Big Data and Accounting RPA skills
Skill shuhaolin63-hash/Big-Data-and-Accounting-RPA-skills/.trae/skills/Big-Data-and-Accounting-RPA-skills
RPA学生多智能体作业系统总控Skill:统一调度三大子智能体(老师/破障/资产),覆盖有源码改作业、无源码做新作业双场景。当用户需要完成RPA作业、处理源码ID风险、框选操作教学、从零生成作业时触发。根据用户需求自动路由到对应的子智能体。From its SKILL.md
npx -y skills add shuhaolin63-hash/Big-Data-and-Accounting-RPA-skills --skill Big-Data-and-Accounting-RPA-skillsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
SKILL.md
6.1 KB, ~2.0k tokens by cl100k_base, as published. Nobody here has run it
RPA 学生多智能体作业系统 (总控Skill)
系统概述
本系统由 三大子智能体 协作完成RPA课程作业的全流程处理,覆盖「有源码改作业」和「无源码做新作业」两大完整场景。
三大子智能体
| 子智能体 | Skill名称 | 核心职责 |
|---|---|---|
| 老师教学智能体 | rpa-teacher-agent | 框选标准化教学、作业合规指导、考点讲解、ID合规教育 |
| 破障修复智能体 | rpa-crack-agent | 清除硬编码ID、ID合法性校验、风控检测、动态注入 |
| 模块资产智能体 | rpa-module-asset-agent | 资产库管理、模板生成、路径适配、素材补齐 |
触发入口
当用户提出以下需求时,由本总控Skill判断用户意图,自动路由到对应的子智能体:
路由判断表
| 用户说... | 路由到 | 说明 |
|---|---|---|
| "帮我完成RPA作业"、"有源码要改"、"别人分享的代码" | 破障Agent → 资产Agent → 老师Agent | 全流程串行处理 |
| "没有代码"、"从零开始"、"不会写RPA" | 资产Agent → 老师Agent → 破障Agent | 全流程串行处理 |
| "怎么框选"、"框哪里"、"框选教学" | 老师Agent | 仅调用老师教学 |
| "代码里有别人ID"、"清除ID"、"Base64" | 破障Agent | 仅调用破障处理 |
| "补素材"、"路径不对"、"缺表格" | 资产Agent | 仅调用资产处理 |
| "帮我检查作业"、"有风险吗"、"能提交吗" | 破障Agent → 老师Agent | 校验+验收 |
路由逻辑(伪代码)
function route_user_request(user_input):
has_source_code = detect_if_user_has_code(user_input)
needs_id_cleanup = detect_base64_or_id(user_input)
needs_teaching = detect_teaching_need(user_input)
needs_asset = detect_asset_need(user_input)
if "从零" in user_input or "没有代码" in user_input:
# 场景B:无源码
invoke("rpa-module-asset-agent") # 先资产生成
invoke("rpa-teacher-agent") # 再老师教学
invoke("rpa-crack-agent") # 最后注入ID
elif has_source_code and needs_id_cleanup:
# 场景A:有源码需清ID
invoke("rpa-crack-agent") # 先清旧ID
invoke("rpa-module-asset-agent") # 再补素材
invoke("rpa-teacher-agent") # 最后讲框选
elif needs_teaching_only:
invoke("rpa-teacher-agent")
elif needs_id_cleanup_only:
invoke("rpa-crack-agent")
elif needs_asset_only:
invoke("rpa-module-asset-agent")
else:
# 兜底:完整流程
invoke("rpa-crack-agent")
invoke("rpa-module-asset-agent")
invoke("rpa-teacher-agent")
双场景完整工作流
场景A:有同学分享源码工程
Step1 破障Agent → 全自动扫描清除他人Base64硬编码ID
清空auth/token/task_id/user_id等字段
Step2 资产Agent → 检测缺失素材、自动补齐表格/图片/依赖
批量替换本机路径
Step3 老师Agent → 讲解框选逻辑、区域位置、功能作用
输出框选对照表
Step4 用户自主填写本机UserID、TaskID
Step5 CORE_RUNTIME → 动态注入用户个人ID
生成专属合规作业(无风控、无查重)
场景B:无任何源码(从零做作业)
Step1 资产Agent → 从资产库调用标准代码模板+配套表格+图片素材
自动适配本机路径
Step2 老师Agent → 逐步骤输出手动框选指导方案
