AI Agent System Framework on GitHub

This is wild. A developer just recreated Anthropic’s Claude Code system… in hours. Not leaked. Not copied. Rebuilt using AI. I explored the viral claw-code repo — and most people are missing the real story 👇 🔗 https://lnkd.in/d24DTZuf 🧠 This isn’t just a “clone” It’s a clean-room rebuild of an AI coding agent. No original code. Just behavior + architecture recreated from scratch. ⚙️ What’s actually happening under the hood: This system runs on a simple but powerful loop: → Think (LLM) → Plan steps → Use tools → Store memory → Repeat That’s it. But that loop changes everything. 🔥 Why this matters (big shift) AI is no longer just answering questions. It’s now: • Thinking in steps • Taking actions • Building things autonomously 💡 The biggest takeaway: We’re moving from: Prompt Engineering ❌ to Agent Engineering ✅ The real skill now is designing systems where AI can: think → act → improve 🚀 My take: This repo isn’t production-ready. But it’s a glimpse into the future of software. And that future is coming fast. If you're building in AI, you should be paying attention. Comment “AGENT” and I’ll share: → Full breakdown → How to run it → How to build your own #AI #GenerativeAI #SoftwareEngineering #TechTrends #Developers #MachineLearning

Anthropic accidentally published their entire Claude Code source code to NPM. Here's what the leak reveals about how AI coding tools actually work: Anti-distillation defenses. Claude Code injects fake tool definitions into API calls. If competitors are recording traffic to train copycat models, they get poisoned data. A creative technical countermeasure. "Undercover mode." When Anthropic employees use Claude Code on external repos, the AI is instructed to never mention internal codenames — and crucially, to not reveal it's an AI. There's no way to turn this off. Frustration detection via regex. Yes, they use a simple regex (not the LLM) to detect if you're swearing at the tool. Sometimes the boring solution is the right one. KAIROS: an unreleased autonomous agent. References throughout the code point to a background daemon with "nightly memory distillation," GitHub webhooks, and 5-minute cron jobs. This appears to be Anthropic's next big feature. The full analysis is worth reading. This is rare visibility into how frontier AI companies actually build products. What surprised you most about these findings? #aiwithsai #ai #claudeai #aitools #machinelearning https://lnkd.in/dWm7wfCd

Recently explored an interesting open-source project: code-review-graph It builds a local knowledge graph of your codebase, mapping functions, classes, and dependencies — so AI tools only read what actually matters instead of scanning entire repos. Result:  ▶ Faster code reviews  ▶ Better context understanding  ▶ Significant reduction in token usage What stood out is its ability to track changes incrementally and quickly identify impacted areas, making large codebases much easier to navigate. Feels like a step towards solving one of the biggest problems in AI coding: context, not capability. 🌐 https://lnkd.in/dRKnbqVj 👏 Kudos to Tirth Kanani for building this! #AI #CodeReview #OpenSource #DeveloperTools #LLM

The Future of AI Apps is NOT Just Prompting… It’s Agents If you’re still building AI apps with just prompts and APIs, you might be missing the bigger shift. Recently, OpenAI released openai-agents-python — and it’s a glimpse into what production-grade AI systems will actually look like. Here’s why this matters 👇 🔹 From Single Prompts → Multi-Step Reasoning Instead of one-shot responses, agents can plan, act, observe, and improve. 🔹 Tool Usage Becomes Native Agents can call APIs, run functions, fetch data — just like a junior developer would. 🔹 Memory & Context Handling No more hacking long prompts. Agents manage context more intelligently. 🔹 Composable Architectures You can build systems where multiple agents collaborate (think: researcher + analyst + executor). 💡 What this means for developers: If you’re working with: Python 🐍 LLMs Automation workflows You should start thinking beyond “chatbots” → and start building AI systems. ⚠️ But here’s the reality: Agents are powerful… But they also introduce: Latency challenges Higher costs Debugging complexity So don’t blindly jump in — use them where they actually add value. 🔥 My Take: The winners in AI won’t be the ones who know the best prompts… They’ll be the ones who can design: 👉 agent workflows 👉 tool integrations 👉 decision systems I’ve dropped the repo here : githubrepo : https://lnkd.in/gfPmW4CA Curious to hear — are you already experimenting with AI agents, or still exploring? #AI #MachineLearning #OpenAI #Python #AIAgents #GenAI #Automation #Developers #Tech

