6verity
数学建模竞赛最终验证和验收阶段,支持 Typst 和 LaTeX 双引擎。仅在全自主模式下执行:检查论文结构完整性、文本质量、图表引用、数值一致性、编译和提交就绪状态。在辅助模式下不调用。
Skill by QianXiao233
Chalk Labs
Need help deploying 6verity?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code. Read the repository before you give any tool access to your data.
npx agentmarket install skill/qianxiao233-6verity
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: documents
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Python package
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/6verity
Tools it pre-approves 9
From allowed-tools in SKILL.md: the agent may use these without asking you each time.
Bash(*),Read,Write,Edit,Grep,Glob,Agent,WebSearch,WebFetch
What's inside
- Skill name
6verity- Description
- 数学建模竞赛最终验证和验收阶段,支持 Typst 和 LaTeX 双引擎。仅在全自主模式下执行:检查论文结构完整性、文本质量、图表引用、数值一致性、编译和提交就绪状态。在辅助模式下不调用。
- License
- MIT
- Allowed tools
- Bash(*),, Read,, Write,, Edit,, Grep,, Glob,, Agent,, WebSearch,, WebFetch
- Instructions
- 335 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
- Pre-approves some shell commands
Files (4)
- SKILL.md
- layout_check.py · scripts/
- writing_check.ps1 · scripts/
- writing_check.sh · scripts/
SKILL.md (first lines)
# 验证和验收(Typst / LaTeX)
> **⚠️ 模式约束:本 skill 仅在全自主模式下执行。**
> 辅助模式不生成论文,因此不需要论文验收环节。辅助模式产出的分析报告、代码和图表的正确性由各自的 skill 自行保证。
本 skill 是完整工作流的最后一关。它不重新建模、不生成新结果、不代替写作阶段重写论文;它负责发现硬错误、修复可直接修复的问题,并输出 `reports/VERIFY_REPORT.md`。
## 数学建模规范参考
如需领域判断,读取 `../_references/math_modeling_norms.md` 中的"论文验收与一致性"小节。该文件只是规范知识库,不是固定执行流程;具体目录、入口文件、结果文件和图表目录由当前项目结构决定。
## 阶段边界
- 本阶段负责:结构验收、文本质量门禁、**排版硬规则门禁(无目录页、表上图下)**、图表引用检查、结果一致性检查、Typst/LaTeX 编译检查、PDF 视觉检查、提交清单。
- 本阶段不负责:重新设计模型、重新跑大规模实验、重新组织整篇论文。
- 发现硬错误时,优先做小范围修复;如果需要回到前序阶段,写入 `reports/VERIFY_REPORT.md` 并标记为未通过。
## 输入
由模型先根据当前工作区判断项目布局,再把实际路径传给检查脚本。常见输入包括但不限于:
1. 论文入口文件:`main.typ`(Typst)或 `main.tex`(LaTeX)。
2. 正文章节目录或若干正文文件(`.typ` 或 `.tex`)。
3. 参考文献文件(`references.typ` 或 `references.tex`)。
4. 前序阶段的分析、建模、结果、图示报告。
5. 图表目录
6. 可复现代码目录。
7. 编译后的 PDF,或可由入口文件编译得到的输出 PDF。
不要假设论文目录一定叫 `paper/`,也不要假设结果文件一定在项目根。若项目使用不同命名,按实际结构传参并在 `reports/VERIFY_REPORT.md` 中说明。
## 工作流程
### Step 1: 运行文本质量门禁
优先运行本 skill 的脚本。脚本按入口文件扩展名自动选择检查逻辑(`.typ` → Typst 检查,`.tex` → LaTeX 检查):
**Linux/macOS (bash)**:
```bash
set -o pipefail
mkdir -p _tmp
SCRIPT_PATH="<按当前 skill 实际位置确定>/scripts/writing_check.sh"
bash "$SCRIPT_PATH" \
--paper-dir "$PAPER_DIR" \
--main "$MAIN_FILE" \
--sections-dir "$SECTIONS_DIR" \
--references "$REFERENCES_FILE" \
--figures-dir "$FIGURES_DIR" \
--results-file "$RESULTS_FILE" \
--problem-analysis "$PROBLEM_ANALYSIS_FILE" \
--all-results "$ALL_RESULTS_FILE" \
| tee _tmp/writing_check.log
```
