automate tracking SDK feature support

[Andrew Kent (realark)]

Problem

• it's hard to know which SDKs support a specific feature/spec
• it's hard to know the level of parity the various SDKs have with each other

Solution

• add a SDK support section to the end of every spec which contains a markdown table (with optional json attribute section)
• add a parser python script to read these tables
• once parsed, we can do whatever we want with this info. For example, here's a generated CSV of various features and their level of SDK support: https://drive.google.com/file/d/1mFv8bkG2e92S5QmRdrb8t_5vbFKD-BSR/view?usp=sharing
• add a bi-weekly CI job which investigates SDKs and opens a PR with findings for unknown support values

Example

# SDK support

| ID                | Capability        | .NET | Go      | Java | JS      | Python  | Ruby    | Rust |
|-------------------|-------------------|------|---------|------|---------|---------|---------|------|
| ai-span-filtering | AI span filtering | no   | partial | yes  | partial | partial | partial | no   |

```json
{
  "id": "filter-ai-spans",
  "name": "Filter AI spans",
  "category": "Tracing"
}
```.

Notes

• The goal is to have coding agents maintain these tables as they go (basically just tell AGENTS.md to keep the tables up to date when features are added)
• CI jobs check the format and values of the tables to make sure everything is parsable and consistent
• I'm not 100% sure about the accuracy of the various tables. If we decide to go forward with this I'll do another pass for accuracy