Firehose for 24 April 2026

Four repositories cover AI evaluation and agent tooling: skills for testing models against real workloads, a collection of evaluation frameworks, a plugin for building curated resource lists, and a snapshot of DeepSeek V4 pricing and availability.

GitHub Repositories

4 repos created

AI Eval GitHub · Agent plugin Agent skills for evaluating AI models for a real job: define the workload, draw candidates from OpenRouter, run wide and shallow, publish findings.
Awesome AI Evaluations Tools GitHub · Evaluation Collection of frameworks and tools for AI evalations, including tool-use, agentic AI, MCP, and multimodal
Claude Resource List Builder Plugin GitHub · Agent plugin Claude Code plugin: build, maintain, and audit curated GitHub resource lists with AI-driven categorisation.
Deepseek 4 Pricing 240426 GitHub · Documentation A point-in-time, incomplete synopsis of DeepSeek V4 model availability and pricing as observed on 24 April 2026.