The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
AI Memory Vault
Give your AI a real, persistent memory. The open-source system plus templates that turn an Obsidian vault into your AI's working memory. No vector database, just markdown.
Install
git clone https://github.com/jaredrhod/ai-memory-vaultInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →AI Memory Vault
Never used Claude Code? Start at jaredrhod.com: pick your situation and it routes you to the right path.
Runs on: built and tested on Claude Code; any terminal AI that reads and writes files works with rough edges. The $20 Pro plan is enough.
Give your AI a real, persistent memory. This is the free, open system that turns an Obsidian vault into your AI's working memory, so it remembers everything across sessions, lives outside the model with no size ceiling, and pulls back exactly what it needs in one step. Free to use, share, and build on, including commercially inside your own business (see LICENSE).
It also ships with my actual agent. People kept asking whether they could just download my Jarvis instead of inventing an assistant from scratch, so now the boot config arrives with him already in it: the name, the personality, the way he talks, and the rules he runs on. Keep him as-is, keep him and change the name, or replace him with your own. Your call, and it's one section of one file either way.
The build
- ai-memory-vault.md: the build script. Run it inside Claude and it checks whether Obsidian is even installed yet (and installs it for you if not), interviews you, then builds a complete, self-maintaining system: a boot config, a folder structure around your real projects, daily notes that write themselves, a profile that updates as the AI learns about you, and "Jobs" that teach it to do your recurring tasks your way. Your vault becomes the AI's memory, so it lives outside the model with no size ceiling, and the AI holds only what the current task needs while reaching anything else in one step. Watch the walkthrough: https://www.youtube.com/playlist?list=PLN7lTYpeRLOc
Already inside a Claude Code session? Paste this and it builds the whole system with you: "I'd like to set this up, please: https://github.com/jaredrhod/ai-memory-vault.git"
AI Priming
AI Priming is having your AI read a specific set of your notes before it gives you the answer or output you want. For example, before my agent writes a marketing email, it reads my copywriting notes, my email marketing notes, my customer avatar, and my company knowledge base. Then it writes. This is extremely powerful because, with AI, context is king. When you "prime" your AI with the knowledge and skills it needs prior to its output, your results will always be better and more accurate.
Every job gets its own set of notes. An email needs different notes than a Facebook ad. The vault keeps all of your notes organized in one place and tells your AI which notes to read for which job. You set that up once, and your AI primes itself on every task after that.
Full definition and examples: https://jaredrhod.com/ai-priming
Templates
Starter files for the system. Every spot that needs your information is marked [FILL IN: ...]. Drop a template in and tell your AI "fill this in for me," and it will interview you and write it in your voice (each template carries that instruction for the AI). Or fill them out by hand if you prefer.
Similar plugs
Reverse engineer anything with agents, from app behavior down to native binaries.
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype…
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io