BC nemotron-3-5-lightning-text2sql-lora
Run the Nemotron-3.5 Lightning Text2SQL LoRA fine-tuning tutorial (NeMo Megatron-Bridge) end-to-end for the user on a single node: data prep, checkpoint conversion, LoRA fine-tuning of the 30B-A3B hybrid Mamba-Transformer MoE, and merging the adapter back to a Hugging Face checkpoint. Use when the user wants to run this cookbook, fine-tune Nemotron-3.5 Lightning with LoRA, or adapt the notebook to their own machine.
The skill promises to run a complete fine-tuning cycle for Nemotron-3.5 Lightning with LoRA: data prep, checkpoint conversion, training the 30B-A3B hybrid Mamba-Transformer MoE, and merging the adapter back to Hugging Face format. Eight modules and six scripts in the files, 1761 tokens of description.
Tests gave it a B grade: quality 80, process 63. No critical errors, linter passed, references intact. Sandbox didn't run. Works across all major platforms—Claude, Cursor, Gemini, Deepseek, and others. This is a workflow type, so a complete recipe rather than a single function.
Install it if you're ready for a long process on a single machine and don't expect flawless execution (the 63 process score hints at organizational rough edges). Good for LoRA experiments on Nemotron.
Run the Nemotron-3.5 Lightning Text2SQL LoRA fine-tuning tutorial (NeMo Megatron-Bridge) end-to-end for the user on a single node: data prep, checkpoint…
As a process C 63/100 · Has gaps — weak spots: result and completion, inputs and preconditions, consistency
How to improve
- 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 · 0
✓ No critical or high findings
Files scanned: 8. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- note
edit-residuethe text marks something as outdated (lines 100): check that old rules are not kept next to new ones — the full check reads the text for contradictions
Process rating: all ten parameters 63/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 40Consistency. Frontmatter name (nemotron-3-5-lightning-text2sql-lora) differs from the folder (nemo-megatron-bridge)
- 65Failures and branches. 3 branches
- 70When it triggers. States when to use, but not when not to
- 100Tools and files. No external tools needed
- 100Steps. 28 steps
- 100Execution cost. Instruction body is 1761 tokens
- 100Running it twice. No mutating operations
- 100Progress reporting. Reports progress
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
- +4No input/output examples
- +1No license
- +2Single-language instructions
- +3Description length 419: enough signal without eating the budget
- +4Structure: 9 headings
- +3Step-by-step instructions: 28 items
Quality base 70; lint remarks subtract, signals add up to 100. Result: 80.