CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

nemoclaw-user-get-started

Installs NemoClaw, launches a sandbox, and runs the first agent prompt. Use when onboarding, installing, or launching a NemoClaw sandbox for the first time. Trigger keywords - nemoclaw quickstart, install nemoclaw openclaw sandbox, nemohermes quickstart, hermes agent nemoclaw, run hermes openshell sandbox, nemoclaw prerequisites, nemoclaw supported platforms, nemoclaw hardware software, nemoclaw windows wsl2 setup, nemoclaw install windows docker desktop.

66

Quality

83%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Medium

Suggest reviewing before use

SKILL.md
Quality
Evals
Security

Quality

Content

67%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is a well-structured, highly actionable quickstart with clear sequencing, error-recovery guidance, and properly signaled one-level-deep references that all exist. Its main weaknesses are redundancy (repeated install command and dashboard-URL explanations) and inlined platform edge-case detail that belongs in the existing references.

Suggestions

Remove the duplicated `curl -fsSL https://www.nemoclaw.sh | bash` block (lines 43-46 keep only the `newgrp docker` recovery step) and consolidate the three explanations of the authenticated dashboard URL into one passage.

Move platform-specific express-install detail (DGX Spark/DGX Station/WSL behavior, `NEMOCLAW_NO_EXPRESS`, `NEMOCLAW_POLICY_TIER` defaults) into references/windows-preparation.md or a platform reference, keeping the main flow to the common Linux/macOS path.

Unify the placeholder sandbox name (`my-assistant` at line 178 vs `my-gpt-claw` elsewhere) so the copy-paste examples are internally consistent.

DimensionReasoningScore

Conciseness

Most content is product-specific detail Claude cannot know, but there is real redundancy: the `curl -fsSL https://www.nemoclaw.sh | bash` command appears twice (lines 24 and 45), and the authenticated `dashboard-url --quiet` workflow is explained three times in the browser section (lines 162-171). This fits 'mostly efficient but could be tightened' rather than the minor-trimming anchor at 4.

3 / 5

Actionability

Commands are concrete and copy-paste ready (`curl | bash`, the headless `NEMOCLAW_NON_INTERACTIVE=1` variant, `newgrp docker`, `nemoclaw my-gpt-claw connect`, `openclaw tui`). Minor gaps: the terminal example switches placeholder sandbox names (`my-assistant` vs the consistent `my-gpt-claw`), and the non-express provider selection path is only named, not shown.

4 / 5

Workflow Clarity

The sequence (prerequisites, install, review summary, web search/messaging, network policy presets, verification, first prompt) is clearly ordered with recovery paths (answer `n` and re-run `nemoclaw onboard`, PATH re-source, Docker socket recovery, port-collision rollback) and installer-side gateway/dashboard verification. It sits below 5 because mid-wizard checkpoints are deferred to the reference and validation is described as installer behavior rather than explicit user validation steps.

4 / 5

Progressive Disclosure

All four referenced files exist (prerequisites.md, quickstart-details.md, quickstart-hermes.md, windows-preparation.md) and are one level deep, each clearly signaled with a 'Load [references/X] when...' condition. Not 5: the inline pointer at line 150 duplicates the bottom References list, and platform-specific install detail (DGX/WSL express install) is inlined rather than split into the existing platform references.

4 / 5

Total

15

/

20

Passed

Description

91%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is strong: it states concrete actions in third person, gives an explicit 'Use when' clause, and carries a comprehensive, natural-sounding trigger keyword list with product synonyms. The only weakness is minor overlap between some trigger keywords and sibling NemoClaw skills, and a 'what' slightly narrower than the skill's full scope.

DimensionReasoningScore

Specificity

"Installs NemoClaw, launches a sandbox, and runs the first agent prompt" lists three concrete third-person actions, matching the 'several specific actions with minor gaps' anchor; the what-coverage is thinner than the skill's full scope (onboard wizard, inference, security policies), keeping it below 5.

4 / 5

Completeness

It explicitly answers what ("Installs NemoClaw, launches a sandbox, and runs the first agent prompt") and when ("Use when onboarding, installing, or launching a NemoClaw sandbox for the first time") with concrete trigger phrases, matching the top anchor exactly.

5 / 5

Trigger Term Quality

"nemoclaw quickstart", "nemohermes quickstart", "hermes agent nemoclaw", "nemoclaw windows wsl2 setup", "nemoclaw install windows docker desktop" give comprehensive natural-phrase coverage with product synonyms (nemoclaw/nemohermes/hermes/openclaw/openshell) and little a user would naturally say that is missing.

5 / 5

Distinctiveness Conflict Risk

NemoClaw onboarding is a clear niche with product-specific triggers, but keywords like "nemoclaw prerequisites" and "nemohermes quickstart" overlap with the related nemoclaw-user skills referenced in the body, so minor overlap risk remains.

4 / 5

Total

18

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

openai/plugins

Path

plugins/nvidia/skills/nemoclaw-user-get-started/SKILL.md

Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.