nagisanzenin/effortmining

30 stars · Last commit 2026-07-10

Benchmark-calibrated per-subagent reasoning effort for Claude Code. Classify the subtask, dispatch the cheapest tier a blind grader still accepts. Pilot-proven: -64.7% tokens vs effort inheritance at equal quality.

README preview

![effortmining — right-effort dispatch for Claude Code subagents · 64.7% fewer tokens, same pass rate](assets/banner.png)

# effortmining

![version](https://img.shields.io/badge/version-0.5.3-blue)
![status](https://img.shields.io/badge/status-benchmark--proven-brightgreen)
![effort](https://img.shields.io/badge/effort-low%E2%80%A6max%20(5%20tiers)-green)
![telemetry](https://img.shields.io/badge/telemetry-100%25%20local-green)
![license](https://img.shields.io/badge/license-MIT-lightgrey)

> Spend the reasoning each subtask deserves — no more, no less.

Claude Code makes every helper agent think as hard as your whole session. Most helpers don't need to. **effortmining right-sizes reasoning effort per subagent and cuts output-token spend ~64.7% with no drop in quality** — measured, not vibes.

## Install

```
claude plugin marketplace add nagisanzenin/effortmining
claude plugin install effortmining@effortmining
```

View full repository on GitHub →