RT Avi Chawla (@_avichawla)What are Agent Skills?An agent with 30 specialized workflows could need ~150k tokens in its system prompt if you loaded everything upfront.With Agent Skills, that drops to ~3k tokens. The agent knows what skills exist but loads full instructions only when the task needs them.Here's how skills exactly work in 7 stages:> Stage 0: Skills discoverySkills are discovered from multiple sources, like:- Bundled (ships with the platform)- project-level (.claude/skills/)- global (~/.claude/s ...

What are Agent Skills?An agent with 30 specialized workflows could need ~150k tokens in its system prompt if you loaded everything upfront.With Agent Skills, that drops to ~3k tokens. The agent knows what skills exist but loads full instructions only when the task needs them.Here's how skills exactly work in 7 stages:> Stage 0: Skills discoverySkills are discovered from multiple sources, like:- Bundled (ships with the platform)- project-level (.claude/skills/)- global (~/.claude/skills/)- plugins- API- comm ...

Core Insights - The article emphasizes that using an Agent without installing Skills is akin to having a smartphone without apps, limiting its capabilities [1] - The essence of Skills is not merely as plugins but as a means to inject industry best practices, real project experiences, engineering constraints, and structured thinking frameworks into the Agent [4] - Installing the first five Skills can significantly enhance productivity [5] Summary by Sections Ordinary Agent Limitations - Ordinary agents have notable deficiencies, such as the inability to write engineering-level code, understand real-world rules, and adhere to design specifications [6] Recommended Skills - The article lists 20 Skills, highlighting the top five that should be installed first for immediate productivity benefits [5] 1. **find-skills**: Enables the Agent to actively search, discover, and recommend other Skills, effectively creating a skills marketplace for the Agent [8] 2. **vercel-react-best-practices**: Contains over 40 performance rules for React/Next.js, helping avoid performance pitfalls [9] 3. **frontend-design**: Teaches the Agent to create high-quality, non-template UI, improving aspects like color schemes and interface hierarchy [10] 4. **web-design-guidelines**: Essential for code review, includes over 100 web accessibility, UX, and performance standards to avoid basic UX errors [11] 5. **remotion-best-practices**: Best practices for creating short videos using React, covering animation, export, and performance optimization [13] Additional Skills - Other Skills mentioned include: - **brainstorming**: Aids in structured brainstorming and workflow planning [14] - **agent-browser**: Automates browser tasks like web browsing and form filling [15] - **supabase-postgres-best-practices**: Optimization guide for Supabase and PostgreSQL databases [17] - **azure-cost-optimization**: Rules for optimizing Azure costs, potentially saving significant cloud service expenses [19] - **code-review-expert**: Enhances code review quality by covering SOLID principles, security, and performance [29] Skills vs. Model Size - The article argues that the real differentiation in Agent performance is not based on model size or parameters but rather on the rules defined through Skills [30][31] - It posits that while models set a lower limit, Skills determine the upper limit of an Agent's capabilities [32] - The future developer's role will evolve from merely writing code to defining rules and shaping AI [34]

RT BNB Chain Developers (@BNBChainDevs)BNB Chain Skills Hub is now live!A community-curated registry where builders can submit agent skills as JSON files and enriched with owner info, latest commit and a @GoPlusSecurity audit report.This makes it easier for BNB developers to discover and use agent skills across the ecosystem.Submit your agents skills 👇https://t.co/vP21ZI3QKY ...

Building #BNBAgents over the weekend?Share them below this post and we'll pick up some cool ones 👇👀https://t.co/R5Ns6Pp03jBNB Chain (@BNBCHAIN):We’re pushing agent innovation forward by supporting ecosystem projects shipping agent skills.If you’re building agent skills for BNB developers, we want to know.Comment below with what your agent skill does, how developers can set it up, a GitHub link and include the hashtag https://t.co/vgph7maxko ...

