---
id: manual-agentic-in-inferoa
name: "agentic-in/inferoa"
nameZh: "agentic-in/inferoa"
sourceCategory: "Loop Engineering Skills"
sourceCategoryZh: "Loop Engineering Skills"
url: "https://github.com/agentic-in/inferoa"
sourceUrl: "https://github.com/search?q=loops+engineering&type=repositories"
tags: ["Loop engineering","Agent harness","Token optimization","Inference"]
tagsZh: ["Loop engineering","Agent harness","Token optimization","Inference"]
categorySlugs: ["dev","meta-tools","data-analysis","ai-machine-learning"]
---

# agentic-in/inferoa

中文名称：agentic-in/inferoa

## Description

Inferoa is an inference-native tokenmaxxing agent harness for loop engineering, with slash-command loops, verification evidence, memory, context control, and routing visibility.

Inferoa 是面向 loop engineering 的 inference-native tokenmaxxing agent harness，支持 slash-command 闭环、验证证据、记忆、上下文控制和路由可见性。

## Why It Was Selected

This repository was selected because it directly frames loop engineering as an inference workload. It connects long-horizon agent loops with context windows, prefix-cache reuse, model routing, token pressure, and serving decisions.

GitHub API snapshot: non-fork repository, TypeScript, 133 stars, updated 2026-06-16 UTC.

## Use This Skill

- Open or install the skill from: https://github.com/agentic-in/inferoa
- Review the GitHub search source: https://github.com/search?q=loops+engineering&type=repositories
- Categories: dev, meta-tools, data-analysis, ai-machine-learning
- Tags: Loop engineering, Agent harness, Token optimization, Inference
- 中文标签：Loop engineering、Agent harness、Token optimization、Inference

## README SEO Report

### Detailed Overview

Inferoa treats loop engineering as more than agent orchestration. Its README argues that every loop is also an inference workload: turns accumulate, prompt prefixes drift, cache reuse can collapse, stale evidence fills context, and model routing becomes harder. The project focuses on keeping loops durable while making token pressure, context control, and serving decisions visible.

It is relevant for loop engineering teams that care about cost, context-window discipline, prefix-cache preservation, model routing, memory, and completion evidence.

### Installation

The README lists a global npm installation path:

```bash
npm install -g inferoa@dev
```

### Usage

Inferoa can be opened as a TUI or run in print mode. The README highlights `/loop`, `/plan`, and `/tokenmaxxing` as core slash commands.

```bash
inferoa setup
inferoa
inferoa "Inspect this repository and list the test entrypoints."
```

```text
/loop Improve this repository and prove it with tests.
/plan
/tokenmaxxing
```

### Capabilities

- Recursive long-horizon loop mode.
- Plans, tests, tool results, research metrics, and completion evidence.
- Memory and context control.
- Prefix-cache discipline and token pressure visibility.
- Model serving and routing awareness.

### Source Archive

- Archive source type: generated-local-file
- Archived docs: None

## 3W FAQ

### What is this skill?

Inferoa is an agent harness for running loop engineering workflows with inference-aware context and token controls.

### Who is this skill for?

This skill is for teams working on development, meta-tools, data-analysis, and AI machine learning workflows where agent loops can become expensive or context-heavy.

### When should you use it?

Use it when loop engineering needs explicit memory, verification evidence, routing visibility, and token/cost pressure management.

## 中文 3W FAQ

### What：这个 skill 是什么？

Inferoa 是一个 agent harness，把 loop engineering 和推理成本、上下文窗口、prefix cache、模型路由一起纳入控制。

### Who：适合谁使用？

适合处理 development、meta-tools、data-analysis、AI machine learning 工作流的团队、开发者或 AI Agent。

### When：什么时候使用？

当闭环任务很长、上下文压力大、成本敏感，或需要保留验证证据和模型路由可见性时使用。