chore: add AGENTS.md for AI coding agent guidance

[bryannaegele]

Summary

• Adds AGENTS.md documenting the multi-app rebar3/mix layout, per-app build/test/lint commands, PR title conventions, and other repo-specific context for AI coding agents working in this repo.
• Adds CLAUDE.md as a pointer to AGENTS.md for Claude Code.

Test plan

• N/A (documentation only)

Assisted-by: Claude Sonnet 5

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 17.59%. Comparing base (57a1154) to head (4eaa2c1).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1016   +/-   ##
=======================================
  Coverage   17.59%   17.59%           
=======================================
  Files          24       24           
  Lines         722      722           
=======================================
  Hits          127      127           
  Misses        595      595           

Flag	Coverage Δ
api	17.59% <ø> (ø)
elixir	17.59% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[yordis]

I am not sure if adding such files is a good idea; you will be affecting everyone's workflow. That file injects information into most hardness, and such a file depends on each individual's workflow and style.

Eventually, it may become stale or add information that the hardness doesn't need.

They are plenty of study proving that such practices aren't as useful and actually a net negative.

I would suggest moving the information to the README or documentation directory.

[bryannaegele]

Yeah, we can definitely trim down on the workflow bits with the exception of the commit/PR formatting which is in the community guidelines.

What this is section I think can be updates/fixes to the readme and the top-level notes in Tests, as well.

In the Notes for agents there's a mix of things that can be pushed to the respective apps. And I agree on adding a top-level docs for this repo but I don't want to scaffold all that out in this.

Part of the exercise of this is to surface all the things that need to be fixed up which would cover some of these topics which it is doing. I meant to mark this as a draft. Similar PR will go up in contrib.

lmk what you think but there is a base requirement of adding the file with per the community guidelines. What else (if anything) goes in it is up to us.

[yordis]

CONTRIBUTING.md file and normal documentation of the project which is why I created https://github.com/open-telemetry/opentelemetry-erlang-contrib/tree/main/docs is far more effective, given the fact that we will no be quantifying the usage of such file, most people don't, and the ones that do, discover that are just not that effective.

Every model release and every hardness change make these static files ineffective. These are just proven situations.

[yordis]

Is this a CNCF policy or is there a direct objective to request this?

I am in the opposite camp unless we will be giving credits to built-in hardness and tools as well, models aren't the final answer, and in some cases, the list is long enough, given the fact that, sometimes, we could spawn multiple harnesses to work on thing (included proprietary hardnesses and models).

[bryannaegele]

I linked it above. It's in the community guidelines