AI Coding Live News — Real-Time 𝕏/Twitter Updates & Breakthroughs

👀51

🔥91

AgentSkillsHub open-sources a scored directory of Claude Skills, MCP servers, and Codex skills

@DanKornas highlighted AgentSkillsHub (github.com/zhuyansen/agent-skills-hub), an open-source directory that classifies MCP servers, Claude Skills, Codex Skills and more into seven categories, scores quality across six dimensions, and offers /compare/ plus /best/{scenario}/ pages, with GitHub Actions syncing every eight hours.

Key Takeaways

Seven-category catalog for MCP, Claude/Codex Skills, coding assistants, and more.
Six-dimension quality scoring plus compare and scenario pages.
GitHub Actions refreshes the directory every eight hours.

Read full story

Linked Models & Agents:

🤖 Claude Code (Anthropic)🤖 codex 🌐 Anthropic 🌐 OpenAI

#AgentSkillsHub#MCP#Claude Skills#Codex#Directory#open source

Langfuse@langfuse·Collected 67d ago

👀31

🔥86

View on 𝕏 ↗

Langfuse Ships CLI v1 for Terminal-Native Agent Evaluation and CI/CD Quality Gates

Langfuse released CLI v1, enabling developers and coding agents to run evaluation suites directly from the terminal, compare prompt diffs, and enforce automated quality gates in CI/CD pipelines.

Key Takeaways

Terminal-first CLI execution allows instant eval suite runs directly within local developer shells and CI jobs;
Integrates with autonomous coding agents to allow self-verification before automated commit creation;
Enforces strict PR merge gates based on quantitative accuracy, latency, and cost regression thresholds.

Read full story

Linked Models & Agents: