The Tessl Registry now has security scores, powered by Snyk

Explore all

Replit launches “Security Agent” to scan and fix vulnerabilities in AI-built apps

Replit's new 'Security Agent' integrates automated vulnerability detection and remediation into its AI coding platform, enhancing security during app development.

Evaluating Kimi 2.5 vs Kimi 2.6: What happens to agent skills when the model gets smarter?

Early signals from benchmarking Kimi K2.5, K2.6, and Sonnet 4.5 on 21 agent skills. Kimi K2.6 is a better model than K2.5, and skills still matter as models improve.

Anthropic, OpenAI, or Cursor model for your agent skills? 7 learnings from running 880 evals (including Opus 4.7)

Explore findings from 880 evaluations comparing Anthropic, OpenAI, and Cursor models, highlighting the impact of agent skills on performance and cost efficiency.

Baptiste Fernandez, Simon Maple

15 min read

21 Apr 2026

Article

Cloudflare introduces “Agent Memory” to help AI agents remember across sessions

Cloudflare's 'Agent Memory' provides AI agents with persistent memory, enhancing their performance by managing context and storing essential information separately.

Google adds subagents to Gemini CLI to handle parallel coding tasks

Google's Gemini CLI now supports subagents, enabling parallel task handling by distributing work across specialized agents, improving efficiency in coding workflows.

Anthropic adds 'routines' to Claude Code for scheduled agent tasks

Anthropic introduces 'routines' to Claude Code, enabling developers to automate and schedule coding tasks, running them without direct interaction or active sessions.

A Proposed Framework For Evaluating Skills [Research Eng Blog]

A new framework evaluates the impact of skills on agent performance, showing a 20% accuracy boost and cost efficiency, while highlighting evaluation challenges.

Vercel open-sources Open Agents to help companies build their own AI coding agents

Vercel has open-sourced Open Agents, a platform for building custom AI coding agents, addressing the limitations of generic tools in large codebases.

The infrastructure gap: what we heard at AI Engineer Europe

AI Engineer Europe highlighted the infrastructure gap in agent deployment, with teams struggling to manage and evaluate skills effectively in production environments.

GitHub brings remote control to Copilot CLI as coding agents move beyond the terminal

GitHub introduces remote control for Copilot CLI, allowing users to manage terminal sessions from web or mobile, reflecting a shift in AI coding agent usage.

GitHub pauses Copilot Pro trials and tightens limits as providers grapple with demand

GitHub pauses new Copilot Pro trials and tightens usage limits due to increased demand, reflecting broader challenges faced by AI tool providers managing system capacity.

I Spent a Week Fixing the Wrong Skill (And Other Lessons from Evaluating an AI PR Reviewer)

The article explores lessons learned from evaluating an AI PR reviewer, highlighting the importance of risk classification and addressing false positives in AI models.