// Ingest Wikipedia HTML, normalize links/media, and archive to knowledge base.
| name | wiki-ingestion |
| description | Ingest Wikipedia HTML, normalize links/media, and archive to knowledge base. |
Continuous improvement: see
continuous_improvement.mdin this folder for a running log of lessons learned and guidance on evolving the wiki-ingestion skill.
Use this skill when importing Wikipedia articles or converting HTML content into Markdown notes.
Converts Wikipedia HTML (or similar web content) into well-formed Markdown with:
%20 for spaces)archives/Wikimedia Commons/Command: uv run -m templates.new_wiki_page
Script prompts for Wikipedia article name (e.g., "Fourier Transform")
Generates YAML frontmatter template:
---
aliases: [Alternative name]
tags: [flashcard/active, language/in/English]
---
Adds Wikipedia link comment: <!-- Source: https://en.wikipedia.org/wiki/Article_Name -->
Copies template to clipboard
Action: Paste into new file general/Article Name.md (or special/ if specialized content)
Command: uv run -m convert_wiki
archives/Wikimedia Commons/ using convert_wiki.filename_rename_map.jsonc for filename renames%20 encoding (not %3A or other encodings)Flashcard creation is managed automatically by the repository’s build workflows; agents and authors are not expected to run any commands to produce flashcards. The generator scans for three kinds of markup:
{@{ hidden text }@} for cloze deletions (hide text within a paragraph),::@:: for two-sided question/answer pairs (line-only, two cards), and:@: for one-sided question/answer pairs (line-only, single card).The source must honour the single‑line restriction for the latter two
formats; use <br/> or <p> for any desired visual breaks. When you add
these markers, the build updates pytextgen regions behind the scenes. See
pytextgen skill for additional details.
aliases and tags in YAML frontmatterarchives/Wikimedia Commons/)archives/Wikimedia Commons/ with %20-encoded filenamesaliases and tags.markdownlint.json settings{@{ }@}, ::@::, or :@: for active recallconvert_wiki%20 encoding for spaces (not %3A or other encodings)# Ingest from clipboard
uv run -m convert_wiki
# Scaffold new wiki-sourced note
uv run -m templates.new_wiki_page
Content conventions, examples, and tooling for course notes under special/academia.
Repository-wide tooling including init wrapper, pack/publish utilities, and all helper scripts. Links to tools-special and tools-templates for subfolder details.
Regenerate programmatically-generated content blocks in knowledge base notes using pytextgen.
Help the user add spaced-repetition flashcards (cloze deletions or Q/A pairs) to existing Markdown notes across the repository. The skill encapsulates the multi‑step process the user follows in their finance lecture notes (e.g. FINA 3103) and elsewhere: read the prose, identify key terms, dates, formulas and logical assertions, and wrap them using flashcard markup (`{@{ }@}`, `::@::`, `:@:`). There are three supported forms: - **Cloze** (`{@{ }@}`) hides arbitrary text inside paragraphs. - **Two-sided QA** (`::@::`) on a single line, yielding two cards. - **One-sided QA** (`:@:`) on a single line, yielding a single card. For the QA formats remember the line-only rule; if visual separation is needed insert `<br/>`/`<p>` instead of newline characters. Representative examples later in this document illustrate all three types. Use existing flashcard files as style guides and adapt the output based on user feedback. The skill also suggests regeneration commands once flashcards are inserted.
Archive online content into archives/ with automatic index.md updates via pyarchivist tool.