Agent-based review
AssessTechniques
An LLM-based review step that evaluates code against architectural and quality constraints.
Why it's here
Placed in Assess: 1 article(s) of evidence from 1 source(s), led by research-stage coverage, with 1 in the last 30 days. Confidence 24%. Low accumulated evidence, so it defaults conservatively pending more signal.
Evidence (1)
- 5Hacker News·8/27/2026researchHarness Engineering for AI-Assisted Coding
The article introduces harness engineering, a workflow for keeping AI-generated code correct and consistent over time by combining context documents, deterministic checks, and agent-based review. It argues that functional tests alone are not enough, and that teams need verification points to enforce architectural decisions, conventions, and constraints.