opensourcesecurity

Extensible MCP server chassis for scoped, guardrailed AI access to real systems
Updated 2026-05-31 05:35:45 -07:00
Production enterprise-grade RAG AI built in 2024.
Updated 2026-05-28 08:09:31 -07:00
Open source transparency shell for Linux servers. 9-layer hardened SSH shell that lets reviewers verify live server state instead of trusting operator claims.
Updated 2026-05-27 15:59:19 -07:00
Self-hosted support portal in Elixir/Phoenix/LiveView with Postgres-everything stack. Replaces Zammad (500K+ LOC) in ~4,600 lines. Single-tenant, no SaaS dependencies.
Updated 2026-05-27 15:58:34 -07:00
A Claude Code skill that reviews a software portfolio via three parallel sub-agents, each applying a different engineering lens.
Updated 2026-05-27 09:32:13 -07:00
MCP server exposing Forgejo as LLM-callable tools
Updated 2026-05-26 11:10:02 -07:00
Multi-model local LLM evaluation for production technical support. A1 prompt template, Prism Porters Puzzle benchmark, failure-mode taxonomy.
Updated 2026-05-25 18:54:48 -07:00
Maximal security project on an open sourse BIOS/OS (OpenWRT) running an ARM based CPU.
Updated 2026-05-25 14:57:31 -07:00
Continuous head-to-head performance comparison of SafeNet vs. Proton and Mullvad. Methodology behind safenetvpn.us/proof.
Updated 2026-05-24 08:31:24 -07:00
Updated 2026-05-24 08:18:54 -07:00
Updated 2026-05-24 07:54:33 -07:00
Updated 2026-05-18 06:56:50 -07:00
Updated 2026-05-08 08:58:08 -07:00
Updated 2026-03-27 12:13:24 -07:00
Updated 2026-03-10 21:58:20 -07:00