← 返回 Skills 市场
天天老师助手
作者
pengdagang1975
· GitHub ↗
· v1.0.0
· MIT-0
326
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install tiantian-grader
功能描述
作业批改专家,支持多份作业连续批改,自动生成班级成绩统计表
安全使用建议
Before installing, confirm the following: (1) Why does the skill need TENCENT_SECRET_ID and TENCENT_SECRET_KEY? Ask the author which Tencent APIs are used and whether you can provide a scoped service account (minimal permissions) rather than full account keys. (2) The skill writes to /root/.openclaw/workspace and persists a MEMORY.md and uploaded student files — if you will process real student data, ensure you have consent and use a sandbox with limited access and encrypted storage. (3) The SKILL.md uses child_process.exec; this can enable command injection if inputs (filenames, URLs) are not sanitized — request code review or run it in an isolated container. (4) If you cannot review the full code, do not provide high-privilege credentials; instead create a dedicated, least-privilege Tencent account and rotate or revoke keys after testing. (5) Ask the developer for a minimal reproduction that shows explicit Tencent usage (network endpoints, SDK calls) and for documentation of data retention and deletion policies.
功能分析
Type: OpenClaw Skill
Name: tiantian-grader
Version: 1.0.0
The skill is a homework grading assistant designed to process student assignments, manage rosters, and generate statistical reports. It implements a structured workflow including OCR simulation, manual correction of recognized text, and CSV export of grades. All file operations are restricted to a dedicated workspace directory (/root/.openclaw/workspace/teacher_tiantian), and no evidence of data exfiltration, malicious command execution, or prompt injection was found in skill.md or the handler logic.
能力评估
Purpose & Capability
The skill claims to be a homework-grading assistant which can ingest uploads, grade, and produce reports — that matches many of the file- and report-related operations in SKILL.md. However, the skill declares required environment variables TENCENT_SECRET_ID and TENCENT_SECRET_KEY while the visible code mostly performs local file I/O and uses child_process.exec; there is no obvious, explicit Tencent API usage in the visible snippets to justify granting cloud credentials. Requiring cloud secrets should be explained (e.g., OCR or cloud storage) — otherwise the credential request looks disproportionate.
Instruction Scope
The SKILL.md contains embedded Node.js runtime logic that creates and writes into a workspace under /root/.openclaw/workspace/teacher_tiantian, reads a roster CSV, reads/writes a persistent memory file, saves uploaded files, and uses child_process.exec. These actions read and persist local data and execute shell commands. That scope goes beyond a simple in-memory grading routine: it persists data on disk, manipulates files in a root-owned path, and executes commands — all of which increase risk (data leakage, tampering, command injection) if inputs are not strictly validated.
Install Mechanism
There is no install spec (instruction-only), so no external binaries or archives are fetched during installation. That reduces supply-chain risk. However, the instructions create persistent directories and files under /root/.openclaw/workspace and will write/execute runtime code there, so there is still local persistence risk even without an installer.
Credentials
The skill requires TENCENT_SECRET_ID and TENCENT_SECRET_KEY. A single cloud credential can be reasonable if the skill uses Tencent OCR, storage, or AI services — but the SKILL.md does not clearly show Tencent API calls in the visible parts. Asking for full cloud credentials without explaining the exact APIs/scopes needed is disproportionate. Also the skill requests access to a web-enabled config and a specific model (deepseek-reasoner), expanding its reach.
Persistence & Privilege
always:false (normal). The skill requests session persistence (persist: true) and writes a MEMORY.md and workspace files, so it will leave lasting data on the host. It doesn't claim to change other skills or system-wide config, but persistent storage of uploaded student data and rosters increases sensitivity and should be considered when granting permissions.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install tiantian-grader - 安装完成后,直接呼叫该 Skill 的名称或使用
/tiantian-grader触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
- Initial release of 天天老师助手 (tiantian-grader) for efficient homework grading.
- Supports continuous grading of multiple assignments.
- Automatically generates class performance statistics.
- Includes comprehensive file and roster management.
- Offers user-invocable grading actions and detailed result storage.
元数据
常见问题
天天老师助手 是什么?
作业批改专家,支持多份作业连续批改,自动生成班级成绩统计表. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 326 次。
如何安装 天天老师助手?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install tiantian-grader」即可一键安装,无需额外配置。
天天老师助手 是免费的吗?
是的,天天老师助手 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
天天老师助手 支持哪些平台?
天天老师助手 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 天天老师助手?
由 pengdagang1975(@pengdagang1975)开发并维护,当前版本 v1.0.0。
推荐 Skills