Tag
Canon is a local-first CLI tool that automatically injects project decisions into AI coding agents like Claude Code and Cursor to prevent them from re-adopting rejected approaches.
This article presents a simple pattern to provide LLM agents with memory for their decisions, enhancing their decision-making process.
A team reflects on six common structural failure points in AI builds: context, identity, decision memory, attention, write-back, governance, and economics, and offers a diagnostic tool based on their experiences.