From 68% to 100%: Optimizing Skills With a Single Command

Explore all

GitHub brings Claude and Codex agents directly into Copilot

GitHub integrates Anthropic's Claude and OpenAI's Codex into Copilot, offering Business and Pro users a choice of coding agents for tasks like code drafting and review.

Ollama and OpenClaw make natural bedfellows for the local AI stack

Ollama and OpenClaw are good examples of a local-first AI stack, reducing cloud dependency by running models and agent frameworks on consumer hardware for enhanced control and efficiency.

The Context Flywheel: Why the Best AI Coding Teams Will Win on Context

The Context Flywheel explores how iterative context development in AI coding teams enhances agent output, expertise, team learning, and organizational alignment.

OpenAI moves beyond SWE-bench Verified as coding benchmarks saturate

OpenAI shifts focus from SWE-bench Verified as coding benchmarks reach saturation, exploring more complex evaluations for AI models in real-world scenarios.

Your AGENTS.md file isn't the problem. Your lack of AI Agent Evaluations is.

The issue with context files like AGENTS.md is not their existence but the lack of evaluation to validate their effectiveness in guiding AI agents.

Anthropic brings ‘frontier cybersecurity’ to Claude Code as cyber stocks slide

Anthropic introduces AI-driven Claude Code Security to help developers identify and fix software vulnerabilities, integrating into existing workflows.

My Coding Agent Needed a Package Manager for Its Own Brain (And I Gave It One Using a Skills Registry)

Explore how a skills registry transformed a coding agent's decision-making, enabling it to read library documentation and make informed stack decisions.

The Context Development Lifecycle: Optimizing Context for AI Coding Agents

The Context Development Lifecycle focuses on optimizing context for AI coding agents, highlighting the shift from code writing to context understanding in software development.

Do Agent Skills Actually Help? A Controlled Experiment

A controlled experiment using Harbor evaluates if Claude Code's Agent Skills improve performance on domain-specific tasks, focusing on database migration scenarios.

Claude Code hid file-access data — a new open-source observability tool emerged

A developer built an observability layer for Claude Code after Anthropic's update hid file-level activity, impacting visibility into agentic coding tool operations.

Introducing Task Evals: Measure Whether Your Skills Actually Work

Today we’re launching Task Evals: a built-in way to measure whether a skill is actually steering agent behaviour. Skills can be well written and still drift as models and surrounding context change. Task evals compare outcomes with and without a skill, making it clear when it’s helping, doing nothing, or working against you.

Claude Can't Write PubNub Functions. We Fixed That.

The article discusses how a Tessl skill was developed to enable Claude Code to write PubNub Functions effectively, improving deployment success from 60% to 100%.