetch. Job UUIDs
are shown in the output of `circleci workflow get` and `circleci job get`.
Use `--step-num` to select which step's output to read.

## Flags

| Flag              | Description                                                                                    |
| ----------------- | ---------------------------------------------------------------------------------------------- |
| `--condensed`     | Fetch error-relevant lines only, filtered server-side (experimental)                           |
| `--execution int` | Parallel execution index to read output from                                                   |
| `--step-num int`  | Step number whose output to fetch (required)                                                   |
| `--strip-ansi`    | Force (or with =false, disable) ANSI stripping; defaults to stripping only when not a terminal |

Global flags: `-c, --config`, `--debug`, `--no-color`, `-q, --quiet` — see `circleci --help`.

## Examples

- Get the output of step 3 in a job: 
  `circleci job output get 8e50c384-0083-43d0-bc8f-93f0db589d6b --step-num 3`
- Get the output of step 3 from the second parallel execution: 
  `circleci job output get 8e50c384-0083-43d0-bc8f-93f0db589d6b --step-num 3 --execution 1`
- Pipe the output to a file: 
  `circleci job output get 8e50c384-0083-43d0-bc8f-93f0db589d6b --step-num 3 > step.log`
- Pipe condensed output to an AI CLI tool: 
  `circleci job output get 8e50c384-0083-43d0-bc8f-93f0db589d6b --step-num 3 --condensed | llm "Why did this fail?"`

## Details

Fetch the raw stdout and stderr of a single step within a job. Both streams
are fetched in parallel and printed together, stdout first.

Output passes through unchanged on a terminal and is rendered to plain text
when redirected (ANSI stripped, progress redraws collapsed). --strip-ansi
forces either behaviour.

