Xpersona
agentCLAWHUBUnverified

日语朗读作业批改模板

日语朗读作业批改（教师模式）——面向 A1/A2/B1 (JLPT N5/N4/N3) 日语学习者的朗读音频作业批改工具。ASR+声学粗筛自动检测漏读/语速/停顿；促音/长音/清浊仅输出候选。支持批量批改与归档，满分10分。 Skill: 日语朗读作业批改模板 Owner: bianmaxingkong Summary: 日语朗读作业批改（教师模式）——面向 A1/A2/B1 (JLPT N5/N4/N3) 日语学习者的朗读音频作业批改工具。ASR+声学粗筛自动检测漏读/语速/停顿；促音/长音/清浊仅输出候选。支持批量批改与归档，满分10分。 Tags: latest:2.2.3 Version history: v2.2.3 | 2026-08-18T11:26:20.526Z | user 相较此前版本的主要改进： - ASR 引擎升级为 faster-whisper（large-v3/medium 自适应） - 评分模型从扣分制重构为四维加权制（完整性/准确性/流畅度/发音质量） - 发音判定改为"可疑候选定位+教师复核"，不自动判错 - 新增置信度分级（高/中/低），分级驱动教师复核投入 - 最低分从6.0调整为完整0-10分区间 v2.2.

OpenClaw
View on ClawHubHomepage

Rank

62

Safety

84

Downloads

1.6k

Updated

Oct 10, 2026

Version

2.2.3

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 1.6K downloads reported by the source. Last updated 10/10/2026.

Avoid when

  • Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor
Clawhubvendor · observed Oct 10, 2026
Protocol compatibility
OpenClawcompatibility · observed Oct 10, 2026
Adoption signal
1.6K downloadsadoption · observed Oct 10, 2026
Latest release
2.2.3release · observed Aug 18, 2026
Handshake status
UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s17f7de24jwye68xxaykgnjqwn8616wn:japanese-reading-grader
  1. Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.
  2. Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-bianmaxingkong-japanese-reading-grader/snapshot"

Documentation

CLAWHUB

77,147 characters of source documentation, loaded on request.

Extracted files

3 files captured from the source.

SKILL.md

--- 
name: japanese-reading-grader
display_name: 日语朗读作业批改
description: A1/A2/B1(N5/N4/N3)日语朗读音频作业批改，faster-whisper ASR+声学双通道四维AI评测；支持单人/批量批改、纠错反馈、成绩归档；AI预批改，教师人工终审，满分10分。

## 适用场景

学生提交音频文件（MP3/MP4/M4A/WAV）朗读日语文本，教师需要：下载 → 转写 → 评分 → 纠错 → 录入成绩 → 生成班级汇总。

## 作业基础信息

| 项目 | 内容 |
| :--- | :--- |
| **适用对象** | A1/A2/B1 日语学习者（操作标签参考 JLPT N5/N4/N3） |
| **作业形式** | 朗读音频 |
| **总分** | 10 分 |

---

## 工作流程概览

> Step0 标准文本准备 → Step1 音频预处理 → Step1.5 音频质量检测 → Step2 ASR+声学双通道分析 → Step2.5 ASR可靠性检测（置信度分级）→ Step3 四维 AI 评分 → Step4 教师复核 → Step5 成绩录入与反馈生成

---

## Step 0：标准文本准备

1.  **标准答案存储路径**：`./answer/{COURSE_ID}_{ASSIGNMENT_ID}_answer.txt`
2.  **文本格式要求**：
    - 每行 1 个完整句子，无多余空格与换行
    - 重点词汇标注假名
    - 不含多余标点，便于音频转写对齐
3.  **版本校验**：批改前确认标准答案与作业要求一致
4.  **标准答案假名预处理**：
    - 默认本地 `pykakasi`：离线生成假名序列用于辅助分析，确定性强、不上传任何学生数据
    - 生成对照文件：`./answer/{COURSE_ID}_{ASSIGNMENT_ID}_kana_ref.txt`，结构：`原文句子｜规范化假名序列`，用于逐假名对齐比对。**该文件是准确性判定/同音词豁免/清浊·长短音核对的必需输入。**
    - 如未安装 `pykakasi`（降级模式）：**不阻断评分，但强制将该生归入「中置信」分流教师复核**，并在批改日志记录降级原因