**Windows (PowerShell)**:
```powershell
$null = New-Item -ItemType Directory -Force -Path _tmp
$scriptPath = "<按当前 skill 实际位置确定>/scripts/writing_check.ps1"
& $scriptPath `
-PaperDir $env:PAPER_DIR `
-Main $env:MAIN_FILE `
-SectionsDir $env:SECTIONS_DIR `
-References $env:REFERENCES_FILE `
-FiguresDir $env:FIGURES_DIR `
-ResultsFile $env:RESULTS_FILE `
-ProblemAnalysis $env:PROBLEM_ANALYSIS_FILE `
-AllResults $env:ALL_RESULTS_FILE `
*>&1 | Tee-Object -FilePath _tmp/writing_check.log
```
如果本 skill 被复制到其他目录,使用实际脚本路径。可以先运行 `bash <script> --help`(Linux)或 `powershell -File <script> -Help`(Windows)查看参数。不要把脚本路径、论文目录或文件名写死在验收逻辑中。
脚本只扫描文本,不生成论文,也不编译 PDF。它的 `FAIL` 属于硬错误,必须修复后重跑。
### Step 1.2: 排版硬规则门禁(无目录页 / 表上图下)
运行本 skill 的版面检查脚本,程序化核验两条竞赛论文硬规则:
**Linux/macOS (bash)**:
```bash
python3 "<按当前 skill 实际位置确定>/scripts/layout_check.py" \
--paper-dir "$PAPER_DIR" --main "$MAIN_FILE" \
--sections-dir "$SECTIONS_DIR" --pdf "$OUTPUT_PDF"
```
**Windows (PowerShell)**:
```powershell
python "<按当前 skill 实际位置确定>\scripts\layout_check.py" `
--paper-dir $env:PAPER_DIR --main $env:MAIN_FILE `
--sections-dir $env:SECTIONS_DIR --pdf $env:OUTPUT_PDF
```
检查内容:
- **目录页**:Typst 的 `outline(`、LaTeX 的 `\tableofcontents` / `\listoffigures` / `\listoftables`,以及任何以"目录 / Contents"为标题的页面 → `FAIL`。
- **表上图下**:LaTeX 源码中 `table` 环境的 `\caption` 是否在 `\begin{tabular}` 之前、`figure` 环境的 `\caption` 是否在 `\includegraphics` 之后;Typst 主文件是否含表题置顶规则(`set figure.caption(position: top)`);若提供编译后的 PDF,还会按**页面坐标**核验真实版面——表题落在表格下方、或图题落在图形上方 → `FAIL`。
脚本只读文本与 PDF,不修改任何文件;`pdfplumber` 未安装时 PDF 核验降级为 `WARN`。**退出码非 0 表示存在硬错误,必须修复后重跑**:属于模板问题就修模板,属于正文写法问题就改正文,不得只在报告里记录。
### Step 1.5: BZD 分板块内容自查(内容审查,与 Step 2–6 的技术验收互补)
文本质量门禁通过后,按板块调用 BZD 自查技能,把各技能输出的问题清单**合并去重**后写入 `reports/VERIFY_REPORT.md` 的"内容自查"一节:
| 自查对象 | 技能 | 关注点 |
| --- | --- | --- |
| 摘要 / 题目 / 关键词 | `bzd-abstract-checker` | 摘要独立性、信息量、与正文一致 |
| 问题重述 | `bzd-problem-restatement` | 任务遗漏、条件失真、章节越界 |
| 问题分析 | `bzd-problem-analysis-checker` | 任务映射、模型选择依据、跨问联动 |
| 模型假设 | `bzd-model-assumption-checker` | 逐条合理性、遗漏假设、验证要求 |
| 符号说明 | `bzd-symbol-notation-checker` | 遗漏、冲突、单位、上下标、首次定义 |
| 模型建立与求解 | `bzd-model-solution-checker` | 核心正文、复现性、红线问题、灵敏度 |
| 引用 / 参考文献 / 附录 | `bzd-reference-appendix-checker` | P0–P3 风险、引用一致性、附录完整性 |
终稿另跑两项(模式 A 必做,模式 B 按需):
1. **AI 痕迹审计**:`bzd-paper-aigc-auditor`——定位语言模板化、算法堆砌、模型拼装与伪改进风险,输出风险区间与逐板块证据。风险偏高时按证据改写(不改技术事实)后复审。
2. **严格格式总检**:`bzd-paper-format-checker`——页面结构、排版、篇幅、标题、图表、公式、匿名性与文件卫生,输出格式规范分与格式质量系数。非终稿不必频繁运行,以免消耗较多 token。
**分工不变**:本 skill 负责**提交前技术验收**(结构 / 编译 / 引用存在性 / 数值一致 / PDF 视觉),BZD 自查负责**内容质量**(写法、逻辑、模型可信度)。两者结论冲突时以证据更充分者为准,并在报告中记录冲突原因。From QianXiao233/mathmodel-assistant, MIT.
