Context Over Control: Rethinking What We Add to Agents
Why I retired my diagramming skill and uninstalled Superpowers: give agents clear goals, the context and tools they need, and explicit acceptance checks—then let them choose how to work.
Notes from practice
Why I retired my diagramming skill and uninstalled Superpowers: give agents clear goals, the context and tools they need, and explicit acceptance checks—then let them choose how to work.
How milestones, architecture decision records, and task specifications helped me keep fast-moving coding agents focused on the next release.
A compute-kernel example of moving state, retries, and acceptance checks from prompt instructions into the software around a coding agent.
Research agents benefit from stable goals and fast, automatic feedback. Application development needs the same foundations, plus an explicit boundary around what tests cannot decide.