---

## Step 1：音频预处理

- 批量转为 wav 格式，统一采样率
- 存放路径：`./audio/{student_id}/`

---

## Step 1.5：音频质量检测

检测项：
- 有效语音比例
- 音量范围
- 背景噪声
- 是否存在多人声音

**异常处理**：音频质量不足 → 不评分 → 请求重新上传。

---

## Step 2：ASR + 声学双通道分析

### 转写工具规格（强制）

- **唯一转写工具**：`faster-whisper`
- **仅允许模型**：`large-v3 fp16` / `medium`；禁用 `small` / `base` / `tiny` 及全部其他规格
- **硬件调度**：有充足 GPU 用 `large-v3 fp16`；无 GPU 自动降级 `medium`
- **固定开启** `word_timestamps=True` 输出词级时间戳

系统同时处理两个分析通道（**双通道校验的基础**）：

| 通道 | 用途 |
| :--- | :--- |
| **A. ASR 文本识别通道** | 辅助判断朗读内容完整性、辅助定位可能存在的发音异常、生成文本对齐候选 |
| **B. 声学分析通道** | 输入原始 wav 音频，分析发音异常候选、音长变化、停顿、语速、节奏；用于发音质量评价及长音/促音/拨音/流畅度分析 |

---

## Step 2.5：ASR 可靠性检测

系统综合判断以下方面：
- 文本匹配程度
- 音频质量
- ASR 置信度
- 对齐情况

输出：**高 / 中 / 低置信度**（用于 Step 4 教师复核的投入分级）。

### 置信度分级判定标准

| 置信度 | 判定条件 |
| :--- | :--- |
| **高** | 文本匹配率 ≥90% 且 ASR 平均置信度 ≥0.8 且 音频信噪比 ≥20dB |
| **中** | 文本匹配率 ≥70% 且 <90%，ASR 平均置信度 ≥0.5 且 <0.8 |
| **低** | 文本匹配率 <70%，或 ASR 平均置信度 <0.5，或音频质量检测不通过 |

> 置信度‑AI评分衔接规则
> 1. **高、中置信度**：四维AI加权分数作为正式预评分，教师仅做异常修正；
> 2. **低置信度**：AI四维加权分数仅作为后台诊断参考，**不计入、不对外展示，最终成绩完全由教师人工给出**。

---

## Step 3：四维 AI 评分

### 评分原则

1.  不影响理解的偏差不扣分
2.  A1/A2 （N5/N4）阶段轻微发音偏差以反馈为主
3.  不确定错误不扣分

> 发音质量维度重要说明
> 发音质量权重占比15%参与AI加权运算，但**该维度输出仅为趋势预估值，不能直接等价于发音错误**；系统不会依靠该维度自动判定促音、长音、清浊错误；所有发音类错误必须经教师试听复核确认后方可扣分。
> 系统可自动化可靠指标仅为：语速、停顿时长、完整性（漏读/多读）、文本对齐。

### 评分前置过滤规则（评分前必须执行）

计算任何维度分数前，先筛除以下情况，**排除出扣分范围**：

1.  **ASR 误识别**——不进入评分计算
    - 人名、地名、汉字词、同音词差异
    - 系统标记为“ASR疑似误识别”的项
2.  **转写存疑**——不进入评分计算
    - 系统置信度 < 0.5 的识别结果
3.  **学习者自然偏差**——反馈但不扣分
    - A1/A2 阶段轻微音长偏差、拨音不到位
    - 语气词添加、促音/长音时长轻微不足

### 音频层评价指标

1.  **发音质量**：清音/浊音区别、拗音准确性、促音闭塞时间、长音持续时间、拨音鼻音实现
2.  **流畅度**：平均语速、停顿次数、非正常停顿长度、重复修正次数
3.  **节奏与韵律**：音节时长比例、长短音节比例、句末语调变化（B1以上参考，不作为 A1/A2 主要扣分依据）

### 能力层级说明（双轨分级）

| 能力层级 | 约对应 JLPT | 朗读评价重点 |
| :--- | :--- | :--- |
| **A1** | N5 | 准确朗读短句，掌握基础假名发音，长音/促音容错空间大 |
| **A2** | N4 | 连贯朗读

_meta.json

{
  "ownerId": "kn7cw3j55n4hxe9ze445qp81ks86150x",
  "slug": "japanese-reading-grader",
  "version": "2.2.3",
  "publishedAt": 1787052380526
}

skill-card.md

## Description:

A1/A2/B1(N5/N4/N3)日语朗读音频作业批改，faster-whisper ASR+声学双通道四维AI评测；支持单人/批量批改、纠错反馈、成绩归档；AI预批改，教师人工终审，满分10分。

This skill is ready for commercial/non-commercial use.

## Publisher:

[bianmaxingkong](https://clawhub.ai/user/bianmaxingkong)

### License/Terms of Use:

MIT-0

## Use Case:

Teachers use this skill to pre-grade Japanese reading audio assignments for A1/A2/B1 learners, generate concise feedback, and support batch grading records. The skill is designed as an AI-assisted workflow with teacher review before final grades are recorded.

