I build agent systems that can think, coordinate, report back, and still have a personality.
- Building emotion-aware agent experiences that make AI intent visible
- Exploring coding harnesses, multi-agent orchestration, and durable verification
- Connecting agent work to the real world through hooks, event routing, Discord, and tmux
- Obsessed with small, sharp tools that compose into dependable systems
|
A service-first emotion layer for AI agents. It reads each response, detects intent and emotion, then attaches the right character image through OpenClaw or Hermes.
|
A red-claw coding-agent harness built around one disciplined loop: deep interview, resilient planning, tmux-native team execution, and durable verification.
|
|
A daemon-first notification router for agent operations, with typed events, provider-native hooks, Git/GitHub/tmux sources, and Discord or Slack delivery.
|
An opinionated multi-model OpenCode harness that orchestrates specialist agents, background work, LSP/AST tooling, tmux sessions, and relentless ultrawork loops.
|
Expressing Agent intent and emotion with Hent-ai
Orchestrating Coding workflows with Gajae-Code and OpenAgent
Routing Hooks, events, and notifications through clawhip
Shipping Agent infrastructure with a visible personality
Sharp tools. Loud personality. Real output.