Multi-agent harness for Claude Code: autonomous Planner-Generator-Evaluator sprints with parallel Agent Teams, harness-engineering framework, and 9 canonical defect diagnoses.
-
Updated
May 19, 2026 - PowerShell
Multi-agent harness for Claude Code: autonomous Planner-Generator-Evaluator sprints with parallel Agent Teams, harness-engineering framework, and 9 canonical defect diagnoses.
GAN-inspired adversarial development loop — Planner directs Generator + Evaluator in fresh cmux panes, gated by hard pass/fail per deliverable. Adapted from Anthropic's harness-design guidance for long-running agentic apps.
Add a description, image, and links to the planner-generator-evaluator topic page so that developers can more easily learn about it.
To associate your repository with the planner-generator-evaluator topic, visit your repo's landing page and select "manage topics."