### Deployment Geography for Use:

Global

## Known Risks and Mitigations:

Risk: Student audio, scores, reports, and progress files may contain sensitive education data.

Mitigation: Use the skill only in authorized teaching contexts, keep generated files in restricted course folders, minimize student identifiers, and follow school retention rules for deletion or archival.

Risk: ASR or acoustic analysis can misidentify pronunciation or reading errors, especially for low-confidence audio.

Mitigation: Use the skill's confidence tiers and require teacher review for low-confidence cases and pronunciation-sensitive deductions before recording final grades.

## Reference(s):

- [ClawHub skill page](https://clawhub.ai/bianmaxingkong/skills/japanese-reading-grader)

## Skill Output:

**Output Type(s):** [text, markdown, guidance, configuration]

**Output Format:** [Markdown or plain text feedback reports with scores, diagnostic notes, correction suggestions, and optional local report/progress file paths]

**Output Parameters:** [1D]

**Other Properties Related to Output:** [May produce per-student report text, batch progress records, failed-audio records, and teacher-review prompts for low-confidence or pronunciation-sensitive findings.]

## Skill Version(s):

2.2.3 (source: release evidence)

## Ethical Considerations:

Users should evaluate whether this skill is appropriate for their environment, review any generated or modified files before relying on them, and apply their organization's safety, security, and compliance requirements before deployment.

Related agents

Same protocol, similar job.

Github ReposUpdated 15h agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW
Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW
Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.

MCPOPENCLAW
Github ReposUpdated 7mo agoRank 70

CopilotKit

The Frontend for Agents & Generative UI. React + Angular

OPENCLAW

Machine-readable data

The same record, as JSON, for agents and crawlers.

{
  "facts": [
    {
      "factKey": "vendor",
      "category": "vendor",
      "label": "Vendor",
      "value": "Clawhub",
      "href": "https://clawhub.ai/bianmaxingkong/skills/japanese-reading-grader",
      "sourceUrl": "https://clawhub.ai/bianmaxingkong/skills/japanese-reading-grader",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-10T07:04:35.208Z",
      "isPublic": true
    },
    {
      "factKey": "protocols",
      "category": "compatibility",
      "label": "Protocol compatibility",
      "value": "OpenClaw",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-bianmaxingkong-japanese-reading-grader/contract",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-bianmaxingkong-japanese-reading-grader/contract",
      "sourceType": "contract",
      "confidence": "medium",
      "observedAt": "2026-10-10T07:04:35.208Z",
      "isPublic": true
    },
    {
      "factKey": "traction",
      "category": "adoption",
      "label": "Adoption signal",
      "value": "1.6K downloads",
      "href": "https://clawhub.ai/bianmaxingkong/japanese-reading-grader",
      "sourceUrl": "https://clawhub.ai/bianmaxingkong/japanese-reading-grader",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-10T07:04:35.208Z",
      "isPublic": true
    },
    {
      "factKey": "latest_release",
      "category": "release",
      "label": "Latest release",
      "value": "2.2.3",
      "href": "https://clawhub.ai/bianmaxingkong/japanese-reading-grader",
      "sourceUrl": "https://clawhub.ai/bianmaxingkong/japanese-reading-grader",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-08-18T11:26:20.526Z",
      "isPublic": true
    },
    {
      "factKey": "handshake_status",
      "category": "security",
      "label": "Handshake status",
      "value": "UNKNOWN",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-bianmaxingkong-japanese-reading-grader/trust",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-bianmaxingkong-japanese-reading-grader/trust",
      "sourceType": "trust",
      "confidence": "medium",
      "observedAt": null,
      "isPublic": true
    }
  ],
  "events": [
    {
      "eventType": "release",
      "title": "Release 2.2.3",
      "description": "相较此前版本的主要改进： - ASR 引擎升级为 faster-whisper（large-v3/medium 自适应） - 评分模型从扣分制重构为四维加权制（完整性/准确性/流畅度/发音质量） - 发音判定改为\"可疑候选定位+教师复核\"，不自动判错 - 新增置信度分级（高/中/低），分级驱动教师复核投入 - 最低分从6.0调整为完整0-10分区间",
      "href": "https://clawhub.ai/bianmaxingkong/japanese-reading-grader",
      "sourceUrl": "https://clawhub.ai/bianmaxingkong/japanese-reading-grader",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-08-18T11:26:20.526Z",
      "isPublic": true
    }
  ]
}

Record generated Oct 10, 2026.

Sponsored

Ads related to 日语朗读作业批改模板 and adjacent AI workflows.