Claude agent with tools and evals
Quest Build a Claude Agent SDK app with its own MCP server, then trace and evaluate it
- Hero
- Claude Code
- Weapon
Claude Opus 5.5- Charge
Anthropic Agent SDK Dev
Anthropic MCP Builder
MLflow Skills
Summoner skill tree
Every spell in this class, in the row where it usually goes. Pick a few from each row for a loadout.
ScoutGather context
CraftMake the thing
TestCheck the work
ShipDeliver it
Nothing equipped here
Loadouts
Combos
Build a Claude agent, then trace and grade it
Summoner · AI and agent engineering- Anthropic Agent SDK Dev
- MLflow Skills
Spells
9 spells
Anthropic skill for building MCP servers in Python (FastMCP) or TypeScript, from API research to tests and an eval set.
182,779 GitHub stars on anthropics/skills
Scaffolds a Claude Agent SDK app in Python or TypeScript with /new-sdk-app, then checks it with verifier agents.
43,650 GitHub stars on anthropics/claude-plugins-official
Hugging Face skills for the hf CLI and Hub, plus on-demand skills for datasets, training, evals, Gradio and local models.
11,195 GitHub stars on huggingface/skills
Codex plugin for building on the OpenAI API: Agents SDK apps, ChatGPT Apps, API keys and API error triage.
7,528 GitHub stars on openai/plugins
A Langfuse skill for tracing setup, prompt management, datasets, evaluations and API queries.
312 GitHub stars on langfuse/skills
Official Hub server: search models, datasets, Spaces and papers, and call Gradio apps on Spaces as tools.
308 GitHub stars on huggingface/hf-mcp-server
MLflow skills to add tracing to an AI app, debug failing traces, build scorers and run agent evaluations.
82 GitHub stars on mlflow/skills
Queries W&B runs and Weave traces, compares and diagnoses runs, summarizes evals and writes W&B reports.
72 GitHub stars on wandb/wandb-mcp-server
One Together AI skill with guides and scripts for inference, fine-tuning, evaluations, batch jobs, media generation and GPUs.
37 GitHub stars on togethercomputer/skills