SKILLEMALL.ai
GitHub ↗ RU Sign in

CC nemotron-nano3

Reference desk for Nemotron 3 Nano / Llama-Nemotron Nano 3 — architecture, training data, recipes, evaluation, quantization, deployment. Use when the user asks facts about the model rather than building a pipeline.

The skillemall take

Reference guide for NVIDIA's Nemotron 3 Nano: architecture, training data, quantization recipes, and deployment. Promises factual information about the model. Contains 18 files, no scripts—documentation only. Checks found 17 medium and low severity issues, no critical ones. Safety score 83, quality 84, but process score dropped to 64. No model runs or sandbox testing performed.

Use it for: architecture details, training parameters, quantization options. Won't help with: automating deployment or building pipelines—no such functions included. Install if you need a reference on Nemotron specs, not a tool to operate it.

NVIDIA-NeMo/Nemotron Agent Skills author: NVIDIA-NeMo 18 files body ≈ 1 819 tokens Open the sourcegithub.com↗ analyzed 5 d ago

Reference desk for Nemotron 3 Nano / Llama-Nemotron Nano 3 — architecture, training data, recipes, evaluation, quantization, deployment.

As a process C 64/100 · Has gaps — weak spots: result and completion, inputs and preconditions, running it twice

ReferenceInfrastructuretype and topics are labelled automatically from the skill text
Runs in: Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) OpenClaw Goose OpenCode
JSON
Technical rating
C
83/100
safety, quality, tests
Safety 60%
83
Quality 40%
84
Run on models
none yet
Process rating
C
64/100
Has gaps
Result and completion w 14
0
Inputs and preconditions w 11
0
Progress reporting w 2
0
the three weakest of ten parameters · all ten

How to improve

    For the model run — optional
    • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
    • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

    Guard findings · 17

    ✓ No critical or high findings

    Medium and low: 17
    • low Secrets in code secret-high-entropy-token context/quick-reference.md:24
      High-entropy token-like string (may be an id, hash or a credential) (documentation table row)
      | `NVID…F16` | base pretrained model |
      table
    • low Secrets in code secret-high-entropy-token context/quick-reference.md:25
      High-entropy token-like string (may be an id, hash or a credential) (documentation table row)
      | `NVID…F16` | post-trained model |
      table
    • low Secrets in code secret-high-entropy-token context/quick-reference.md:26
      High-entropy token-like string (may be an id, hash or a credential) (documentation table row)
      | `NVID…FP8` | quantized post-trained deployment model |
      table
    • low Secrets in code secret-high-entropy-token context/quick-reference.md:66
      High-entropy token-like string (may be an id, hash or a credential)
      | Nemo…-v2 | refreshed GitHub/code corpus plus synthetic code |
    • low Secrets in code secret-high-entropy-token context/quick-reference.md:68
      High-entropy token-like string (may be an id, hash or a credential)
      | Nemo…-v1 | synthetic STEM reasoning, textbook, scientific coding, InfiniByte cross-domain code |
    • low Secrets in code secret-high-entropy-token model-card.md:23
      High-entropy token-like string (may be an id, hash or a credential) (documentation table row)
      | Base checkpoint | `NVID…F16` |
      table
    • low Secrets in code secret-high-entropy-token model-card.md:24
      High-entropy token-like string (may be an id, hash or a credential) (documentation table row)
      | Post-trained checkpoint | `NVID…F16` |
      table
    • low Secrets in code secret-high-entropy-token model-card.md:25
      High-entropy token-like string (may be an id, hash or a credential) (documentation table row)
      | Quantized checkpoint | `NVID…FP8` |
      table
    • low Secrets in code secret-high-entropy-token paper/_overview.md:90
      High-entropy token-like string (may be an id, hash or a credential)
      | Qwen…nRM | reward model used for RLHF |
    • low Secrets in code secret-high-entropy-token paper/_overview.md:98
      High-entropy token-like string (may be an id, hash or a credential)
      | Nemo…-v2 | refreshed GitHub/synthetic code corpus |
    • low Secrets in code secret-high-entropy-token paper/_overview.md:99
      High-entropy token-like string (may be an id, hash or a credential)
      | Nemo…-v1 | synthetic specialized pretraining data |
    • low Secrets in code secret-high-entropy-token paper/data.md:50
      High-entropy token-like string (may be an id, hash or a credential)
      | Nemo…-v2 | refreshed GitHub/code plus synthetic augmentation |
    • low Secrets in code secret-high-entropy-token paper/data.md:52
      High-entropy token-like string (may be an id, hash or a credential)
      | Nemo…-v1 | specialized synthetic content for STEM, textbooks, scientific coding, and reasoning |
    • low Secrets in code secret-high-entropy-token paper/data.md:82
      High-entropy token-like string (may be an id, hash or a credential)
      ## 2.2.2 Nemo…-v2
    • low Secrets in code secret-high-entropy-token paper/data.md:168
      High-entropy token-like string (may be an id, hash or a credential)
      ## 2.2.4 Nemo…-v1
    • low Secrets in code secret-high-entropy-token paper/data.md:247
      High-entropy token-like string (may be an id, hash or a credential)
      | teacher model | Qwen…507 |
    • low Secrets in code secret-high-entropy-token paper/rl.md:109
      High-entropy token-like string (may be an id, hash or a credential)
      - Qwen…507 is used as judge

    Files scanned: 18. Evidence is masked. Grey chips explain why severity was lowered.

    Against the Agent Skills spec

    ✓ No remarks against the Agent Skills spec

    Process rating: all ten parameters 64/100

    • 0Result and completion. Does not say what the result is
    • 0Inputs and preconditions. Does not say what the process needs to start
    • 0Progress reporting. Says nothing while it works
    • 30Running it twice. 4 mutating operations with no state check
    • 70When it triggers. States when to use, but not when not to
    • 85Steps. 65 steps, 1 vague phrases
    • 100Tools and files. No external tools needed
    • 100Failures and branches. 2 branches, has a failure section
    • 100Consistency. Name and required fields are in place
    • 100Execution cost. Instruction body is 1819 tokens

    Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

    Quality signals

    • +5Description has no quoted example phrases that should trigger the skill
    • +4Description does not say when NOT to use the skill (false activations)
    • +3Output format is not stated: the model decides each time
    • +1No license
    • +2Single-language instructions
    • +3Description length 214: enough signal without eating the budget
    • +4Structure: 18 headings
    • +3Step-by-step instructions: 65 items
    • +4Has examples (1 code blocks)

    Quality base 70; lint remarks subtract, signals add up to 100. Result: 84.