菜单
CI pipeline for auto-generating documentation and architecture diagrams. Use when setting up CI, generating diagrams, updating README automatically, or validating metadata. Triggers include CI, auto-generate, diagram, Mermaid, dependency graph, architecture.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Statistical Context Engineering algorithm implementation and verification. Use when verifying code against paper equations, checking cross-fitting logic, auditing for leakage, reviewing Algorithm 1, or debugging statistical computations. Triggers include cross-fitting, leakage, equations, paper, algorithm, phi, context operator, variance reduction, out-of-fold.
Productionalize R&D ML projects with the craft of an experienced engineer. Use when: "wrap up my ML project", "make this production ready", "productionalize", "I need to ship this model", "prepare for release", "add tests to my pipeline", "make my code look professional", "this needs to not look AI-generated", "review my ML code", "make this reproducible". Covers testing, reproducibility, documentation, code quality, and release prep.
ICML 2025 paper visuals, figures, and LaTeX tables. Use when generating figures, formatting tables, or preparing appendix materials. Triggers include figure, M1, M2, M3, publication, LaTeX, consolidated, ICML, paper.
SCE dataset management, loading, and anonymization. Use when loading datasets, anonymizing sources, managing parquet files, choosing storage strategy, or validating checksums. Triggers include dataset, parquet, download, anonymize, checksums, Hugging Face.
SCE Python library architecture and coding standards. Use when creating new modules, refactoring code, checking dependencies, defining module responsibilities, or writing module headers. Triggers include module, architecture, dependency, imports, engine.py, pipeline.py, meta.py, component.
| name | ci-autodocs |
| description | CI pipeline for auto-generating documentation and architecture diagrams. Use when setting up CI, generating diagrams, updating README automatically, or validating metadata. Triggers include CI, auto-generate, diagram, Mermaid, dependency graph, architecture. |
| Document | Purpose |
|---|---|
| references/METADATA_SPEC.md | Module header specification |
| docs/autogen/autodocs_design.md | System design document |
| docs/autogen/tasks/ | Implementation task files |
| Task | Status | Description |
|---|---|---|
| Task 01 | 🔲 TODO | Parser models and core implementation |
| Task 02 | 🔲 TODO | Metadata validator |
| Task 03 | 🔲 TODO | Mermaid diagram generator |
| Task 04 | 🔲 TODO | Markdown API docs generator |
| Task 05 | 🔲 TODO | README injection system |
| Task 06 | 🔲 TODO | CI/CD integration |
<!-- AUTODOC:NAME:START --> markers for injection.@component decorators.sce/
├── autodocs/ # Autodocs package (to implement)
│ ├── models.py # Data classes
│ ├── parser.py # Parsing logic
│ ├── validator.py # Validation rules
│ ├── generators/
│ │ ├── mermaid.py # Diagram generation
│ │ ├── markdown.py # API docs
│ │ └── injector.py # README injection
│ └── cli.py # CLI interface
scripts/ci/
├── generate_docs.py # CI script (wraps autodocs)
└── validate_metadata.py
docs/generated/ # Output directory