My evolving delegation rules and Eval Harness for agentic build tools, to manage and cost-check + quality-check delegating work to sub-agents, including cheaper models, to save token costs while ensuring quality work. Includes paired-agent Eval tests. Uses Subagents in Claude code. Another approach: https://github.com/Asymptote-Labs/agent-beacon
hooks evaluations evaluation evaluation-metrics evaluation-framework delegating evals agentic-workflow agent-orchestration evaluation-harness
-
Updated
Nov 19, 2026 - Python