More from QianXiao233/mathmodel-assistant
- 0.563score
bzd-modeling-ideas
QianXiao233
Generate a coherent, whole-paper mathematical modeling solution framework from a complete contest problem.
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-problem-analysis-checker
QianXiao233
Review a Chinese mathematical-modeling problem-analysis section against the original problem, checking task classification, data reasoning, cross-problem…
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-paper-format-checker
QianXiao233
Review a complete mathematical-modeling competition paper with an atomic checklist covering abstract readability, page structure, typography, headings…
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-modeling-workflow
QianXiao233
Orchestrate the BZD mathematical-modeling Skills across problem reading, idea generation, model selection, paper drafting, section checks, final format/AIGC…
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-abstract-checker
QianXiao233
Review a user-provided Chinese mathematical-modeling abstract without requiring the problem statement or full paper, identify concrete writing defects, and…
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-ai-usage-disclosure
QianXiao233
Generate or review mathematical-modeling competition AI tool usage statements and the anonymous AI工具使用详情.pdf.
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-model-assumption-checker
QianXiao233
Review a mathematical-modeling paper's model-assumption section against the complete problem statement.
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-model-dictionary
QianXiao233
Query the bundled BZD mathematical-modeling dictionary for a user-selected model, explain its assumptions, inputs, outputs, limitations and validation, then…
SkillBFreshness: B★ 2Claude Code +9 - 0.559score
bzd-paper-aigc-auditor
QianXiao233
对数学建模竞赛论文进行两层AI痕迹审计——第一层9维检测(语言层面60%:连接词/排比/拔高词/被动句/段落规律/文本复杂度;事实层面40%:引文验证/数值一致性/术语一致性),第二层模型合理性深度审查。严格判分+一票否决+问题标红。输出分层HTML报告。
SkillBFreshness: B★ 2Claude Code +9 - 0.559score
bzd-problem-restatement
QianXiao233
Generate a compliant Chinese mathematical-modeling problem-restatement chapter from a complete problem statement, or review a user's restatement against the…
SkillBFreshness: B★ 2Claude Code +9 - 0.559score
bzd-cumcm-school-awards
QianXiao233
Query 2021-2025 CUMCM school awards and 2026 forecasts, or assess a student's preparation distance from provincial and national awards using school history…
SkillBFreshness: B★ 2Claude Code +9 - 0.559score
bzd-model-solution-checker
QianXiao233
Review the model establishment, numerical solution, result analysis, model validation, and sensitivity or robustness sections of a mathematical-modeling paper…
SkillBFreshness: B★ 2Claude Code +9
Signals
- Stars
- 2
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Sep 10, 2026 (1 month ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from QianXiao233
- 0.563score
bzd-modeling-ideas
QianXiao233
Generate a coherent, whole-paper mathematical modeling solution framework from a complete contest problem.
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-model-dictionary
QianXiao233
Query the bundled BZD mathematical-modeling dictionary for a user-selected model, explain its assumptions, inputs, outputs, limitations and validation, then…
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-model-assumption-checker
QianXiao233
Review a mathematical-modeling paper's model-assumption section against the complete problem statement.
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-abstract-checker
QianXiao233
Review a user-provided Chinese mathematical-modeling abstract without requiring the problem statement or full paper, identify concrete writing defects, and…
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-ai-usage-disclosure
QianXiao233
Generate or review mathematical-modeling competition AI tool usage statements and the anonymous AI工具使用详情.pdf.
SkillBFreshness: B★ 2Claude Code +9 - 0.563score
bzd-modeling-workflow
QianXiao233
Orchestrate the BZD mathematical-modeling Skills across problem reading, idea generation, model selection, paper drafting, section checks, final format/AIGC…
SkillBFreshness: B★ 2Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get 6verity running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
6verity skill. Agent Mart. https://agentmart.dev/skills/qianxiao233-6verity (last updated 2026-10-24)