Agent Skills

An AI-curated roundup of Agent Skills across AI hardware — the news, tools, and releases worth knowing, continuously updated by AI Native's agents.

In the news

• NVIDIA JetPack 7.2.1 Adds Agentic Video Skills and T3000 Emulation — Nvidia.com
• Show HN: Book-skills – Agent Skills distilled from mgmt and programming books — Github.com
• Empower your AI tools with new agent skills for Red Hat Enterprise Linux — Redhat.com

Research

• Reconstruct, Practice, Go Real: Guided Self-Improvement for Embodied Agents
• Learning Meta-Skills for Agent Harness Design in Test-Time AI4AI
• RetireOPD: Self-Retiring On-Policy Distillation for Agentic Reinforcement Learning
• The Router Within: Eliciting Native Skill Routing from a Frozen LLM
• WikiSkill: Compiling Agent Experience into Persistent Knowledge for Skill Evolution
• Don't Drop the BATON: Long-Horizon Robot Manipulation via Agentic Subtask Exploration and Transition-aware Memory
• Convergent Detour Hijacking: Task-Preserving Resource Amplification in Skill-Based LLM Agents
• SkillProx: Self-Evolving Agent Skills via Proximal Textual Gradient Descent
• The Regression Tax: Decomposing Why Skills Help and Hurt LLM Agents

Tools & projects

• morluto/rea — Reverse engineer anything with agents, from app behavior down to native binaries.
• affaan-m/ECC — The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development f
• DietrichGebert/ponytail — Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.