Core Insights - The paper "SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks" reveals critical truths about the development of AI agents, emphasizing that agents cannot self-teach new skills effectively [2][40] - The research involved 36 scholars from top institutions and tech companies, highlighting the collaborative effort in understanding agent skills [2] Group 1: Agent Skills Overview - Agent Skills are structured knowledge packages that enhance LLM Agents during inference, differing fundamentally from traditional prompts and tools [5] - The Skills ecosystem is experiencing rapid growth, with a total of 84,192 skills created within 136 days, averaging 810 new skills daily [8] - The paper establishes a benchmark for evaluating the effectiveness of these skills, addressing the lack of standard methods in the industry [9] Group 2: Research Design and Methodology - The research design involved three phases: aggregation of skills, quality screening, and evaluation across various conditions and agent models [14][15] - A total of 86 high-quality tasks were selected from 322 candidates, covering 11 domains, ensuring rigorous testing standards [15][18] Group 3: Key Findings - Finding 1: Expert-built skills resulted in a significant average success rate increase of 16.2 percentage points, demonstrating the value of human expertise in skill development [20] - Finding 2: AI-generated skills were found to be ineffective, leading to a decrease in success rates by 1.3 percentage points, challenging the narrative of self-evolving agents [22][23] - Finding 3: The effectiveness of skills varies significantly across domains, with healthcare and manufacturing showing the highest leverage effects [24][26] - Finding 4: Smaller models combined with skills outperform larger models without skills, indicating a shift in strategy towards optimizing skill integration rather than solely focusing on model size [27][29] Group 4: Engineering Insights and Industry Implications - The research indicates that providing 2-3 skills yields peak performance improvements, while excessive skills lead to cognitive overload and diminished returns [31] - A focus on detailed and targeted skills documentation enhances performance, contrasting with comprehensive documents that may hinder effectiveness [32] - The findings suggest a strategic shift in AI development, emphasizing the importance of high-quality vertical skills over merely scaling model parameters [35][36]

We're still compiling agent skills for BNB builders 👀👇Share under this thread with #BNBAgents:• What your agent skill does• How devs can set it up• Repo linkNote: Always DYOR and assess potential security risks before using any tools shared in this thread. ...

Core Insights - Anthropic has completed a $30 billion Series G funding round, achieving a post-money valuation of $38 billion and an annual recurring revenue of $14 billion, with an average growth rate exceeding 10 times over the past three years [1][2]. Funding and Investment - The funding will support advanced research, product development, and infrastructure expansion, with Claude being the only AI model available on major cloud platforms: Amazon Web Services, Google Cloud, and Microsoft Azure [2]. - Key investors in this round include the Ontario Teachers' Pension Plan and Coatue, along with significant contributions from firms like Blackstone, Goldman Sachs, JPMorgan, and Sequoia Capital [4][5]. Customer Growth and Revenue - The number of Claude customers spending over $100,000 annually has increased sevenfold in the past year, with over 500 clients now spending more than $1 million annually, including 8 out of the top 10 Fortune 500 companies [6]. - Claude Code, launched in 2025, has generated over $2.5 billion in annual recurring revenue, doubling its growth since early 2026, and now accounts for more than half of total revenue [6][11]. Product Development and Features - Anthropic has released the Claude Opus 4.6 model, which features a 1 million token context window, enhancing its capabilities in long-context queries and reasoning [7][10]. - Claude Code has become a popular tool among developers, allowing for automated engineering tasks and code management, with a graphical UI version called Cowork launched to cater to non-technical users [9][10]. Market Position and Strategy - Unlike OpenAI's consumer-focused approach, Anthropic emphasizes the B2B market, achieving significant commercial success with a focus on transforming technological advancements into market advantages [11]. - The company is preparing for an IPO, potentially in 2026, having engaged legal and investment banking firms for preliminary preparations [10].

Core Viewpoint - The article discusses a significant review paper titled "AI Meets Brain," which bridges cognitive neuroscience and artificial intelligence, focusing on how human memory mechanisms can inform the development of human-like memory systems in agents [2][6]. Summary by Sections Memory Definition - Memory is redefined as not just data storage but as a cognitive link that connects past experiences with future decisions, involving a two-stage process in the human brain [6]. Perspectives on Memory - From a cognitive neuroscience perspective, memory serves as a bridge between past and future [6]. - For large language models (LLMs), memory exists in three forms: parametric memory, working memory, and explicit external memory [7]. - Agent memory transcends simple storage, functioning as a dynamic cognitive architecture that integrates experiences and environmental feedback [8]. Importance of Memory - Memory plays a crucial role in enhancing agent capabilities by overcoming context window limitations, building long-term personalized profiles, and driving experience-based reasoning [12][13]. Memory Classification - The review categorizes memory based on cognitive neuroscience definitions, distinguishing between short-term and long-term memory, with long-term memory further divided into episodic and semantic memory [15][21]. Memory Storage Mechanisms - Memory storage in the human brain involves dynamic cooperation across brain regions, while agent memory systems are explicitly engineered to optimize data structure selection for computational efficiency [31][32]. Memory Management - Memory management in agents is a continuous process involving extraction, updating, retrieval, and application, contrasting with the static nature of traditional memory systems [33][34]. Future Directions - Future agent memory systems should aim for omni-modal capabilities, integrating various data types beyond text, and facilitating skill transfer across different agents [49][50].

Claude can now connect to @benchling, PubMed, and https://t.co/poBaIt4Qz9, among other platforms. And our new Agent Skills can help Claude follow scientific protocols and procedures consistently. ...