Introduce ScriptTrigger & CommandsTrigger for the Julia plugin

## Context

Part of #313. `plugin-scripts` wants every submodule to expose a `ScriptTrigger` and a `CommandsTrigger`, so a flow can be triggered whenever the underlying script/commands execution completes, with an **exit condition** deciding whether the trigger fires based on the result (e.g. only on failure).

`plugin-script-julia` is missing both. Existing implementations in `plugin-script-go`, `plugin-script-node`, `plugin-script-python`, `plugin-script-ruby`, and `plugin-script-shell` (same package as the task classes, no nested subpackage) can be used as a reference.

## What to implement

### 1. `ScriptTrigger`

- Triggers when `io.kestra.plugin.scripts.julia.Script` execution completes.
- Emits an event containing: command executed, arguments, env (if available), exit code, stdout/stderr (if capture enabled), and contextual metadata (execution id, namespace, flow id).

### 2. `CommandsTrigger`

- Triggers when `io.kestra.plugin.scripts.julia.Commands` execution completes.
- Emits an event containing: commands executed (or the matched one), exit code(s), stdout/stderr if available, and contextual metadata (execution id, namespace, flow id).

## Trigger configuration requirements

Each trigger should accept at least:

- `commandPattern` (string / regex / glob)
- `argumentsPattern` (optional)
- `exitCondition` (string expression — can match on exit code, output, or a literal condition)

## Example YAML

```yaml
triggers:
- id: on_julia_fail
type: io.kestra.plugin.scripts.julia.ScriptTrigger
commandPattern: ".*"
exitCondition: "exitCode != 0"

triggers:
- id: on_julia_commands
type: io.kestra.plugin.scripts.julia.CommandsTrigger
commandPattern: ".*"
exitCondition: "exitCode == 0"
```

Compare the existing ScriptTrigger and CommandsTrigger implementations in plugin-script-go, plugin-script-node, plugin-script-python, plugin-script-ruby, and plugin-script-shell, then inspect the Julia Script and Commands task classes in the same package. Trace how completed executions expose command data, results, and contextual metadata before defining the matching Julia triggers. Done means both triggers support the stated patterns and exit conditions and emit the requested execution details.

Written by the indexing model from the issue text.