Tyler Kendrick
Community@tyler-r-kendrick · Chicago, IL. USA
Former Cloud Solution Architect @microsoft. I empower companies that build software and push the frontier of AI.
Agent Skills by Tyler Kendrick
Showing 76 vetted skills indexed across 2 GitHub repositories.
Tyler-R-Kendrickskill-creator
Create, evaluate, and iteratively improve Agent Skills with benchmark-driven testing loops.
Tyler-R-Kendricklearn
Capture user corrections and encode them into durable repository guidance files.
Tyler-R-Kendrickengineer-skill
Validate and optimize agent skill structure, frontmatter, and SKILL.md body content.
Tyler-R-Kendricktrainer-synthesize
Generate evals.json manifests and APO train/val JSONL datasets from grounded source data.
Tyler-R-Kendricktrainer-train-agent
Orchestrates the trainer optimization loop for agent contract files.
Tyler-R-Kendrickjudge-rubric
Generate formalized judging rubrics with locked dimensions, pass-partial-fail boundaries, and evidence requirements.
Tyler-R-Kendrickcreate-workflow
Create and validate GitHub Agentic Workflows using the gh aw CLI.
Tyler-R-Kendrickresearcher-research
Research public datasets and benchmarks to produce ranked source briefs for eval authoring.
Tyler-R-Kendricktrainer-optimize
Optimizes markdown prompt files using Agent Lightning APO with explicit JSONL train and validation datasets.
Tyler-R-Kendricktrainer-train-skill
Orchestrates the trainer optimization loop for SKILL.md agent skill targets.
Tyler-R-Kendricktrainer-train-code
Orchestrates optimization loops for Python code targets using Microsoft Trace and test-based feedback.
Tyler-R-Kendricktrainer-train
Orchestrates iterative optimization loops for prompt files, skill contracts, and agent contracts.
Tyler-R-Kendricktrainer-train-prompt
Orchestrates the trainer optimization loop for prompt-like markdown and prompty files.
Tyler-R-Kendrickengineer-prompt
Diagnose prompt failures and rewrite markdown prompts using minimal prompt-engineering techniques.
Tyler-R-Kendrickjudge-outcome
Evaluate final outputs and candidate responses against locked rubrics for outcome-based comparison.
Tyler-R-Kendrickengineer-copilot-agent
Validate and optimize GitHub Copilot custom-agent contracts with discovery, routing, and minimization scripts.
Tyler-R-Kendrickengineer-code
Convert Python functions and prompts into trainable Microsoft Trace components with feedback loops.
Tyler-R-Kendrickjudge-trajectory
Evaluate agent trajectories and tool-use traces with process-aware rubrics.
Tyler-R-Kendrickagentv
Create, run, and manage AgentV evaluations for AI agents using EVAL.yaml files.
Tyler-R-Kendricktrainer-election
Elect the strongest prompt or skill candidate from scored evaluation workspace artifacts.
Tyler-R-Kendrickplaywright-cli
Automate browser interactions and web testing workflows via Playwright CLI.
Tyler-R-Kendrickopenui-autoresearch
Orchestrate machine learning research campaigns with hypothesis generation and contract enforcement.
Tyler-R-Kendrickhuggingface-papers
Retrieve structured metadata for AI research papers via the Hugging Face API.
Tyler-R-Kendrickhuggingface-local-models
Search Hugging Face Hub and launch GGUF models via llama.cpp.
Frequently Asked Questions About Tyler Kendrick
FAQPage SchemaWhat tasks can I accomplish with Tyler Kendrick's skills?▼
You can train and fine-tune Hugging Face models, deploy SageMaker endpoints, build Gradio Spaces, estimate VRAM for GGUF and Safetensors models, optimize prompts via Agent Lightning APO, author EVAL.yaml agent evaluations, and audit codebases for minimalist simplification.
Who should use these skills?▼
ML engineers running Hugging Face training jobs, cloud architects deploying SageMaker inference, prompt engineers refining SKILL.md and agent contracts, and researchers orchestrating NeMo-RL or OpenUI symbolic diffusion experiment campaigns benefit most.
What are the installation and runtime requirements?▼
Most trainer and engineer skills require Python 3.11+; trainer-optimize needs agentlightning and openai packages, trainer-train-code needs the trace-opt package, and agentv requires Node.js 18+ with agentv installed globally via npm.
Are these skills open source and what do they cost?▼
The new skills carry MIT licenses at no cost. Usage expenses come only from underlying infrastructure you choose, such as Hugging Face Jobs, ZeroGPU, or AWS SageMaker compute, not from the skills themselves.
What prerequisites or repository structure do the trainer skills expect?▼
Trainer loops expect repositories keeping artifacts in a .trainer-workspace/ directory next to the target file, a validation command, and stage capabilities. Agent targets use .agent.md files; skill targets follow the agentskills.io specification.