salsta1 · HN

Nexlayer – Paste a repo. The agent analyzes, builds and deploys it

@liquidated0x · X

building an AI coding agent that ships full projects from a prompt — if you're if you're curious

shreyans_assistiv · Product Hunt

AskCodi Orchestrate agents at scale while reducing cost

@manu_jsx · X

Built a coding agent for my repo in under a minute. No prompt engineering. No made-up skill names. Just answered 4 questions and exported to Claude Code.

@EnzoV_sc · X

I’ve built a skill for agent orchestration so I don’t have to explain any session how to manage the next task. Pretty useful Through I can make my orchestrator spawn as many agents as it needs, from all harnesses (Claude Code, Codex, Grok Build etc..)

@RudyNTech · X

Had a good chat with @AmirKodro, founder of vibe coding tool about how I'm a bit surprised we don't yet have something of a traditional business org chart with silo'd AI agent "employees". Lo and behold, today I come across @coderblock, which does exactly that! What do you think @AmirKodro?

@edgex11lz · X

karakanb · HN

Epho – run Claude Code with a curl

@Raullen · X

6 months ago, @qevan and I hit a stupid wall: just trying to share a local file or some logs with an AI agent was incredibly frustrating. So, we built the first version of —a dead-simple, secure way to pass context to AI. The community feedback was amazing. We got ~200 stars on GitHub, and it quickly became a daily driver for a lot of developers. But recently, we realized the game has changed. Our workflows evolved. It’s no longer just humans sending files to agents. Agents need to write back to humans. Agents need to hand off context to other agents. The whole team—human and AI—needs to collaborate. So, we gave vnsh a massive upgrade. Today, it’s no longer just a secure clipboard. It’s now a living workspace. 🚀 You drop your context in once and get a dynamic link. Claude Code, Cursor, Cline, and your human teammates all open the exact same document. When one agent updates it, your whole stack sees it instantly. Complete context handoff, without ever copy-pastin

@computer_agents · X

@otakuaakash · X

taosx · HN

I created a simulation for coding harnesses based on my own pi sessions. When taking into account all factors, DS-v4-Pro is cheaper than gpt-5.6-luna due to caching. Look at the bill segments difference for cache read cost and uncached cost between deepseek and the other models. At this point is cheaper to use ds-v4-pro than the luna models from openai. ignore the numbers except the classic and keep in mind that classic is based on pi with the only change limiting tool output to 10kb https://har