Add Explain Vulnerability AI action to GraphQL API

What does this MR do and why?

This change introduces an AI completion for generating an explanation of a vulnerability using an intelligent chat service.

#406633 (closed)

Screenshots or screen recordings

How to set up and validate locally

If you do not already have an OpenAI API key you will need to set one up

• Request a key via https://forms.gle/LPLXtgVjuGL4JCYFA
• Follow the sign up process in the resulting invite email
• Once signed in, go to 'View API keys' in the top right user menu
• Select 'Create Secret Key' and copy the generated key
• In the rails console:

  Gitlab::CurrentSettings.update(openai_api_key: "<your-key>")

Following that:

1. Enable feature flags (consult https://gitlab.com/gitlab-org/gitlab/-/issues/405161 for the latest list):

   Feature.enable(:explain_vulnerability)
   Feature.enable(:openai_experimentation)
   Feature.enable(:ai_experimentation_api)

2. Import a test project:
3. Pick a specific SAST vulnerability with a location: { file: "", start_line: 0 }
4. Execute from a rails console:

   [1] pry(main)> ::Llm::CompletionWorker.new.perform(<user_id>, <vulnerability_id>, "Vulnerability", :explain_vulnerability, {})
   [ActionCable] Broadcasting to graphql-event::aiCompletionResponse:resourceId:Z2lkOi8vZ2l0bGFiL1Z1bG5lcmFiaWxpdHkvODY2NDY:userId:Z2lkOi8vZ2l0bGFiL1VzZXIvMQ: "{\"id\":\"acc55795-b110-468e-a5b8-c2eafb2b9b91\",\"model_name\":\"Vulnerability\",\"response_body\":\"As an AI language model, I do not have access to the actual code or context of the vulnerability mentioned in the README.md file. Therefore, I cannot provide a detailed explanation of how the vu...

MR acceptance checklist

This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.

• I have evaluated the MR acceptance checklist for this MR.