One big problem with AI coding tools today, they repeatedly scan large parts of the codebase for every prompt in which token usage shoots up (cost + latency), context gets bigger and complex, response hallucinates. Came across a neat repo solving this: code-review-graph It builds a structured graph of your code (functions, classes, dependencies), so AI tools can fetch only what’s actually relevant instead of re-reading everything. Result is much lower token usage, Cleaner, focused context Feels like a simple but powerful step towards better context engineering for AI dev workflows. https://lnkd.in/dCwABgN9        https://lnkd.in/duXNQaj2 #AI #LLM #DeveloperTools #OpenSource

While scrolling through LinkedIn, I found an interesting article (Sebastian Raschka, PhD Sebastian Raschka that clearly explains the real architecture behind coding agents and how they are built) about AI agents. The author clearly explained the structure of agents, how they perform tasks, and how they are built. Below is the link to the article, and here is what I understood from it. The article explains that AI coding agents are not just large language models they are systems built around LLMs using tools, memory, and context. In real-world applications, the surrounding system plays as important a role as the model itself. From my understanding, the architecture of a coding agent mainly includes these components: 1. Live Context (Workspace Awareness) The agent first understands the project environment — repository structure, files, and instructions — before taking action. This prevents it from working blindly and improves accuracy. 2. Prompt Structure & Cache Reuse Instead of rebuilding prompts every time, agents reuse stable instructions and only update new inputs. This makes them faster and more efficient. 3. Tool Usage Agents don’t just generate text; they can use tools like reading files, running commands, or editing code. This makes them act more like real developers rather than chatbots. 4. Context Management To avoid overload, agents compress old information and keep only relevant data, ensuring better performance over long sessions. 5. Memory System Agents maintain both working memory and full transcripts, allowing them to remember past actions and continue tasks logically. 6. Sub-agents & Delegation Advanced agents can delegate smaller tasks to sub-agents, improving efficiency and parallel problem-solving. My Key Takeaway: AI agents are not just “smart models” they are structured systems combining memory, tools, and reasoning loops. The future of AI productivity will likely depend more on how we design agent systems rather than only improving models. Here’s the article: https://lnkd.in/gdD7dEdX github: https://lnkd.in/gmCf_H5e. #AIAgents #CodingAgents #LLM #MachineLearning #AI #GenerativeAI #DeepLearning

I'm sharing with you this very interesting AI code assistant project called Nexus (link below) it analyses your entire project written in any of 30 languages and outputs basically an optimized AST (abstract syntax tree). but it doesn't sit there doing nothing: it provides your AI a skill to navigate that tree with a query language. It has the huge advantage of your AI not guessing anymore from the few files it reads and "imagines the rest" the AI has the real situation accessible in a token saving manner. meaning you can use your context window to do what you are paying for: write correct code. It starts making you productive in new discussions immediately, sparing you the overhead of discovery on each discussion start. Additionally, it and your git history to detect the hot spots you're focused on. I'm sharing my journey writing a unique multidimensional multi-tenant platform with multiple innovations. It is materialized in Svelter(.me) as a proof of concept playground. A platform dedicated to NIS2 is in the oven and will hopefully bring huge value. Follow me for more feedback from real world AI use. #AI #SKILLS #Claude https://lnkd.in/dTWEUm_B

Reduce your Claude AI code token usage by 6.8× fewer tokens on reviews and up to 49× on daily coding tasks with a single command. AI coding tools often re-read your entire codebase for every task. That means unnecessary token consumption, slower responses, and higher costs. code-review-graph solves this. It builds a structural map of your code using Tree-sitter, tracks changes incrementally, and feeds your AI assistant only the relevant context via MCP. The result? ⚡ Faster responses 💸 Lower token usage 🎯 More precise code understanding 🔗 GitHub: https://lnkd.in/gbeCdG8z Commands: pip install code-review-graph code-review-graph install code-review-graph build #tree #claude

Stop burning AI credits on context AI doesn't need. If you use Claude or Cursor, you've felt this: You ask one question. AI reads your entire codebase. The result? → Credits gone → Slow answers → Still average output You try writing .md files to help it. AI still misses what actually matters. The real problem isn't the AI. It's the input. AI doesn't know what to focus on so it reads everything. That's where two tools are changing the game: 1.Graphify Maps your entire codebase so AI understands connections, not just files. Link : https://lnkd.in/d2wg9xsb 2. Code Review Graph (officially contributed to by Claude) Filters your code so AI reads only what's relevant to your question. Link : https://lnkd.in/d8W77tEa The shift happening right now in AI-assisted coding: ❌ AI reads everything ✅ AI reads only what matters Less noise. Faster answers. Better output. Simple way to think: Use Graphify → to understand the system Use Code Review Graph → to work faster with AI Once you start using this approach you will see a big difference in output and credit usage