Topic
Designing the code, tools, permissions, validation and recovery machinery that surrounds an LLM so an agent can act on real systems.
A dev.to post credits Anthropic with running the same model and the same prompt under two harnesses, 20 minutes and $9 for a broken result against six hours and $200 for a working one. The hourly spend barely moved.
Reality
A dev.to piece on harness engineering splits agent degradation into five separable failure modes. The two it actually describes both turn on where a token sits in the assembled prompt, and a bigger window leaves both in place.
Reality
A dev.to post's case for harness engineering rests on one PM's half hour and one investigation where directed context cut a task from about 50k tokens to 15k, a figure the author himself calls an observation.
Reality
A dev.to engineer argues the shipping bottleneck has moved from prompt wording to the environment around the loop, and his own postmortems carry that case a good deal better than the 40% failure figure he opens with.
Reality
A dev.to write-up splits agent work into prompts, context and harness. The interesting part is the harness: tool execution, permissions, validation and recovery, all of it code you own.
Reality