Pinned Loading

1. alexandria alexandria Public

   A personal knowledge engine that synthesizes instead of transcribes — every claim carries a citation back to its source. Write-time synthesis + cited retrieval for LLM agents.

   Python

2. atlas atlas Public

   QLoRA/SFT/DPO fine-tuning program and the 711-case eval harness (judge ensembles, live agentic stress tests) that proved a free hosted baseline wins

   Python

3. model-routing model-routing Public

   A PreToolUse gate for agent harnesses: classifier, human confirmation, Ed25519-signed single-use intents, red-team harness at recall 1.0

   Python

4. pi-king pi-king Public

   tmux-backed session manager and supervisor for the Pi coding agent. Background a session with /bg, see every session on the machine in one dashboard, reattach with a keypress.

   TypeScript

5. ring-mcp ring-mcp Public

   JavaScript

6. proposer-verifier proposer-verifier Public

   Split the agent in two: an LLM brain that can only propose, and a deterministic fail-closed signer. No LLM in the signing path.