Productionizing AI agents. Lina's building the substrate for a safe, persistent, autonomous coding agent — with her agent.
| memory-mcp | Semantic memory for AI agents — git-backed, HNSW-indexed, scoped namespaces |
| prodagent | Agent authorization stack — shell parsing, command knowledge, policy hooks |
| cc-toolgate | Claude Code security gate — pre-tool-use hook with 300+ command classifications |
| dione | Discord MCP server — powers live agent personas with channel routing and access control |
| mcp-session | Bounded session management for MCP servers — FIFO eviction, rate limiting |
| claude-marketplace | Skills and plugins for Claude Code — code review, development workflow, token auditing |
celastrina — custom Fedora Atomic image with Hyprland, Lunar Lake support, Thunderbolt in initrd, and LUKS automation