By Michael McGarrah ·
13 min read
Kiro's AI agent can execute shell commands in WSL2, but the execution layer misbehaves without explicit Linux terminal profiles and shell integration. Here's how to fix the garbled output, path mangling, and false error codes.
By Michael McGarrah ·
19 min read
From v0.1.0-dev to v0.1.1 stable: building automatic update checking, session-aware restart logic, and testing infrastructure to ship reliable releases. This article documents the post-development improvements that turned a working prototype into production-ready software.
By Michael McGarrah ·
18 min read
Getting the Media Integrity Scanner into production — installing in Proxmox LXC containers, configuring for CephFS storage, setting up monitoring and alerting, and building the CI/CD pipeline for automated plugin releases.
By Michael McGarrah ·
9 min read
I wrote about managing AI context files across multiple agents in theory. Here's how it actually plays out in a real project — a Jellyfin plugin where Kiro, Claude Code, Copilot, Gemini, and Codex all need to know the same things.
🇪🇺 This site uses cookies to improve your experience and for analytics. As an EU visitor, you can control your privacy preferences. Read our Privacy Policy