adopt
Take a folder that already holds research — papers, notes, code, half-written drafts — into MAGI, without breaking what is already there.
Repair what the mechanical passes cannot: broken LaTeX from conversion, sprawling tags, and concepts that are secretly the same concept.
$ npx -y skills add Misaka16384/Wikify --skill tidy --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/tidyContext preview
The summary Claude sees to decide when to auto-load this skill.
Repair what the mechanical passes cannot: broken LaTeX from conversion, sprawling tags, and concepts that are secretly the same concept.
name: tidy description: "Repair what the mechanical passes cannot: broken LaTeX from conversion, sprawling tags, and concepts that are secretly the same concept." commands: tidy: "Fix conversion damage, tag sprawl and duplicate concepts." origin: magi
`magi math check` reports errors, tags have sprawled into near-duplicates, or two concept cards are about the same thing.
**Maths.** `magi math format` first — it fixes mechanically what it can, free. Then `magi math check --json` for what is left. Work one file at a time: fix the *first* entry, re-check, and watch the rest of that file's errors vanish — one unclosed `$$` swallows everything after it and reports as many errors. `likely-macro` is usually a false positive; check the source with `magi ingest crop <pdf>` before touching it.
**Tags.** `magi tags extract .`, read the counts, and write the mapping that collapses synonyms, acronyms and plurals into one tag. Show it to the human before `magi tags apply . <tag_map> <alias_map>` — it rewrites every card. Two tags that are aliases of each other usually mean two cards about one concept.
**Concepts.** `magi link --dedup-only` proposes merges. Confirm each is really one concept, then `magi wiki refactor-concept --project-dir . --old <a> --new <b>`. A sub-concept merges into its parent; read it back with `magi wiki context --name <c>` first.
Close with `magi lint --fix` and `magi index`.
by hand.
only route that fixes the links too.
at the end and report what is still broken as **partial**, not as done.
MAGI 是一个 agent-native 的科研工作环境:人是驾驶员,LLM agent 是机体,确定性的 magi CLI 是拘束具——三者同步率越高,科研越快。它把学术论文(PDF/LaTeX)摄入、编译为 Obsidian 兼容的概念卡片项目,并用三核架构管理完整的科研状态: 进入任意项目先跑 magi sync——它输出同步率、三核状态和逐条可执行的修复提示。magi radar 是文献雷达:定时发现相关新论文,并侦察"知识上应引用我方论文却未引用"的候选。同一份 skills 通吃 Claude
Repo: Misaka16384/Wikify
Take a folder that already holds research — papers, notes, code, half-written drafts — into MAGI, without breaking what is already there.
Answer a question from this project's own knowledge — retrieved, read and cited, never from memory.
Compile raw sources into dense, interlinked reference and concept cards, and mine the concepts a compiled card left implicit.
Turn whatever the human has — a PDF, a link, a DOI, a screenshot of a citation — into raw/ sources this project can compile.
Entry point for a MAGI research project: run magi next, do what it says, and call the skill it names.
Triage a literature-radar digest: judge each candidate against what this project is actually doing, and queue only what earns it.