nemoclaw-contributor-onboard

par nvidia

Préparez un checkout source NemoClaw pour une contribution conforme via la configuration en une commande et le contrôle de préparation du dépôt. À utiliser lorsqu'un nouveau contributeur demande…

npx skills add https://github.com/nvidia/nemoclaw --skill nemoclaw-contributor-onboard

Onboard a NemoClaw Contributor

Use the repository setup script as the executable source of truth. Do not duplicate its dependency, build, hook, CLI-exposure, or readiness logic in agent commands.

Establish Trust First

  1. Read the root AGENTS.md and CONTRIBUTING.md completely.
  2. Inspect the worktree and current branch without discarding or overwriting existing changes.
  3. Refresh the trusted origin/main reference, then compare the entire checkout/worktree diff against that up-to-date base before executing any checkout-local code. Include staged, unstaged, and untracked files; review lockfiles and all transitively executed source, not only the entry script or package manifests.
  4. If any execution surface differs from trusted origin/main, review the diff and obtain explicit approval before running it.

Route by Intent

  • Readiness only: run ./scripts/dev-setup.sh --doctor and never run setup, repair, CLI exposure, runtime onboarding, or the pinned agent. Use ./scripts/dev-setup.sh --doctor --json when a machine-readable report helps.
  • Initial checkout setup: run ./scripts/dev-setup.sh from the repository root.
  • Explicit repository repair: run ./scripts/dev-setup.sh --repair only when the user asks to repair or retry repository-local setup.
  • CLI exposure: after explicit approval, run ./scripts/dev-setup.sh --expose-cli.
  • Runtime onboarding: after explicit approval, run ./scripts/dev-setup.sh --with-runtime.

The default and repair modes may update repository-local dependencies, builds, hooks, and the root Python environment. They must not create a gateway or sandbox or expose a host-visible nemoclaw command. CLI exposure is an explicit opt-in that may use an npm link or a user-local shim.

Handle User-Controlled Changes

Pause and obtain explicit approval before installing or changing host packages, starting or replacing a container runtime, accepting a license, generating or registering a signing key, changing GitHub state, or changing global Git configuration.

  • Ask for contributor name and email only when the doctor reports that identity is missing.
  • Prefer repository-local Git identity changes when the user approves them.
  • Use gh auth login -h github.com for missing GitHub authentication and pause for browser or device authentication.
  • Let the user choose and register a Git-supported commit-signing key.
  • Follow ../_shared/git-github-hard-stop.md for authentication, authorization, SSH, remote-access, or push failures.
  • Never print tokens, credential values, private keys, or command output that may contain them.
  • Never place secrets in command arguments, generated reports, or tracked files.

After an approved host, account, identity, or signing remediation, rerun npm run dev:doctor or ./scripts/dev-setup.sh --doctor instead of rerunning setup. Reserve setup and --repair for repository-local dependency, build, or hook repair.

Decide on Runtime Onboarding

Ask whether the intended issue requires a live gateway or sandbox after source setup is ready. Documentation work and isolated unit tests normally do not require runtime onboarding.

If runtime validation is required and the user approves it, run:

./scripts/dev-setup.sh --with-runtime

This delegates to interactive nemoclaw onboard and also opts into development CLI exposure as part of that approved flow. Do not preselect third-party software acceptance, inference provider or model, credentials, sandbox name or resources, messaging integrations, or network policy unless the user already supplied those decisions.

Launch the Pinned Agent Only on Request

When the user specifically asks to use the repository-pinned coding agent, run the doctor first. If readiness fails, report the remediation and obtain authorization for the matching setup or repair mode rather than mutating the checkout automatically. When readiness passes, run:

npm run agent

Pass user-supplied Pi arguments after --. Do not install or invoke a global Pi binary.

Hand Off to the Contributor Lifecycle

This workflow ends after the doctor reports that every contributor-readiness check passes and every approved CLI-exposure, runtime-onboarding, or pinned-agent step is complete. Name the workflow that owns the next stage instead of restating its rules:

  • nemoclaw-contributor-plan-issue refines an issue into capability slices.
  • nemoclaw-contributor-implement-issue implements a slice and owns its tests.
  • nemoclaw-contributor-create-pr publishes the change and owns the branch-state, commit-verification, DCO, template, and review follow-up requirements.

Do not create a branch, commit, push, or PR unless the user's request includes that action.

Report the Result

Summarize repository-local setup performed, doctor status, user-controlled remediations still needed, whether CLI exposure or runtime onboarding ran, and the next safe contributor action.

Plus de skills de nvidia

compileiq-debug
nvidia
Utilisez quand quelque chose ne va pas : Search() bloque, toutes les évaluations retournent INVALID_SCORE, les scores ne s'améliorent pas, chaque configuration retourne le même nombre, erreurs ptxas…
create-github-pr
nvidia
Créer des pull requests GitHub en utilisant l'interface en ligne de commande gh. Utiliser lorsque l'utilisateur souhaite créer une nouvelle PR, soumettre du code pour révision, ou ouvrir une pull request. Mots-clés de déclenchement -…
nemoclaw-maintainer-cross-issue-sweep
nvidia
Analyse les autres problèmes ouverts pour trouver ceux qu’une PR donnée pourrait également corriger ou casser accidentellement. Génère des opportunités de correctifs adjacents et des risques de contradiction avec fichier:ligne…
fhir-basics
nvidia
Apprend aux agents comment fonctionnent les API FHIR R4, quelles ressources sont disponibles, comment les interroger avec des paramètres de recherche, et comment analyser correctement tous les formats de réponse…
compileiq-validate-result
nvidia
Utiliser APRÈS qu'une recherche soit terminée et AVANT de réclamer un accélérateur ou d'expédier un ACF. Charge le CSV dump_results, extrait les K meilleurs candidats (mono-objectif)…
changelog-audit
nvidia
Auditer le CHANGELOG.md de Warp avant une publication : récupérer les entrées perdues, trier par impact utilisateur, affiner le langage des entrées, ajuster les retours à la ligne et (en mode branche de publication) mettre à jour la comparaison…
maintain-dynamic-plugins
nvidia
Maintenir les chargeurs de plugins dynamiques NeMo Relay, les manifestes, les SDK natifs Rust, le protocole worker gRPC, le SDK worker Python, la documentation, les tests et la couverture du workflow de publication
dgx-diagnose
nvidia
Diagnostiquer les problèmes courants du DGX Station GB300 — plantages CUDA, ciblage incorrect du GPU, bugs de conteneur vLLM/SGLang, problèmes d'état MIG, erreurs NVLink/Fabric Manager,…