CtrlK
BlogDocsLog inGet started
Tessl Logo

torchscript-exporter

Torchscript Exporter - Auto-activating skill for ML Deployment. Triggers on: torchscript exporter, torchscript exporter Part of the ML Deployment skill category.

35

0.95x
Quality

3%

Does it follow best practices?

Impact

95%

0.95x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./planned-skills/generated/08-ml-deployment/torchscript-exporter/SKILL.md
SKILL.md
Quality
Evals
Security

Evaluation results

100%

3%

Deploy Sentiment Classifier for C++ Inference Service

Production-ready TorchScript export

Criteria
Without context
With context

Export script exists

100%

100%

Uses torch.jit API

100%

100%

Model set to eval mode

100%

100%

Model saved correctly

100%

100%

Correct output path

100%

100%

No gradients during export

100%

100%

Step-by-step documentation

100%

100%

Export log exists

100%

100%

Script is self-contained

100%

100%

Reproducibility consideration

100%

100%

Error handling present

57%

100%

Output directory created

100%

100%

90%

-10%

Validate Exported TorchScript Model for QA Sign-Off

TorchScript output validation

Criteria
Without context
With context

Validate script exists

100%

100%

TorchScript export step

100%

0%

Eval mode set

100%

100%

Multiple test cases

100%

100%

Numerical tolerance check

100%

100%

Max diff computed

100%

100%

Validation report exists

100%

100%

Report has pass/fail field

100%

100%

Report has tolerance field

100%

100%

Report has test count

100%

100%

No grad during inference

100%

100%

Deterministic seed

100%

100%

97%

-3%

Set Up Repeatable Model Export Pipeline for Inference Serving

MLOps pipeline and serving configuration

Criteria
Without context
With context

Pipeline script exists

100%

100%

TorchScript serialization

100%

100%

Eval mode before export

100%

100%

Model artifact saved

100%

100%

Metadata file exists

100%

100%

Metadata includes version

100%

100%

Metadata includes input/output shape

100%

100%

Metadata includes export timestamp

100%

100%

Metadata includes framework version

100%

100%

Pipeline log exists

100%

100%

Log has timestamped stages

100%

100%

No gradient computation in export

100%

62%

Artifacts directory created

100%

100%

Repository
jeremylongshore/claude-code-plugins-plus-skills
Evaluated
Agent
Claude Code
Model
Claude Sonnet 4.6

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.