ITWorx

I recently took part as a mentor in the CASSINI Hackathon "Space for Water", organised by Fundació KIMbcn and Universitat Politècnica de Catalunya (UPC) at the Campus del Baix Llobregat. Over 100 people working through a weekend on water-related challenges, leveraging Copernicus and Galileo data. Beyond the results, what I'd like to highlight is the value of the process itself: students and industry professionals sharing perspectives, languages and ways of solving problems. At Zenithal Blue , we believe these spaces are essential for knowledge and experience transfer. Thanks to the organisers for the invitation and looking forward to collaborating with the next generation of Earth observation experts. #cassiniEU #EUSpace #UPCtech

Thinking about selling but not sure where to start? You don’t need a month-long renovation to make a big impact. If you’ve got 48 hours this weekend, here are 3 quick things you can do to boost your home’s value and catch a buyer’s eye: 1. Declutter the Entryway: First impressions are everything. Clear out the extra shoes, coats, and mail. A clean, open entrance makes the home feel larger and more inviting the moment someone walks in. 2. Swap Out Old Light Bulbs: Replace any flickering or mismatched bulbs with daylight or warm white LEDs. Consistent, bright lighting makes every room look cleaner and more modern in photos and in person. 3. Refresh the Front Door: A quick wipe-down of the door and polishing the hardware can work wonders. If you're feeling ambitious, a fresh coat of paint in a classic color adds instant curb appeal. Small moves can lead to big results when you're getting ready to hit the market. If you want a personalized checklist for your specific home, I’m always here to help. Just send me a text or give me a call! #gratituderealtor #guidingyouhome #lovewhereyoulive #denverrealestate

Claude Code tips I wish I had from day one. 📁 CLAUDE.md is non-negotiable Run /init on any new repo and keep the file concise — bloated context files hurt more than they help. For large repos, scope module-level CLAUDE.md files and reference them from the root. Pro tip: if you use multi-agent tools, add an AGENTS.md and a BOOTSTRAP.md and point both to CLAUDE.md as a single source of truth — works across Claude Code, Cursor, Codex, Gemini CLI. ⌨️ Shift+Tab — cycle through modes Always start complex tasks in Plan mode. The flow that works: Explore → Plan → Code → Commit. Don't skip the plan step. 📊 /context — your window into what Claude actually sees Run it regularly as you build. It'll show you how much context you're burning and where you can optimize. Sub-agents don't share the main thread context — use them for non-core tasks where the main agent only needs the result, not the full detail. This is one of the biggest token savers. 💭 /insights - Generate a report analyzing your Claude Code sessions, including project areas, interaction patterns, and friction points 🗜️ /compact — use it in long sessions Don't wait until you're hitting limits. Compact proactively. 🔁 /skill-creator — package what you keep repeating If you find yourself re-typing the same prompt, running the same multi-step process, or wanting to standardize a workflow across your team — use /skill-creator. It interviews you, then generates a SKILL.md file: a portable, invokable slash command with optional scripts and reference files bundled inside. Works across Claude Code, Cursor, Codex, and Gemini CLI. Not vendor-locked. 🔍 /review before any agentic PR Always run it in a fresh session — not the one that wrote the code. If you have GitHub / Azure DevOps / GitLab MCPs or their CLIs enabled, Claude can create the PR, add inline comments, open review threads, and flag issues automatically. Keep a human in the loop for the final merge. 🧠 /advisor — native Opus consultation Type /advisor inside Claude Code and pick Opus. From that point, Sonnet handles execution and silently consults Opus when it hits a decision it can't confidently resolve alone — architecture choices, non-trivial refactors, recurring failures. No manual wiring needed. Quality close to Opus, cost close to Sonnet. ⚙️ settings.json vs settings.local.json — project-level vs user-level Stay tuned, covering this one separately. What tip would you add? Drop it in the comments. #ClaudeCode #AIEngineering #DeveloperProductivity #Anthropic #AgenticAI #AIWorkflows #SoftwareDevelopment #BuildingWithAI #ClaudeAI #DevTools

Two tools I use daily just became significantly more powerful together. Claude Code is already my go-to agentic workflow client — the way it designs and executes tasks, the advisor strategy, the built-in reasoning patterns. Most people think of it as a coding tool, but honestly the ceiling is much higher than that. You can wire it into workflows that have nothing to do with code. NotebookLM on the other hand is my default for presentations, deep research, ideation, or getting up to speed on a topic fast. Upload your sources, ask anything, get grounded answers backed by what you fed it — no hallucinations. Now imagine both linked through a bridge. @hooeem published a full guide (link in comments) on exactly this — using an open-source CLI tool called notebooklm-py (reverse-engineered by Teng Ling, repo below) to connect Claude Code and NotebookLM programmatically. The memory workflow stood out most to me: Claude wraps up each session, pushes a structured summary into a "Master Brain" NotebookLM notebook, and queries it at the start of the next session via CLAUDE.md . Your context survives. Your token budget doesn't take the hit. The other use case that immediately clicked for me: building on-the-fly custom agents that run on Claude's models but pull their knowledge base directly from NotebookLM. Domain-specific, grounded, cheap to run. Key resource from the guide 👇 🔗 https://lnkd.in/eQRjeZwX Full credit to @hooeem for putting this together — original post linked in comments. #ClaudeCode #NotebookLM #AIWorkflows #Anthropic #GoogleAI #AIAgents #DeveloperProductivity #KnowledgeManagement #FrontierAI #BuildingWithAI