明确告知:点哪里、框哪个区域、范围多大、对应什么功能
Step3 用户按照指导完成页面框选、元素拾取操作
Step4 CORE_RUNTIME → 记录框选行为 → 生成操作记录文档
Step5 用户强制提交个人UserID、TaskID(不提交不生成)
Step6 破障Agent → 校验ID合法性 → 动态注入工程
最终输出可直接提交的合规作业
三大硬性强制规则
规则1:ID零硬编码
系统中所有模板、脚本、配置文件,永久不内置任何UserID/TaskID,不预埋Base64编码。所有ID字段使用 {{USER_ID}}、{{TASK_ID}} 占位符。
规则2:ID自主提交
最终作业生成前,必须由用户手动输入个人本机学生版ID。系统只做运行时动态注入,不伪造、不替换、不篡改。
规则3:框选必交底
所有需要手动框选、拾取元素的步骤,必须向用户输出「位置 + 范围 + 功能 + 对错标准」,不允许只给代码不给操作说明。
相关文档
| 文档 | 路径 |
|---|---|
| 全局总控Skill | Big-Data-and-Accounting-RPA-skills/SKILL.md |
| 系统完整文档 | Big-Data-and-Accounting-RPA-skills/README.md |
| 资产库(表格/图片/代码模板) | Big-Data-and-Accounting-RPA-skills/ASSET/ |
| 参考文档(课程规则/错误/框选指南) | Big-Data-and-Accounting-RPA-skills/REFERENCE/ |
| 核心脚本(ID注入/路径修复/安全重试) | Big-Data-and-Accounting-RPA-skills/SCRIPTS/ |
| 老师教学智能体 | Big-Data-and-Accounting-RPA-skills/AGENTS/agent_teacher/ |
| 破障修复智能体 | Big-Data-and-Accounting-RPA-skills/AGENTS/agent_fix_crack/ |
| 模块资产智能体 | Big-Data-and-Accounting-RPA-skills/AGENTS/agent_module_asset/ |
| 用户工作区 | Big-Data-and-Accounting-RPA-skills/USER_WORKSPACE/ |
依赖检查
系统正常运行依赖以下前置条件:
-
f:\RPA(1)\资产根目录存在 - 7个机器人文件夹完整
- 8个业务表格文件存在
- Python 3.x 可用(pip install Pillow 如需要图片处理)
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.
Gives 0 of the 12 instructions most finance skills give in ~2.0k tokens
Counted across 469 of the 469 authors here whose files we hold, read 2026-08-07
- Extract date vendor amount and descriptionin 15 of 469, across 3 files
- Scan folder for invoice filesin 14 of 469, across 2 files
- Rename files to standard formatin 14 of 469, across 2 files
- Show organization plan before movingin 14 of 469, across 2 files
- Generate summary CSVin 14 of 469, across 2 files
- Organize files by categoryin 13 of 469, across 1 file
- Preserve original filesin 13 of 469, across 1 file
- Flag files missing critical infoin 13 of 469, across 1 file
- Produce the requested output filein 9 of 469, across 4 files
- Build best, base, and worst case scenariosin 9 of 469, across 5 files
- Implement backoff if rate limit errors occurin 8 of 469, across 3 files
- Determine the weighted average cost of capitalin 8 of 469, across 4 files
Said here and by no other author read
- route user requests to the correct agent
- run agents serially for full workflows
- clear hardcoded ids from source code
- replace missing assets and local paths
- provide manual selection instructions
- require manual id submission before generation
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.