CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

physical-ai-neural-reconstruction

Router for NVIDIA NuRec/NRE: USDZ rendering, NCore conversion, 3DGS, gRPC sensor sim, PhysicalAI HF datasets. Do NOT use for SimReady or infra setup.

64

Quality

80%

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

Fix and improve this skill with Tessl

tessl review fix ./plugins/nvidia/skills/physical-ai-neural-reconstruction/SKILL.md

The canonical home for this skill is physical-ai-neural-reconstruction in NVIDIA/skills

SKILL.md
Quality
Evals
Security

Quality

Content

86%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 well-engineered router body: highly actionable, clean progressive disclosure into six real reference files, and clear workflow sequencing. The only meaningful deductions are minor redundancy (repeated SimReady exclusion and 150 GB teardown note) and validation checkpoints that live in referenced files rather than inline.

Suggestions

State the SimReady exclusion once (Purpose) and the 150 GB/teardown note once (Cross-skill teardown) instead of repeating each in Limitations/Hard Rules — this would tighten conciseness to a 5.

Add one line of inline validation guidance to the fetch workflow (e.g. what to do when `test -f` fails) so checkpoints don't depend entirely on referenced files.

Trim the 'When to Use' keyword list of near-duplicates (e.g. 'nurec', 'nurec router', 'nurec index') to reduce token cost without losing trigger coverage.

DimensionReasoningScore

Conciseness

The body is dense and disciplined — terse tables, hard rules, and a troubleshooting matrix — and its concept explanations ("What is NuRec?" for NRE/USDZ/NCore V4/3DGUT) are product-specific knowledge Claude does not already have. Not a 5: there is noticeable redundancy — the SimReady exclusion is stated in Purpose, Hard Rules, and Limitations, and the 150 GB teardown note appears in both Limitations and the Cross-skill teardown section.

4 / 5

Actionability

Fully executable guidance throughout: a copy-paste-ready clone/refresh block with env-var fallbacks and a `test -f` verification, executable secret-length check commands, a concrete 4-path local lookup order, a goal→skill picker table, and a troubleshooting table pairing exact error strings with solutions — covering the router's common cases end to end.

5 / 5

Workflow Clarity

Clear sequencing (data → conversion → train → render → cleanup), an ordered local-lookup-then-upstream-clone procedure with an explicit verification step, and error-recovery feedback via the troubleshooting table. Not a 5: per-workflow validation checkpoints are deferred to the upstream sibling skills and references/workflows.md rather than stated inline, so checkpoints are present but not fully explicit.

4 / 5

Progressive Disclosure

The body keeps the core routing content inline (picker tables, fetch recipe) while clearly signaling one-level-deep references — workflows.md, mix-ups.md, upstream-fetch.md, secrets-handling.md, teardown.md, maintenance.md — all of which exist in the bundle and none of which nest further. Navigation is easy and content is appropriately split.

5 / 5

Total

18

/

20

Passed

Description

70%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 terse, domain-specific router description with strong distinctiveness and good keyword coverage. Its main weakness is completeness: it states what the skill routes to and when NOT to use it, but never positively states when to use it.

Suggestions

Add a positive trigger clause, e.g. "Use when the user mentions neural reconstruction, NuRec, NRE, 3DGS, or USDZ rendering" — the current description only says when not to use it, capping completeness at 3.

Spell out one or two natural-language synonyms ("neural reconstruction", "3D Gaussian splatting", "sensor simulation") alongside the abbreviations so users who don't know the acronyms still trigger the skill.

Use verb-led phrasing ("Routes…", "Identifies which sibling skill…") to strengthen specificity over topic-noun listing.

DimensionReasoningScore

Specificity

The description lists several concrete capabilities — "USDZ rendering, NCore conversion, 3DGS, gRPC sensor sim, PhysicalAI HF datasets" — covering the domain with only minor gaps. It falls short of a 5 because the items are topic nouns rather than verb-driven actions (cf. "Extract text and tables from PDF files, fill forms, merge documents"), and above a 3 because coverage spans multiple distinct capabilities rather than just 1-2.

4 / 5

Completeness

The "what" is clear ("Router for NVIDIA NuRec/NRE: USDZ rendering, NCore conversion, 3DGS, gRPC sensor sim, PhysicalAI HF datasets"), but the "when" is only stated negatively ("Do NOT use for SimReady or infra setup") — there is no positive "Use when…" clause, which the guidelines cap at 3. A negative exclusion implies rather than explicitly states when to use the skill.

3 / 5

Trigger Term Quality

"NuRec", "NRE", "USDZ", "NCore", "3DGS", "gRPC sensor sim", and "PhysicalAI HF datasets" are terms a user in this domain would naturally say. Not a 5: spelled-out synonyms such as "neural reconstruction", "3D Gaussian splatting", or "sensor simulation" are missing, though the abbreviations partially compensate.

4 / 5

Distinctiveness Conflict Risk

A clear NVIDIA NuRec/NRE niche with product-specific trigger terms, plus an explicit exclusion clause ("Do NOT use for SimReady or infra setup") that minimizes overlap with adjacent skills — matching the anchor for minimal conflict risk.

5 / 5

Total

16

/

20

Passed

Validation

81%

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

Validation — 13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

13

/

16

Passed

Repository

openai/plugins

Path

plugins/nvidia/skills/physical-ai-neural-reconstruction/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.