Home

YAMS (Yet Another Memory System) started as a practical need: I wanted a dead-simple way to store and retrieve files, snippets, and research for LLM-driven workflows — without losing context between sessions. What I use daily is now ready to share. Note: YAMS is v0.7.x - experimental software under active development, not yet production-ready. Updated docs (2025-10-13): Site: https://yamsmemory.ai CLI: https://yamsmemory.ai/user_guide/cli/ MCP server: https://yamsmemory.ai/user_guide/mcp/ Deployment: https://yamsmemory.ai/operations/deployment/ Why YAMS? # Persistent memory for LLMs and tools Content-addressed storage (SHA-256) with block-level deduplication (Rabin) Compression (zstd/LZMA), crash safety (WAL) Fast search: SQLite FTS5 and semantic vector search Simple CLI and TUI; MCP server for Claude/Desktop and other MCP clients Now with alpha plugin support Quick install # Docker:

So I wanted to post a follow-up to my previous introduction to the umbrix platform, where I was using DSPy for entity extraction in cyber threat intelligence. If you missed that, you can find it here . This recap comes a week after I launched the platform. Two days into getting Umbrix running optimally, I started running into many problems. I have been extracting a significant volume of entities, relationships, and nodes from my agents to populate in the graph. However, I quickly realized my system was burning through resources - both in terms of rate limits and agent costs.

Over the past month, I’ve been working on a project for the Google ADK agent hackathon. This post provides an overview of my current multi-agent system, used for threat intelligence gathering, processing, and analysis. The motivation for Umbrix emerged when I was using a small language model to find A LOT of pcaps. I was attempting to seed the model with instructions on how to google dork, and feeded it with search terms to expand find publically reachable network security datasets. From that experiement, it dawned on me, there are many interesting applications for LLM’s. From there, timing and motivation was on my side. I set out to build this system with a simple thesis : *if the future is truly agentic, there are small building blocks and systems that need to be built to improve the efficiently gather and organize sources into a graph. From here I set out to design / vibe-code a system, able to improve how we fundemntally access information from security feeds by creating agents to efficiently gathering and organized sources into a graph.

Abbreviations Used # IMSI: International mobile subscriber identity RAN: Radio area network NGAP: Next-Generation Application Protocol GTP-U: GPRS tunneling protocol user plane I found it convenient to document and order 5G network attacks that I have come across in my ongoing research. I am writing this post as both a reference and a tool to focus on the bigger picture of 5G network security. I approach this work from both a network security and a security perspective, as I have relatively little knowledge of what is required to be a telecom operator. If there are any places where corrections are necessary, feel free to email me or contact me on my socials. Like other posts on this site, this will be a living document until it’s completed.

I have used FreeBSD as a server on and off for a couple years. After recently corrupting a research server running Ubuntu server, I found it imperative to migrate back to FreeBSD, setup ZFS snapshots and segment hosted services using the systems native hypervisor to improve my systems recoverability. If you have not checked out sanoid , I would highly recommend it. This post has been developed to capture that process and point to helpful resources that I have found along the way.