Hongsen He

If you are building serious AI workflows, don't skip reading “The Complete Guide to Building Skills for Claude” playbook. 𝗙𝗶𝗿𝘀𝘁, 𝘄𝗵𝗮𝘁 𝗶𝘀 𝗮 “𝗦𝗸𝗶𝗹𝗹”? A Skill is simply a structured folder that teaches Claude how to perform a specific workflow consistently. Instead of repeating instructions in every chat, you define them once inside: - SKILL.md (required) - scripts/ (optional logic or validation) - references/ (docs loaded only when needed) - assets/ (templates, files, etc.) Think of it like this: - MCP = The kitchen (tools, connectivity, data access) - Skills = The recipes (how to use those tools properly) 🧠 𝗖𝗼𝗿𝗲 𝗗𝗲𝘀𝗶𝗴𝗻 𝗣𝗿𝗶𝗻𝗰𝗶𝗽𝗹𝗲𝘀 1. Progressive Disclosure Skills load in 3 levels: - YAML frontmatter → Always loaded. Decides when to trigger. - Main instructions → Loaded when relevant. - Linked files → Loaded only if needed. 2. Composability - Multiple skills can run together. - Each skill should work well with others. 3. Portability One skill works across: - Claude.ai - Claude Code - API - Build once, reuse everywhere. 𝗧𝗵𝗿𝗲𝗲 𝗖𝗼𝗺𝗺𝗼𝗻 𝗦𝗸𝗶𝗹𝗹 𝗧𝘆𝗽𝗲𝘀 1️⃣ Document & Asset Creation Example: Generate consistent frontend designs or structured reports. 2️⃣ Workflow Automation Example: Multi-step sprint planning across Linear. 3️⃣ MCP Enhancement Example: Code review using Sentry MCP with domain rules embedded. 🧩 𝗧𝗵𝗲 𝗠𝗼𝘀𝘁 𝗜𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁 𝗣𝗮𝗿𝘁: 𝗬𝗔𝗠𝗟 𝗙𝗿𝗼𝗻𝘁𝗺𝗮𝘁𝘁𝗲𝗿 This decides whether your skill triggers. Minimum required: --- name: sprint-planner description: Manages sprint planning workflows in Linear. --- Rules: - Must include WHAT it does - Must include WHEN to use it - nder 1024 characters - No XML tags - Name in kebab-case File must be named exactly SKILL.md If this is vague, your skill will never trigger correctly. 🧪 𝗧𝗲𝘀𝘁𝗶𝗻𝗴 𝗠𝗮𝘁𝘁𝗲𝗿𝘀 𝗠𝗼𝗿𝗲 𝗧𝗵𝗮𝗻 𝗪𝗿𝗶𝘁𝗶𝗻𝗴 You must test 3 things: 1️⃣ Triggering - Does it load on obvious queries? - Does it load on paraphrased queries? - Does it avoid unrelated queries? 2️⃣ Functional Accuracy - Are tool calls successful? - Is output correct? - Are edge cases handled? 🧱 𝟱 𝗣𝗼𝘄𝗲𝗿𝗳𝘂𝗹 𝗗𝗲𝘀𝗶𝗴𝗻 𝗣𝗮𝘁𝘁𝗲𝗿𝗻𝘀 1️⃣ Sequential Workflow Step 1 → Step 2 → Step 3 Perfect for onboarding, billing, account setup. 2️⃣ Multi-MCP Coordination Figma → Drive → Linear → Slack Move data across systems automatically. 3️⃣ Iterative Refinement Generate → Validate → Fix → Repeat Great for report generation or structured outputs. 4️⃣ Context-Aware Tool Selection Choose tool based on file size, type, or context. 5️⃣ Domain-Specific Intelligence Embed compliance rules or business logic before calling tools. This is where real differentiation happens. If you are serious about building production-grade AI workflows, this guide is worth your time. Save this post for later. Share it with your AI team. Follow Shubham Vora to learn more about scaling your business with AI.

101

32 Comments

Here are some simple habits that make your GPU profiling more useful when making AI models run faster Profilers (e.g. Nsight Compute) report tiny timing/counter diffs. So, run‑to‑run variance, different inputs, or host jitter can swamp the signal. If you profile different shapes/inputs each time, or switch between your macOS viewer and a remote Linux GPU box without locking inputs, you’ll “see” a lot of changes that are just noise. So, here's your 5-minute checklist for your next profiling run: 1. Freeze the workload a. Choose 1 representative input/shape (e.g., batch=16, seq=4096, hidden=8192). b. Fix seeds: CUDA_LAUNCH_BLOCKING=0, torch.manual_seed(1337), np.random.seed(1337) c. Disable randomness 2. Pin the environment a. Same container image, driver, CUDA toolkit, clocks. b. Make sure the remote GPU box is otherwise idle. 3. Script it a. One command that: warms up → runs the golden case → emits an NCU report. 4. Name your reports by hypothesis a. golden_baseline.ncu-rep, golden_prefetchL2.ncu-rep, golden_unroll4.ncu-rep. b. Compare only golden vs golden. Don’t mix shapes. 5. Use stable metrics a. Prefer counters less sensitive to jitter: achieved occupancy, eligible warps per cycle, DRAM throughput, L2 hit rate, stall breakdowns, executed instructions. b. Treat small wall‑time changes (<2–3%) as noise until multiple runs agree. #GPUProfiling #NsightCompute #CUDA #PerformanceEngineering

35

** How I use multiple LLMs for interview prep — as a system, not a shortcut As interviews become more senior and nuanced, relying on a single AI tool stopped making sense. Instead, I use different LLMs for what they do best—almost like a comparative matrix. 1. Gemini (Live + Deep Research): real-time mock interviews, STAR story stress tests, company and strategy decoding 2. NotebookLM: personal knowledge base—story banks, interview debriefs, pattern detection across feedback 3. ChatGPT: structuring responses, sharpening executive framing, testing logic and trade-offs 4. Claude: refining narrative clarity, tone, and leadership presence in written and spoken answers Why this works: Each model plays a distinct role—interviewer, editor, strategist, or mirror. Together, they surface blind spots, tighten signal, and turn preparation into a continuous improvement loop. AI doesn’t replace preparation. It raises the floor and sharpens the ceiling—if used intentionally. For anyone preparing for senior TPM, Program Ops, or CoS-style roles, this multi-LLM approach has quietly worked in my favor. #lifeofTPM #AILearning #LLMs

5

If you’re an AI engineer or an AI leader looking to level up your skills in LangChain and LangGraph, this one is for you. I’m teaming up with Arvind Narayanamurthy for a beginner–intermediate, 1.5 hour hands-on workshop where you’ll learn how to move from simple prompts to real AI applications and agentic workflows. Why this matters - Most engineers know how to prompt. Few know how to orchestrate. - LangChain and LangGraph are becoming the backbone of production AI systems. - If you want to build or lead AI engineering teams, you need these mental models. What you’ll learn - Foundations of LLM orchestration and where LangChain fits - Essential building blocks: chat models, prompting patterns, structured outputs, LCEL - Tool calling with MCP adapters - RAG basics end to end - LangGraph workflows and agentic architectures - Two live labs: a micro-app and a simple agent - Quick dive into LangSmith for debugging The workshop is recorded, and everyone gets the replay. If you want practical, real-world skills to build modern AI systems, this is the one to sign up for. Grab your seats: https://lnkd.in/eW2WZdcf

328

36 Comments