Tag:agentic-engineering
All the articles with the tag "agentic-engineering".
What's actually in your agent's context
How coding agents assemble model input, what fills the context across 264 real Pi sessions, and what compaction keeps or leaves out.
A minimal agent loop and the four ways it breaks
(updated )I build a minimal agent loop in TypeScript and run it on a local model. It stops too early, crashes on a bad tool call, overflows its context and trusts what it reads. Each failure points to a part of the harness.
Linux Lost the Human Desktop. Can It Win the Agent Desktop?
The config files, shell commands, manuals, and source code that frustrated normal users may be exactly the handles AI agents need.
Pi: A coding agent for engineers who own their tools
Claude Code, Codex, OpenCode - great places to start, not best places to finish. Pi is the thin harness that gets out of your way. Here's why experienced engineers should pay attention.