adopt
Take a folder that already holds research — papers, notes, code, half-written drafts — into MAGI, without breaking what is…
Write the one thing a person reads after an unattended run — its report — and the lecture notes that prepare them for it, in words they already have.
Claude Code
$ npx -y skills add misaka16384/magi --skill brief --agent claude-code
Codex
$ npx skills add misaka16384/magi --skill brief -a codex -y
Cursor
$ npx skills add misaka16384/magi --skill brief -a cursor -y
Gemini CLI
$ npx skills add misaka16384/magi --skill brief -a gemini-cli -y
GitHub Copilot
$ npx skills add misaka16384/magi --skill brief -a github-copilot -y
OpenCode
$ npx skills add misaka16384/magi --skill brief -a opencode -y
Windsurf
$ npx skills add misaka16384/magi --skill brief -a windsurf -y
name: brief description: "Write the one thing a person reads after an unattended run — its report — and the lecture notes that prepare them for it, in words they already have." commands: brief: "Write a run's report, or lecture notes on a method it used." origin: magi
A run has enough to write up, or only its report is left; or `magi next` lists lecture notes the person asked for.
1. Report: `magi run outline <run> --write` (add `--empty` when nothing big was found: one page). The claims and whether each stands, the chain, the steps and the glossary are in it already; every `<!-- … -->` is yours to replace. 2. Write for one sitting, start to end: motivation → what is and is not shown → what is new → the picture and the core insight (brute force said honestly) → the derivation in chain order, keeping what was decided or is not obvious → decisions → failures worth knowing (at most 5, three lines each) → is it still promising. 3. `magi review drafts/runs/<run>.md` has another agent read it whole. Fix what it names, then `magi run report <run> drafts/runs/<run>.md`; the gate says what a reader would still trip on. 4. Lecture notes: `magi familiar list` shows what was asked for. Write `drafts/lectures/<concept>.md` from the concept card, the `raw/` sources and *the place a claim of this run uses it* — "in [[claim]] it enters like this" — pitched at what the list marks known. Inside a run, register a step first.
(`magi wiki add-concept --name "…" --source "…" --content "…"`) or plain words.
appears. A slug means nothing to the person reading.
linked to its lecture notes.
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
Take a folder that already holds research — papers, notes, code, half-written drafts — into MAGI, without breaking what is…
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.
Talk a run over with the person until it can be written down — motivation, directions, what is not worth it, when to stop, their…
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.
Use this skill to extract structured Markdown/JSON from PDFs and document images—tables with cell-level precision, formulas as…
Estimates intracellular metabolic fluxes from steady-state carbon-13 isotope-tracing measurements using validated atom maps,…