@Teknium Most AI code tools send u'r code 2 the cloud. Every diff. Every secret.
Run 3 agents locally โ Reuse, Quality, Efficiency โ review your git diff. Zero data leaves ur machine.
And a 4th agent audits the output 4 secrets & dangerous calls before u accept a line.
Local. Private. ๐
forked repo. pinned dependency. built from source. git-protected. isolated workspace.
same functionality. zero blind trust. no auto-updates. no network fetches.
if u'r connecting AI to u'r filesystem, treat it like an intern with root access.
own u'r tools or u'r tools own you
saw that "AI brain" Obsidian setup going around
cool concept. terrible security.
it runs raw npx on every launch โ that's a supply chain attack waiting to happen. one compromised package and your local filesystem is cooked.
so i rebuilt it the right way.
THIS GUY CONNECTED HIS AI AGENTS TO HIS OBSIDIAN AND BUILT A BRAIN THAT LEARNS ON ITS OWN. HERE'S HOW TO BUILD IT
Obsidian is just markdown files sitting in a folder. That turns out to be the perfect memory for an AI agent, because an agent can read and write those files directly. He wired his agents into the vault so they pull context from it, do the work, and write what they learned back. The notes aren't the point. The loop is, and it gets sharper every cycle
How to build it:
1. Point an agent at your vault. The fastest way, no plugins, no API keys: open a terminal and run npx obsidian-mcp /path/to/your/vault. That exposes your Obsidian folder to Claude as a tool it can read, search, and write to. Add it to your Claude Code or Cowork config and restart
2. Confirm it can see the brain. Ask it: "list the notes in my vault and summarize what's in them." If it reads them back, the connection is live. Now it starts every task with everything the vault already holds instead of from zero
3. Give each agent one job and a write-back rule. Tell it: "research this, then save what you found as a new note in /brain with links to related notes." One agent researches, one summarizes, one plans. Each writes its output back into the vault
4. Close the loop. Add one line to every agent's instructions: "read /brain before starting, write your result back when done." Now each task leaves the vault richer, and the next run reads that before it works. It compounds instead of resetting
5. You only steer. Review what the brain produces, point it at the next thing. The agents handle the reading, writing, and connecting
The edge isn't better notes. It's a brain that feeds itself, so the work gets sharper every cycle instead of starting over
Bookmark this
@Aaronbuilds_@AlexFinn Its ok brodi. Some people just are peggers and some are pegged. One day you'll do the pegging. Until then, fries in the bag ser...
@undefinedKi This is sauce but very dangerous if you follow the exact directions. Not enough emphasis on the security protocol. If you are going to run this, make sure you triple check your security plan before executing.
Built my own AI agent orchestration layer tonight. #Hermes Agent + #Obsidian vault + local LLMs. Different models for different tasks, vault as shared memory. #Gemma 4 for general ops, #Qwen for research. No APIs, no cloud, everything local. Phase 1 starts tomorrow.
Algo flooding my feed with GitHub links, I built my first AI agent to filter the noise. ๐ค
It automatically scans repos to tell me if they're worth a look or a total dodge.
Running 100% locally qwen2.5:14b (9GB) / Ollama. ๐
Check it out ๐
#buildinpublic#localAI#opensource
If im using #qwen locally and #claude cloud to write all the code, but im the one navigating the terminal and executing the code or edits. Is that vibe coding?
In the past 4 days, I went from using ai chatbots to building my own local triage agent that I communicate with via telegram.
I have zero code experience just #Claude, #VSCode and my local #llm's.
I dont know what got into me but im am locked the fuck in...
@Humanarewild Excellent weight distribution, buddy threw everything he had into that overhand right. That's years of repression all released into that dudes jaw...