LangChain Cloud for traces, evals, and deploys. Plus is $39/seat. Startup is $1,400/month. LangSmith Engine is $2,000/month.

Open-source LLM tracing and evals. Cloud Hobby is $0. Core is $29/seat. Pro is $199/seat. Plus usage after included units.

Superpowers' code review skill has coding agents dispatch a fresh reviewer subagent with a git range and requirements, then fix issues by severity.

Confidentiality-first peer review skill: authorization gate, claim-evidence and statistics checks, and local CLIs that draft and lint a review.

Apply ScholarEval framework to systematically evaluate scholarly work - comprehensive analysis across research quality dimensions with structured scoring and feedback.

Evaluate research rigor systematically - assess methodology, experimental design, statistical validity, biases, confounding, and evidence quality using GRADE.

AWS's AI-first observability experience: one workspace to trace, evaluate, and troubleshoot applications and agents across accounts, regions, and clouds.

Jeff is an open-source family of 0.8B and 2B decision models that return a calibrated probability per option in about 22 ms on a single GPU.

Open-source Microsoft sandbox and benchmark that verifies AI agents reliably complete stateful business tasks, exposing the discovery-reliability gap.

NVIDIA's open-source LLM vulnerability scanner with dozens of plugins and thousands of prompts for probing jailbreaks, prompt injection, and data leakage.

NVIDIA's open-source toolkit for adding programmable guardrails to LLM conversational applications, with Colang flows for input, dialog, retrieval, and output rails.

Open-source terminal coding agent that makes agents justify 'done': a council of three auditors votes on whether the evidence backs the finish.

Bespoke Labs' open-weight 9B model and training recipe that turn Jev-style decisions into single-token choices with probabilities, runnable locally.

Jev is TypeSafe's System One model for typed software decisions with calibrated probabilities, priced at $0.042 per million input tokens.

3Dviz Pro Max is an MIT Claude Code and Codex skill that turns one sentence into an inspectable Three.js or Blender scene.

Hugging Face's community-evals skill runs benchmark evaluations of Hub models on local hardware with inspect-ai or lighteval, picking vLLM or Transformers.

Linus Torvalds Skill is a CC0 code-review pack distilled from 31,397 LKML emails into four model-specific SKILL.md files.