About Me

Asaqe Lee — operator of this site. Two public channels: reproducible infrastructure records, and alignment essays on AI agents.

About

Day-to-day work sits between self-hosted ops and AI engineering: Cloudflare Tunnel, Nginx Proxy Manager, new-api, Headroom, OpenClaw / MCP assistants, and the failure modes that appear only after something has been running for weeks.

Publishing is low-frequency. Infra posts ship after the command has been run on this host. Essays ship with a boundary and a way to tell whether the system is actually reliable.

Alignment with Interaction

Long-form notes on agent design: routing and ownership, memory and inheritance, when to split work across agents, and why prompt quality is mostly a thinking problem.

Open /writing/

System Record / Infrastructure Stack

Deployment and incident logs. Reverse proxies, tunnels, API gateways, Discord/Telegram adapters, Docker DNS, permission drift. Follow-along commands, plus rollback.

Open /lab/

Writing scope

Self-hosted OpenAI-compatible gateways and proxy layers
Reverse proxy, tunnel, and production hardening
Reproducible Docker / Compose configuration
Agent routing, boundaries, memory, and reliability
Verified commands and failure notes — not opinion dumps

Site architecture

Ghost, reading-first layout, two lanes that never share a stream. Canonical URL: https://me.asaqe.org. Mail identity lives on @asaqe.org.

Contact

Corrections and better approaches are welcome. Point them out directly.