Agent Skills
PulseAugur coverage of Agent Skills — every cluster mentioning Agent Skills across labs, papers, and developer communities, ranked by signal.
- 2026-08-09 product_launch Agent Skills has been launched, providing production-grade engineering capabilities for AI coding agents. source
13 day(s) with sentiment data
-
Microsoft Power BI adds Agent Skills for natural language data interaction
Microsoft has introduced Agent Skills for Power BI, a new feature announced at Build 2026. This capability allows users to describe reports in natural language or provide screenshots, enabling the agent to understand an…
-
Anthropic's Claude Code requires minimal changes to support Agent Plugins 1.0
Anthropic's absence from the Agent Plugins 1.0.0 standard, released by AWS, Microsoft, OpenAI, Vercel, and Cursor, is notable given the spec's reliance on Anthropic's own Agent Skills format. A technical analysis reveal…
-
Agent Skills launches production-grade engineering for AI coding agents
A new product called Agent Skills has been launched, offering production-grade engineering capabilities for AI coding agents. The tool aims to enhance the development process for AI-driven coding assistants.
-
Vercel launches Agent Plugins standard for AI agent portability · 2 sources tracked
Vercel has introduced Agent Plugins, a new open and vendor-neutral standard for packaging AI agent skills and MCP servers into portable plugins. This specification aims to create a unified format for agent capabilities,…
-
OpenAI launches Agent Plugins standard for cross-client compatibility
OpenAI has introduced Agent Plugins, an open standard designed to allow developers to build a plugin once and deploy it across various compatible agent clients. This new standard, developed in collaboration with AWS Dev…
-
AI developer streamlines agent rules with reusable "Agent Skills"
An AI developer is exploring ways to streamline project rules for AI agents by using "Agent Skills." This approach converts repetitive prompts into shareable, reusable instructions that can be applied across different t…
-
Genkit Go uses Agent Skills to reduce token costs and context bloat
Genkit Go has introduced Agent Skills, a modular approach designed to reduce context window bloat and token costs. This feature utilizes SKILL.md bundles to separate specialized functionalities from the main context, al…
-
Open-source project turns novels into playable games using AI · 4 sources tracked
A new open-source project called "novel-to-game" is gaining traction, utilizing Agent Skills to transform novels into playable games. This project leverages AI models like Claude Code, Codex, and Kimi Code to ground the…
-
MCP and Agent Skills: Complementary AI Technologies, Not Competitors
The Model Context Protocol (MCP) and Agent Skills are distinct but complementary technologies, not competing ones. MCP, a connection protocol for agent interactions with external tools, recently finalized its 2026-07-28…
-
Developers explore Claude Code sub-agent strategies and Antigravity CLI capabilities
A developer found that cataloging failures was more effective than providing clever instructions when using Claude Code for sub-agent operations. Another developer explored the capabilities of Antigravity CLI, which fea…
-
New tool enables coding agents to generate consistent diagrams locally
A developer has created a local tool called tldraw-canvas-kit, which consists of 19 Agent Skills designed to enable coding agents like Claude Code and Cursor to interact with the offline Tldraw desktop application. This…
-
New Skillware ontology aims to standardize AI agent behavior artifacts
This paper introduces Skillware, a new software ontology and engineering lifecycle designed for persistent behavioral artifacts in AI agent systems. Skillware defines these artifacts as independent software objects, ena…
-
AI integration explored across development tools and review platforms · 3 sources tracked
Three separate articles discuss the integration and application of AI in various development and operational contexts. One piece details updates to New Relic in June 2026, highlighting its use with AWS, AI, and Kubernet…
-
Progressive disclosure aids long-context AI agents in multi-document tasks
A new study investigates the effectiveness of progressive disclosure for long-context AI agents, particularly in tasks involving multiple documents. The research found that while progressive disclosure can improve conte…
-
Anthropic launches free Claude API training academy with certificates
Anthropic has launched Anthropic Academy, a free online learning platform offering over 18 self-paced courses on its Claude API, Model Context Protocol (MCP), Claude Code, and Agent Skills framework. The courses, hosted…
-
New framework detects misalignment in LLM agent skills
Researchers have developed a new framework called Progressive Loading-Aware Hierarchical Contrastive Learning (PL-HCL) to detect inconsistencies between the descriptions and actual behavior of Large Language Model (LLM)…
-
New framework detects inconsistencies in LLM Agent Skills
A new framework called Progressive Loading-Aware Hierarchical Contrastive Learning (PL-HCL) has been developed to detect inconsistencies between the descriptions and actual behaviors of LLM Agent Skills. This method mod…
-
Cursor AI uses Markdown files to teach ML conventions
Cursor has developed a new feature called Agent Skills, which allows teams to encode their machine learning conventions. This enables the AI to understand and adhere to these conventions consistently, improving the accu…
-
Evaluating AI Agent Skills: A Comparison of Three Tools
This article compares three tools designed to evaluate the effectiveness of "Agent Skills," a concept related to AI agent capabilities. The author explores how these tools assess whether an agent's skills are genuinely …
-
COMFYCLAW framework enhances agent control for image generation workflows
Researchers have introduced COMFYCLAW, a new framework designed to enhance agent control over ComfyUI image generation workflows. This system formulates workflow construction as typed graph editing and uses a vision-lan…