add model builder skill#1745
Conversation
|
@mnabian just curious: is it a 1-to-1 conversion of the older cursor rules into skills? |
melo-gonzo
left a comment
melo-gonzo
left a comment
There was a problem hiding this comment.
This looks good overall. You will not need to include skill-card.md or BENCHMARK.md - these will be generated by the nvskills-ci pipeline. Once you sync with the main branch we can invoke the CI pipeline and have a look at the reports generated and make any necessary changes.
To @CharlelieLrt's point - i think its ok to leave the rules as is, and simply point to them in the skills like you do here. For other skills, i've takne the approach of allowing the agent to self discover the rules by only pointing to the main folder/files. Adding specific references as you do here (e.g, pointing to MOD-00b may make the skill more token efficient, but also if we do end up changing the location of these rules in the future we'll have to go back an update all of the skills.
|
/nvskills-ci |
|
/nvskills-ci |
I agree with Carmelo. I think the rules can stay there for now and point the skills to them (or have them discover the rules). |
…icsnemo into model-builder-skill
| patterns in `references/reuse_map.md`; verify each path before citing it. Say | ||
| explicitly "import `X` from `physicsnemo.nn` instead of writing your own." Keep | ||
| genuinely novel, model-specific pieces local to the model. | ||
|
|
There was a problem hiding this comment.
@mnabian Something that can come up at this step is "layer X from phyiscsnemo.nn can help me but it needs a minor change/extension Y -- do I make that change to X or should I make a new module to support my needs?". For example this is an important design decision point in nearly every Earth-2 NVR model integration I've done, and unfortunately it can end up quite subjective on what the best path is.
Should we add some language here guiding that decision making (or redirect decision to user, possibly with some agent-generated proposals), or is this skill intended to be used strictly without making changes to the core library other than adding the new model?
4 participants
PhysicsNeMo Pull Request
Description
Checklist
Dependencies
Review Process
All PRs are reviewed by the PhysicsNeMo team before merging.
Depending on which files are changed, GitHub may automatically assign a maintainer for review.
We are also testing AI-based code review tools (e.g., Greptile), which may add automated comments with a confidence score.
This score reflects the AI’s assessment of merge readiness and is not a qualitative judgment of your work, nor is
it an indication that the PR will be accepted / rejected.
AI-generated feedback should be reviewed critically for usefulness.
You are not required to respond to every AI comment, but they are intended to help both authors and reviewers.
Please react to Greptile comments with 👍 or 👎 to provide feedback on their accuracy.