nemoclaw-contributor-create-pr

par nvidia

Créer des pull requests GitHub qui suivent le modèle PR NemoClaw, puis surveiller les retours CI et de révision automatisée. Utiliser lorsque l'utilisateur souhaite créer une nouvelle PR,…

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

Create or Update a Pull Request

Publish the requested NemoClaw change and follow required CI and scheduled automated reviews until one unchanged latest PR commit has a complete disposition. A successful push alone is not completion. Preserve the user's scope and any explicit instruction to stop at a draft or withhold approval.

Select the stage

  • Initial publication: read Validation, then Publication. Use the implementation evidence already collected.
  • Update an open PR: first complete the PR follow-up contract. Repair valid in-scope findings through nemoclaw-contributor-implement-issue, then apply the validation and publication references. These are stages of the same authorized task.
  • Inspect CI or review feedback: use the shared follow-up contract. Load publication procedures only if an authorized branch or PR write is needed.
  • Mark a draft ready: use the ready-state requirements in Publication after the latest commit completes follow-up.

Publication requirements

Use the canonical NVIDIA/NemoClaw base, template, sensitive-path policy, and trusted validation surface described in the references. Publish from a clean feature branch. Every published commit must be GitHub Verified; the PR body must contain the configured identity's DCO declaration.

Before declaring the source repository, apply the trusted-base source-repository eligibility gate in Publication. Never open a fork PR when a required CI or E2E path for the candidate is restricted to same-repository pull requests.

Bind branch writes to the declared repository, branch, local commit, and expected remote state. Preserve the atomic prior-state guard, fast-forward ancestry check, and readback requirements. Reconcile inconclusive writes before any permitted bounded retry. Never infer success from a write response alone or weaken the reference's concurrency and recovery rules.

Open code-changing or sensitive-path PRs as drafts. Record available review context without claiming unobserved approval. Do not select labels or request maintainer reviews in this workflow. A workflow or live E2E file change does not require separate approval. Manual live E2E dispatch still requires authority for the selected target. Missing dispatch authority does not block draft publication unless the accepted scope requires that live evidence before publication. When the diff adds, expands, or repairs live E2E evidence, require the implementation handoff to apply Define the Live Contract. When it prunes or relocates live assertions, require the handoff to apply Move or Remove Evidence. Route a handoff that omits any required field back to implementation before publication. Follow GitHub access for access errors and Writing and review for PR text.

Completion

Carry the original objective, accepted scope, deferred scope, tested commit, and validation evidence through repairs and publication. Continue authorized work without asking the user to restate the PR request. Ask only for a decision outside that scope or a required authorization that is still missing.

Report the PR URL, current commit, CI and automated-review results, and any remaining human decision or external blocker. Distinguish pending evidence from completed evaluation. This skill does not grant merge authority.

Plus de skills de nvidia

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…
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,…
aicr-managing-openvex
nvidia
Use when adding, updating, or removing CVE/GHSA suppressions in `.openvex.json` — the OpenVEX document consumed by the daily image vulnerability scan workflow.…
aicr-creating-slide-decks
nvidia
À utiliser lors de la création d'un diaporama HTML autonome ou d'un support visuel pour un concept technique ou un flux de travail (par exemple un demos/*.html) — affiché en plein écran ou…
aicr-creating-guided-demos
nvidia
Génère un script de démonstration guidée interactive (demos/*.sh), en direct ou à son rythme, avec le modèle Frame → Tell → Show → Close. Se déclenche sur « script de démonstration », « guidé…
aicr-analyzing-snapshots
nvidia
À utiliser lors de l'analyse d'un fichier YAML de snapshot AICR, de l'examen de l'état du cluster, de la comparaison des caractéristiques des fournisseurs, de l'extraction d'informations sur la topologie GPU/réseau, ou…