News for hackers. Stuff that matters.

Bayram

This is a deep dive into how claude-sneakpeek's "swarm mode" coordinates multiple Claude agents to build software in parallel.

Show HN: Multi-agent deliberation plugin for Claude Code

https://github.com/BayramAnnakov/agent-tower-plugin

I kept copy-pasting the same question between Claude, Codex, and Gemini CLIs - one reviews the other's work, back and forth until I'm satisfied. Decided to automate myself out of the loop.This is Scott Page's "many model thinking" in action: each model has its own character and blind spots,so triangulating between them often catches things a single model misses.Three modes: - Deliberate: one...

( Read More... | no comments | 2 points | on HN )
Nope, at least not yet
I believe i cant edit that comment
Please try & let me know what you think - thanks
Yes, thats what I meant. Sorry if it is misleading
Thank you :-)
(thinking about cross-platform version, recalling my Java past)
sorry about that :(

Show HN: Retain – A unified knowledge base for all your AI coding conversations

https://github.com/BayramAnnakov/retain

Hey HN! I built Retain as the evolution of claude-reflect (github.com/BayramAnnakov/claude-reflect).The original problem: I use Claude Code/Codex daily for coding, plus claude.ai and ChatGPT occasionally. Every conversation contains decisions, corrections, and patterns I forget existed weeks later. I kept re-explaining the same preferences.claude-reflect was a CLI tool that extracted learnings...

( Read More... | 16 comments | 2 points | on HN )
v2 of the plugin is released: now it analyzes your session history to find repeating tasks that could become reusable commands.
Released v2 that addresses some of the issues reported in comments + windows compatiblity
Interesting approach. Since this tool suggests claude.md entries based on your corrections and feedback, and there is a human in the loop for you to precisely control whats added, I think it should work fine for your setup.

Wdyt?

Created an issue to track , will fix tomorrow: https://github.com/BayramAnnakov/claude-reflect/issues/2
Sorry, i will fix that
Agree, thats why I did 2 things: 1) HITL - you review/approve changes 2) Added some instructions around limiting the claude.md size and keeping it concise
2 thoughts: 1) they are, but they wont catch unique patterns very specific to your project 2) privacy - afaik they only train if you have model improvement enabled.

here is their privacy policy: https://privacy.claude.com/en/articles/10023580-is-my-data-u...

but, again, in my case this is more about my project/user specific corrections and preferences

Your workflow sounds solid - and honestly more disciplined than most. The difference I was going for: catching implicit corrections you don't consciously document.

So probably complementary to what you're doing, not a replacement. If you're already doing post-phase reviews, you might catch most of this anyway.

Re:Show HN: Agent/Claude skill for creating ChatGPT Apps on Sunday December 21, 2025 @10:52PM (#46351140)
This skill guides AI agents through the complete lifecycle of building ChatGPT Apps: 1) Fit Evaluation - Determine if a ChatGPT App is right for your product 2) Architecture Design - Plan MCP servers, tools, and widgets 3) Implementation - Build with Node.js, TypeScript, and React 4) Testing - Local development with ngrok, ChatGPT integration testing 5) Deployment - Production hosting and App Store submission
Re:How do salespeople use AI? on Friday December 12, 2025 @05:59AM (#46242932)
Yep, totally agree

Older »