CtrlK
BlogDocsLog inGet started
Tessl Logo

triton-inference-config

Triton Inference Config - Auto-activating skill for ML Deployment. Triggers on: triton inference config, triton inference config Part of the ML Deployment skill category.

41

0.98x
Quality

11%

Does it follow best practices?

Impact

98%

0.98x

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/triton-inference-config/SKILL.md
SKILL.md
Quality
Evals
Security

Evaluation results

97%

-3%

Deploy Sentiment Analysis Model on Triton Inference Server

Production-ready Triton model config

Criteria
Without context
With context

config.pbtxt exists

100%

100%

Correct backend

100%

70%

Model name field

100%

100%

Max batch size set

100%

100%

Input tensor definition

100%

100%

Output tensor definition

100%

100%

Instance group for GPU

100%

100%

Dynamic batching

100%

100%

Deployment guide exists

100%

100%

Step-by-step structure

100%

100%

Validation/verification steps

100%

100%

Production completeness

100%

100%

Without context: $0.2349 · 1m 9s · 13 turns · 14 in / 3,779 out tokens

With context: $0.4217 · 1m 37s · 25 turns · 59 in / 5,482 out tokens

100%

Build a Triton Ensemble Pipeline for Text Classification

Ensemble pipeline config and step-by-step guidance

Criteria
Without context
With context

Ensemble backend field

100%

100%

Ensemble scheduling block

100%

100%

Tensor routing defined

100%

100%

Tokenizer backend

100%

100%

Classifier backend

100%

100%

Correct tensor dtypes

100%

100%

Step-by-step guide

100%

100%

Tensor routing explanation

100%

100%

Validation/testing steps

100%

100%

No placeholders

100%

100%

Ensemble name matches

100%

100%

Without context: $0.4686 · 1m 55s · 23 turns · 24 in / 7,275 out tokens

With context: $0.5793 · 2m 16s · 30 turns · 31 in / 8,462 out tokens

98%

-2%

Optimize Triton Deployment for High-Throughput Production

Production optimization and monitoring config

Criteria
Without context
With context

max_batch_size increased

100%

100%

Dynamic batching added

100%

100%

Multi-GPU instance group

100%

100%

Batching delay or queue settings

100%

100%

Monitoring coverage

100%

100%

Optimization explanations

100%

100%

Validation checklist

100%

100%

Health check or readiness validation

100%

100%

Existing fields preserved

100%

100%

Step-by-step rollout guidance

100%

75%

No placeholders in config

100%

100%

Without context: $0.4257 · 2m 14s · 18 turns · 19 in / 7,237 out tokens

With context: $0.3491 · 1m 47s · 18 turns · 436 in / 5,097 out tokens

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.