For agent users and builders
Field notes from agent work
Arguments, lessons, and teardowns from building and reviewing agent workflows. Read these when you want a point of view rather than a procedure.
Control loop
- 2026-08-13
Agent activity is not progress
Tool calls and busy run logs measure agent activity; progress begins when the workflow can prove that the target state moved closer and stayed there.
- agent-systems
- evals
- autonomic-swe
- 2026-08-13
Approval fatigue is a workflow design problem
When an agent asks for approval too often and without decision-ready context, the workflow trains people to click through instead of protecting the actions that deserve judgment.
- agent-systems
- trust-boundary
- runtime-infra
- 2026-08-13
Rollback belongs in the plan
An agent should know how a consequential action will be contained or reversed before it acts, because a rollback invented after failure is only another untested production change.
- autonomic-swe
- agent-systems
- runtime-infra
- trust-boundary
- 2026-08-13
Your agent is creating evidence debt
When an agent changes a system faster than you preserve why its work was accepted, each successful run leaves future operators with uncertainty they will have to repay.
- agent-systems
- evals
- trust-boundary
- 2026-08-12
A passing test is not always a good result
Tests tell an agent whether known checks passed; they do not automatically prove that the change helped users, respected every boundary, or is safe to keep.
- agent-systems
- evals
- trust-boundary
- 2026-08-11
Your agent should not be its own reviewer
Let an agent explain and improve its work, but use separate evidence and authority for the decision to accept, publish, merge, or deploy it.
- agent-systems
- evals
- trust-boundary
- 2026-08-06
Build the checks before you build the agent
Decide how you will judge the result before building the agent, or your examples and tests will slowly start to favor whatever the agent already does well.
- autonomic-swe
- agent-systems