fix: addresses stack too deep error

[arthurgousset]

Description

This PR aims to address the stack too deep issue described here: #3373

It's a fix, I imagine it could be refactored and improved, but @priyankc and I thought we'd share it here.

We'd love feedback and suggestions on how to improve the PR.

Test

Reproduce the bug here #3373

1. Add a file here modules/hello/plans/test.pp

   plan hello::test {
     fail_plan(
       'That should fail properly',
       'test/failure',
       foo => Error('foo'),
     )
   }

2. Install and run the test plan (see stack too deep error)

   $ bundle install --path .bundle --without test
   $ bundle exec bolt plan run hello::test --format json
   
    bundler: failed to load command: bolt (/Users/arthur/Documents/hyperdrive-eng/customers/forks/bolt/.bundle/ruby/3.2.0/bin/bolt)
    /Users/arthur/Documents/hyperdrive-eng/customers/forks/bolt/.bundle/ruby/3.2.0/gems/puppet-8.10.0-universal-darwin/lib/puppet/pops/types/types.rb:3009:in `block in instance?': stack level too deep (SystemStackError)

3. Apply changes in this PR

4. Run test plan again (resolved stack too deep error)

   $ bundle exec bolt plan run hello::test --format json
   {"kind":"test/failure","msg":"That should fail properly","details":{"foo":{"msg":"foo"},"file":"/Users/arthur/Documents/hyperdrive-eng/customers/forks/bolt/modules/hello/plans/test.pp","line":2,"column":3}}

[CLAassistant]

All committers have signed the CLA.