test(benchmark): define benchmark contribution protocol

[michael-johnston]

The goal is to define the multi-stage process for submitting and validating a models benchmarking requirements and for executing the benchmark.

The process must satisfy the requirements in #3

✍️ Scope of the Protocol Document

The document will be structured around the three core stages of a contribution. For each stage, it will describe the user/admin actions and map to the relevant requirements from #3.

Stage 1: The Pull Request (PR) Submission

This stage describes the contributor's responsibility when opening a PR - What do they need to include to have PR been accepted for merging from benchmark.

This will include the various files and their formats etc.

Stage 2: The Pull Request Review & Validation

This stage describes the automated and/or manual checks that will validate that the artefacts provided in Stage 1 meet specific requirements

Stage 3: Post-Merge - Benchmark Execution

This stage describes what happens after a contribution has been approved and merged - again what is automated and what is manual.

---

✅ Definition of Done:

• A document (perhaps part of existing CONTRIBUTING_MODEL.md) that clearly defines what the three stages of the contribution process (Submission, Validation, Post-Merge) - what is specifically required as input to each, and what is required to happen with those inputs, what is the criteria to move between stages.