FLA MR Readiness Skill

Use this skill before opening a pull request to make sure the change is well-scoped, well-tested, and well-documented.

Pre-flight checklist

1. Read CONTRIBUTING.md

   • Confirm code style, docstring format, commit message conventions.
   • Make sure your branch is up to date with main (or the target branch).

2. Confirm change scope

   • List the files you modified.
   • If the change spans multiple layers (kernel + model + benchmark script), note the dependency chain in the PR description.

3. Check for duplicate work

   • Search open issues and PRs:

     gh pr list --repo fla-org/flash-linear-attention --state open --search "<keywords>"
     

   • If a related PR exists, comment on it rather than opening a competing one.

4. Run dependent tests

   • Find tests affected by your change:

     python scripts/find_dependent_tests.py <changed_file_or_dir>
     

   • Run those tests locally and ensure they pass.
   • If a test is flaky, retry once; if it still fails, explain why in the PR.

5. Performance evidence (if touching kernel code)

   • See fla-nvidia-performance skill for the full evidence requirements.
   • At minimum: before/after benchmark on the same hardware, dense + varlen workloads if applicable, and a summary of any NCU profiling you did.

6. Write PR summary

   • Use the CI-enforced structure below. The source of truth is .github/pull_request_template.md; the check-pr-title workflow rejects bodies that drop the checklist, so never trim it to save space.

7. Code style review

   • Follow CONTRIBUTING.md for Python style, docstrings, comments, and commit prefixes.
   • In tests and public code, use device/platform wrappers from fla.utils (device, device_platform, IS_NVIDIA, IS_NVIDIA_HOPPER, IS_NVIDIA_BLACKWELL, IS_AMD, IS_INTEL) instead of new direct torch.cuda platform checks. Add a small fla.utils helper first when the existing wrappers are not enough.
   • Keep NVIDIA-only profiling commands in performance docs or scripts, not in generic correctness tests.

PR body structure

Follow .github/pull_request_template.md exactly, checklist included:

## Summary
One-paragraph description of what changed and why.

## Test plan
- Unit tests added/modified: `<list>`
- Dependent tests run: `<list>`
- Varlen / CP / model tests: `<yes/no + details>`

## Benchmark / NCU (kernel changes only)
- Hardware: `<e.g., H100>`
- Workload: `<batch, seq_len, dtype>`
- Before: `<throughput or latency>`
- After: `<throughput or latency>`
- Conclusion: `<improvement / neutral / trade-off>`
  (state "neutral" when the change is not performance-related)

## Breaking changes
- None / list any API or behavior changes.

## Checklist

- [x] I have read [CONTRIBUTING.md](../CONTRIBUTING.md) and follow its conventions (code style, docstrings, commit prefixes).
- [x] I have read [AGENTS.md](../AGENTS.md) and, where my change matches its scope, the relevant skill under [.agents/skills](../.agents/skills).
- [x] Dependent tests pass locally or in CI, and new behavior is covered by tests where applicable (tick as N/A for changes with no testable code, e.g. docs-only).
- [x] Kernel changes include same-hardware before/after benchmark numbers, dense + varlen where applicable (tick as N/A when no kernel code changed).
- [ ] This PR is minor/cosmetic-only (typo, formatting, style-only tweaks) — tick only if it is, and justify below.

### If you ticked the "minor" box above

<justification — required when the "minor" box is ticked; otherwise delete this section>

What check-pr-title (.github/workflows/check-pr-title.yml) enforces:

• The first four checklist boxes must always be ticked; each item carries its own N/A reading, so a tick means "considered — done or not applicable" (e.g. benchmark numbers on a docs-only PR).
• The "minor" box is the inverse: leave it unticked for normal PRs. Tick it only when the PR genuinely is a typo/formatting/style-only tweak — then a justification of at least a sentence (≥ 20 non-whitespace characters, HTML comments stripped) under ### If you ticked is required.
• Editing the body re-triggers the check. To edit a PR title/body on this repo, use the REST API, not gh pr edit (see AGENTS.md "Opening PRs").

Important reminders

• Do not put raw performance numbers without context. Always include:

  • workload shape (batch, seq_len, heads, dims, dtype)
  • hardware model
  • benchmark command used
  • before vs after
  • your conclusion

• Do not commit .ncu-rep files or raw profile dumps. Summarize results in the PR body and keep artifacts local.

• No busywork PRs: bundle trivial cleanups into a substantive change; do not open a PR for a single typo unless it is part of a larger fix.