Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

doc: add infra/aws-gpu Terraform bootstrap and AGENTS.md runbook

Open
  • Is something wrong?
#1,419 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 2 days

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
20/100
Issue type
Feature
Clarity
Clearly specified
Activity status
Active
Tech stack
aws, bash, rust, terraform
Domain
cloud, devops, documentation, infrastructure

Research direction

The original work is described in the issue, and the repository’s AGENTS.md documents how to transplant PR 136 from NVlabs/cutile-rs, including mapping moved paths. Start by reading that section and checking whether the archived repository’s pull/136/head can still be fetched; done means the work is recreated against NVIDIA/cuda-rust with authorship and sign-off preserved. The original scope spans infra/aws-gpu Terraform, scripts, guardrails, tests, and a runbook.

Written by the indexing model from the issue text.

Description

cutile-rs post-migration-pr

Tracks pull request NVlabs/cutile-rs#136 (https://github.com/NVlabs/cutile-rs/pull/136) by @kevinkurek, opened 2026-05-08: doc: add infra/aws-gpu Terraform bootstrap and AGENTS.md runbook. The branch was rebased onto cutile-rs main on 2026-10-07 and merges cleanly.

@kevinkurek: that repository has moved here and is being archived, so the pull request is closed there, not rejected. If you want to continue the work, please recreate it against NVIDIA/cuda-rust. AGENTS.md in NVIDIA/cuda-rust has a "Transplanting a cutile-rs pull request" section that lets an agent do the replay (it fetches pull/136/head from NVlabs/cutile-rs, keeps your authorship and sign-off, and maps the moved paths).

Original description:

Summary

Video of PR explanation: https://www.tella.tv/video/gpuforcutile-1-bjpl

Adds infra/aws-gpu/: a self-contained Terraform stack + helper scripts that provision a temporary g5.xlarge (NVIDIA A10G, sm_86, Ubuntu 24.04 DLAMI), install CUDA 13.2 / LLVM 21 / Rust on it, clone cutile-rs, and run `cargo run -p cutile-examples --example add_refs` as a smoke test.

Aimed at contributors who don't already own a CUDA 13.2 + sm_80+ workstation. Not part of the regular build — `scripts/run_all.sh` is unaffected.

Contents

  • Terraform (`main.tf`, `variables.tf`, `outputs.tf`, `versions.tf`) with Ubuntu 24.04 DLAMI SSM lookup, auto-detected /32 SSH ingress, fresh per-run RSA-4096 key (mode 0600, gitignored), and `instance_initiated_shutdown_behavior=terminate`.
  • `up.sh` / `down.sh` / `ssh.sh` / `run-cutile-smoke.sh` wrappers. Smoke script accepts `CUTILE_REPO` / `CUTILE_REF` / `CUTILE_EXAMPLE` env vars to test a fork or branch.
  • `bootstrap-iam.sh` + `iam-policy.json`: idempotent one-shot least-privilege IAM user setup. Captures the new access key in shell variables and writes via `aws configure set` without echoing the secret. Supports `--dry-run` and `--destroy` (with self-destruct fallback).
  • `AGENTS.md`: runbook for LLM coding agents driving this flow on behalf of a contributor (persona triage, quota check, no-secret-on-stdout contract).
  • `.guardrails/secret-guard.sh`: optional Claude Code PreToolUse hook denying common credential-leak commands. Behavior pinned by a 26-case test harness.
  • `.github/workflows/aws-gpu-guardrails.yml`: runs the secret-guard test suite. Path-filtered to `.guardrails/**` so it doesn't interact with existing CI.

Test plan

  • `bash scripts/run_all.sh` is unaffected — no workspace crates touched.
  • `terraform fmt -check` clean; `bash -n` clean on all scripts.
  • Guardrail tests: 26/26 pass (`bash infra/aws-gpu/.guardrails/secret-guard.test.sh`).
  • Tested end-to-end on a fresh AWS account: `bootstrap-iam.sh` → `up.sh` → `run-cutile-smoke.sh` → `down.sh`. `add_refs` example built and ran on the A10G; teardown clean.
  • No secrets, `.pem`, `tfstate`, or real `tfvars` committed (verified by `infra/aws-gpu/.gitignore`).

Happy to break this into smaller PRs (e.g. infra-only first, runbook second, guardrails optional) if that fits your review process better.

Dominant language
Rust
Stars
3.7k
Forks
312
Avg merge
6d 5h
Merged PRs (30d)
68

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from NVIDIA/cuda-rust

All issues in NVIDIA/cuda-rust

Similar issues

More Rust issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.