What does this skill do, and when should you use it?
This skill guides installation of NeMo Relay for the CLI, Python, Node.js, Rust, OpenClaw, Hermes, LangChain, LangGraph, and Deep Agents. It first inspects the target manifest, environment, operating system, architecture, and existing installation before selecting and running an installation path. Installation is kept separate from runtime configuration, plugin files, scopes, middleware, and first-application examples. For Codex Desktop, it adds safeguards for a provider-filtering issue that can temporarily hide threads after persistent installation.
Inspects project manifests, environment, operating system, architecture, and existing installations; selects a CLI, maintained integration, language-package, or source-checkout path based on the requested outcome; reads the reference for that path; shows the exact installation command; runs installation and basic availability checks; uses nemo-relay doctor for CLI configuration or readiness issues, including JSON and persistent-plugin modes; reports what was installed, where it was installed, and the verification result.
- A developer who wants to try Relay temporarily or run the local gateway and needs the CLI path selected.
- A Python, Node.js, or Rust application owner whose code directly controls tool or model call sites and needs the matching language package.
- A LangChain, LangGraph, or Deep Agents user who needs a maintained framework integration.
- An OpenClaw or Hermes user who needs installation guidance for the existing harness.
- A Codex Desktop user considering persistent Codex provider installation and needing continuity safeguards before proceeding.
How do you install this skill?
- Before running a remote installer, pin and confirm a specific release, review the script, and assess reachability of GitHub/raw.githubusercontent.com in the target network environment.
- Persistent Codex Desktop installation changes global configuration and may hide threads; complete the recovery-file, dry-run, and explicit-confirmation gate, and do not restart before providing the recovery-file path.
- The benchmark pass rates are project-supplied evidence and do not constitute execution or independent reproduction in this review.
- Shell / CLI
- Network access
- Local filesystem
NeMo Relay CLIPythonNode.jsRust
Install this skill into an Agent Skills-compatible client with: npx skills add nvidia/skills --skill nemo-relay-install --yes. The collection can be installed without cloning the repository or copying folders manually. Then ask the agent to install NeMo Relay for the desired target; if no target is named, it asks one short clarification question. The referenced installation files are not included in the supplied material, so their specific NeMo Relay commands are not reproduced here.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/NVIDIA/skills.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/skills/nemo-relay-install" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Install NeMo Relay for my Python application.
- I want to try the NeMo Relay CLI temporarily.
Try a prompt such as “Install NeMo Relay for my Python application.” or “I want to try the NeMo Relay CLI temporarily.” The agent should inspect the target and existing installation, show the installation command, perform installation and a basic availability check, and stop. Runtime configuration, plugins.toml, scopes, middleware, and first-app examples are outside this skill’s installation scope.
What are this skill's strengths and limitations?
- Covers the CLI, Python, Node.js, Rust, and several maintained framework integrations.
- Clearly separates installation from runtime configuration and first use.
- Includes a continuity-protection workflow for persistent Codex Desktop installation.
- Provides scoped nemo-relay doctor guidance for CLI configuration and readiness problems.
- The supplied material omits the referenced installation files, so it does not provide command-level installation details for each path.
- It stops after installation and basic availability checks and does not create scopes, register middleware, or build a first application.
- No specific operating-system, architecture, or version compatibility matrix is provided.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| NeMo Relay Installation Guide this page ✓ NVIDIA · Official | 62 · Recommended | ★ 3.8k | today | Apache-2.0 |
| NeMo Relay Quick Start ✓ NVIDIA · Official | 61 · Recommended | ★ 3.8k | today | Apache-2.0 |
| NeMo Relay Typed Wrappers & Codecs ✓ NVIDIA · Official | 55 · Use with care | ★ 3.8k | today | Apache-2.0 |
| NeMo Relay Migration Assistant ✓ NVIDIA · Official | 57 · Use with care | ★ 3.8k | today | Apache-2.0 |
| debug-skill: Interactive Debugger for Coding Agents | 56 · Use with care | ★ 335 | 6mo ago | MIT |
How did FollowSkills review this skill?
The skill explicitly limits scope, shows commands before execution, requires review of remote installers, and provides preview, confirmation, recovery-file, doctor, uninstall, and rollback guidance for persistent Codex changes. Deductions apply because remote scripts, global Codex configuration, and package-manager writes still create external effects, while credential, network data-flow, and dependency-security guidance is incomplete.
Installation routing, package-manager preservation, version checks, and doctor-based failure diagnosis are reasonably explicit, with separate verification rules for CLI and language packages. Deductions apply because this is a static review with no executable reproduction; BENCHMARK.md contains self-reported results rather than verifiable CI coverage, and platform, version, and command-failure cases remain incomplete.
The skill targets CLI, Python, Node.js, Rust, OpenClaw, Hermes, and maintained framework integrations, with generally clear triggers, exclusions, and post-install handoff boundaries. Deductions apply because Chinese-language support is unspecified and core paths depend on GitHub, raw.githubusercontent.com, and NVIDIA documentation without mainland-China reachability or fallback guidance.
The documentation is well organized and includes references, license, author, version information, an evaluation report, and a recovery asset, with clear scope boundaries. Deductions apply because static validation records missing recommended Instructions and Examples sections, the author format fails the checker, and maintenance ownership, changelog practice, and formal update procedures are not sufficiently explicit.
The skill can route users to appropriate installation commands and deliberately stops at installation and basic availability checks instead of configuring runtime behavior prematurely. Deductions apply because static review cannot establish that commands work in practice; the benchmark reports positive results but lacks independently reproducible third-party execution evidence, so the static ceiling applies.
The files provide concrete commands, verification steps, public reference entry points, evaluation tasks, and version information, giving the skill some auditability. Deductions apply because there is no verifiable CI test suite or independent multi-source reproduction; the benchmark is author-supplied limited evidence and cannot justify a higher static score.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →