Tag
A long-form engineering article discussing harness design for vertical-domain Agents: how to integrate project knowledge, workspaces, identity authentication, and long-task recovery on top of a Deep Agents foundation, and how to iterate on adding or removing capabilities based on evaluation results.
CLAP proposes a closed-loop method for domain agent post-training that converts noisy business data into structured SFT and preference samples, integrates reward/KL diagnosis, offline gates, and application-chain replay to decide adapter release. Experiments on five manufacturing batches show modest average gains and highlight that regression and high KL risks require an integrated data-training-evaluation-release loop rather than relying on a single score.