📝 Add comprehensive AGENTS.md for Browser SDK, scripts, and E2E testing

[thomas-lebeau]

Motivation

Makes working with AI a bit more efficient. Most of these instruction are things that AI can already infer, but it has to read lot of file every time so it's slow and consume lots of tokens (and it can get things wrong).

Adding AGENTS.md file help AI agents work in the repository. They were in my unstaged changes since a few days/weeks and I've found them useful. They are by no mean perfect nor complete but it's a start and we can tweak them and add more (e.g. developer extension, better unit test best practice,...) with time and experience.

AGENTS.md are compatible with Cursor and most other AI tools. Claude code read instruction from CLAUDE.md files, but can reference other file such as @./AGENTS.md.

Changes

• Added root level AGENTS.md.
• Added test/e22/AGENTS.md for working on e2e tests.
• Added scripts/AGENTS.md for working on scripts.
• Added corresponding CLAUDE.md files for claude code compatibility.

note: I've used AI to bootstrap these files and refined them manually.

Test instructions

Testing this is probably easier on the long run. Let's merge it and give it a try, we can fix it when we find improvements.

Checklist

• Tested locally
• Tested on staging
• Added unit tests for this change.
• Added e2e/integration tests for this change.

[datadog-datadog-prod-us1]

✅ Tests

🎉 All green!

❄️ No new flaky tests detected
🧪 All tests passed

🎯 Code Coverage
• Patch Coverage: 100.00%
• Overall Coverage: 77.25% (-0.08%)

View detailed report

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: a0558de | Docs | Datadog PR Page | Was this helpful? Give us feedback!}

[cit-pr-commenter-54b7da]

Bundles Sizes Evolution

📦 Bundle Name	Base Size	Local Size	𝚫	𝚫%	Status
Rum	165.78 KiB	165.78 KiB	0 B	0.00%	✅
Rum Profiler	4.33 KiB	4.33 KiB	0 B	0.00%	✅
Rum Recorder	24.48 KiB	24.48 KiB	0 B	0.00%	✅
Logs	56.21 KiB	56.21 KiB	0 B	0.00%	✅
Flagging	944 B	944 B	0 B	0.00%	✅
Rum Slim	123.05 KiB	123.05 KiB	0 B	0.00%	✅
Worker	23.63 KiB	23.63 KiB	0 B	0.00%	✅

🚀 CPU Performance

Action Name	Base CPU Time (ms)	Local CPU Time (ms)	𝚫%
RUM - add global context	0.0039	0.0038	-2.56%
RUM - add action	0.0129	0.013	+0.78%
RUM - add error	0.0123	0.0124	+0.81%
RUM - add timing	0.0025	0.0024	-4.00%
RUM - start view	0.003	0.0029	-3.33%
RUM - start/stop session replay recording	0.0006	0.0006	0.00%
Logs - log message	0.0143	0.0141	-1.40%

🧠 Memory Performance

Action Name	Base Memory Consumption	Local Memory Consumption	𝚫
RUM - add global context	27.89 KiB	27.22 KiB	-690 B
RUM - add action	50.08 KiB	50.97 KiB	+914 B
RUM - add timing	27.36 KiB	25.86 KiB	-1.50 KiB
RUM - add error	55.79 KiB	53.36 KiB	-2.44 KiB
RUM - start/stop session replay recording	24.97 KiB	24.47 KiB	-516 B
RUM - start view	427.11 KiB	425.20 KiB	-1.92 KiB
Logs - log message	45.38 KiB	94.23 KiB	+48.85 KiB

🔗 RealWorld

[BeltranBulbarellaDD]

IMO, maybe add format, lint and typecheck? To test changes is interesting to run linting and typecheck. WDYT?

[BeltranBulbarellaDD]

I like it because it serves as internal documentation for the SDK too.