CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

nemoclaw-setup

Install and configure NVIDIA NemoClaw (sandboxed OpenClaw agent platform) on Linux. Handles cloudflared tunnels, Docker cgroup fixes, OpenShell, sandbox creation, remote access via Cloudflare Tunnel, and known bug workarounds. Use whenever the user mentions installing NemoClaw, setting up OpenClaw, configuring an NVIDIA Spark or DGX for sandboxed agents, or troubleshooting NemoClaw deployment.

Quality

92%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

81%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.

A strong, highly executable setup guide: every step has concrete commands, validation checkpoints and error-recovery paths are thorough, and the structure is easy to navigate. The main gaps are the non-expanding $USER placeholders in the systemd unit, some padding/time-sensitive version pins, and a large single-file body that could be slimmed with a references bundle.

Suggestions

Fix the systemd unit heredoc so it produces an executable service: substitute the real username/home path at write time (e.g., printf with $USER/$HOME, or explicitly instruct the user to replace $USER), since systemd does not expand shell variables in unit files.

Trim token overhead: drop or compress the 'What You Get' section and move pinned version specifics (OpenClaw ≤ v2026.3.11 bug, Nemotron 3 Super 120B) into a clearly labeled 'Known issues / version notes' subsection so dated content is isolated and easy to update.

Offload self-contained sub-procedures — the keepalive script + systemd service, Cloudflare remote access, and custom-skill authoring — into references/ files (e.g., references/remote-access.md, references/persistent-forward.md) and keep SKILL.md as a lean overview pointing one level deep to them.

DimensionReasoningScore

Conciseness

The body is command-dense and assumes Claude's competence (no basic concept explanations), but the 'What You Get' section is promotional padding and pinned versions ('OpenClaw ≤ v2026.3.11', 'Nemotron 3 Super 120B') are time-sensitive content outside a dedicated deprecated/old-patterns section, matching 'efficient; minor instances of over-explanation that could be trimmed' rather than the every-token-earns-its-place level 5.

4 / 5

Actionability

Nearly everything is copy-paste executable (curl installers, onboarding, keepalive script, full troubleshooting fixes), but the systemd unit heredoc is quoted ('SERVICE') so it writes literal 'User=$USER' and '/home/$USER/...' that systemd cannot expand — a minor but real executability gap that keeps it at 'mostly executable guidance with minor gaps' rather than fully copy-paste ready.

4 / 5

Workflow Clarity

Eleven clearly sequenced steps with a pre-flight check step, conditional skip ('If nemoclaw is already installed, skip to Step 4'), a dedicated Verify step, a health-check feedback loop that auto-restarts the tunnel, and a 13-row troubleshooting table for error recovery — matching the level-5 anchor of explicit validation with feedback loops; no destructive step lacks warnings (the update flow flags sandbox destroy/recreate and required reconfiguration).

5 / 5

Progressive Disclosure

Good structure: well-labeled sections, prerequisites table, architecture diagram, and a clearly signaled References list, but this ~320-line single file inlines substantial self-contained material (keepalive script + systemd unit, Cloudflare tunnel setup, custom-skills authoring) that a references/ bundle could offload — 'most content is appropriately placed; minor organization gaps' rather than the well-split level 5.

4 / 5

Total

17

/

20

Passed

Description

100%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.

A model description: concrete action list, explicit third-person 'Use whenever...' trigger clause with natural synonyms, and a well-defined niche of named products. Both 'what' and 'when' are answered clearly and concisely with no padding.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — 'Install and configure', 'Handles cloudflared tunnels, Docker cgroup fixes, OpenShell, sandbox creation, remote access via Cloudflare Tunnel, and known bug workarounds' — giving comprehensive coverage of the skill's capabilities, matching the level-5 anchor rather than the 'several specific actions; minor gaps' level-4 anchor.

5 / 5

Completeness

Explicitly answers both questions: the 'what' ('Install and configure NVIDIA NemoClaw... Handles cloudflared tunnels...') and the 'when' ('Use whenever the user mentions installing NemoClaw, setting up OpenClaw, configuring an NVIDIA Spark or DGX...') with concrete trigger phrases, exactly matching the level-5 anchor; the missing-trigger cap of 3 does not apply.

5 / 5

Trigger Term Quality

The 'Use whenever the user mentions' clause covers the natural phrasings users would actually say — 'installing NemoClaw', 'setting up OpenClaw', 'configuring an NVIDIA Spark or DGX', 'troubleshooting NemoClaw deployment' — including the install/setup/configure/troubleshoot synonym set, which is comprehensive for this domain; no natural term is conspicuously missing.

5 / 5

Distinctiveness Conflict Risk

Names a clear niche via specific product names (NemoClaw, OpenClaw, NVIDIA Spark, DGX) with distinct triggers, so it is unlikely to fire for unrelated skills — matching the 'clear niche with distinct triggers; minimal conflict risk' anchor.

5 / 5

Total

20

/

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

jezweb/claude-skills

Path

plugins/integrations/skills/nemoclaw-setup/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.