Skip to content
Back to search

sci-brain

Brain storming like a senior research scientist

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/quantumbfs--sci-brain/badge.svg)](/listings/quantumbfs--sci-brain)

HTML

<a href="/listings/quantumbfs--sci-brain"><img src="/api/listings/quantumbfs--sci-brain/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

**Claude Code:**

/plugin marketplace add QuantumBFS/sci-brain

**Codex / OpenCode:** clone this repo and symlink its skill directories into your agent's discovery path — see [`.codex/INSTALL.md`](.codex/INSTALL.md) or [`.opencode/INSTALL.md`](.opencode/INSTALL.md). Or simply tell your agent:

Install the plugin/skills from https://github.com/QuantumBFS/sci-brain

**pi:** install the npm package:

pi install npm:sci-brain

or clone the repo and symlink its skill directories into `~/.agents/skills/` (pi's global skill path, shared with Codex) — see [`.pi/INSTALL.md`](.pi/INSTALL.md).

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 14 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "name": "sci-brain",
  3. "owner": {
  4. "name": "Jin-Guo Liu"
  5. },
  6. "plugins": [
  7. {
  8. "name": "sci-brain",
  9. "description": "Scientific research brainstorming partner: survey the literature, find good problems, and shape concrete research ideas together",
  10. "version": "0.5.0",
  11. "source": "./"
  12. }
  13. ]
  14. }
README.md · README · 138 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # sci-brain
  2. Research skills for [Claude Code](https://claude.ai/claude-code), [Codex](https://github.com/openai/codex), [OpenCode](https://github.com/opencode-ai/opencode), and [pi](https://github.com/earendil-works/pi). Give it a research topic — it surveys the literature into a citable knowledge base on your disk, brainstorms ideas with you, and writes the paper.
  3. ## Install
  4. **Claude Code:**
  5. ```
  6. /plugin marketplace add QuantumBFS/sci-brain
  7. ```
  8. **Codex / OpenCode:** clone this repo and symlink its skill directories into your agent's discovery path — see [`.codex/INSTALL.md`](.codex/INSTALL.md) or [`.opencode/INSTALL.md`](.opencode/INSTALL.md). Or simply tell your agent:
  9. ```
  10. Install the plugin/skills from https://github.com/QuantumBFS/sci-brain
  11. ```
  12. **pi:** install the npm package:
  13. ```
  14. pi install npm:sci-brain
  15. ```
  16. or clone the repo and symlink its skill directories into `~/.agents/skills/` (pi's global skill path, shared with Codex) — see [`.pi/INSTALL.md`](.pi/INSTALL.md).
  17. ## Skills
  18. Two kinds, told apart by the first words of each skill's description. **User trigger** skills are the needs you type (`/survey`, `/write-paper`). **Agentic trigger** skills (`how-to-*`) are procedures the agent reaches for while serving one of those needs; you can still invoke them directly. Each table row is the skill's frontmatter description (test-enforced).
  19. ### User trigger — invoke by need
  20. | Skill | Use when |
  21. |---|---|
  22. | [`survey`](skills/survey/) | surveying a research topic into a knowledge base or writing a grounded literature, technology, or field assessment. |
  23. | [`brainstorm-ideas`](skills/brainstorm-ideas/) | brainstorming research ideas with a Socratic collaborator who learns your background and helps find a problem worth attacking. |
  24. | [`know-me-better`](skills/know-me-better/) | you want the agent to learn your research style from your papers (Zotero, PDF folder, or Google Scholar) so it speaks your language. |
  25. | [`autoresearch`](skills/autoresearch/) | starting, resuming, or checking an autoresearch campaign — topics, evidence base, sealed validator, then validator-scored attempt loops. |
  26. | [`write-paper`](skills/write-paper/) | drafting or revising a scientific manuscript with real results. |
  27. | [`review-paper`](skills/review-paper/) | reviewing, commenting on, or fact-checking an existing manuscript, including its references. |
  28. | [`write-slides`](skills/write-slides/) | building a Typst + Touying slide deck for a scientific talk, lecture, or briefing. |
  29. | [`dump-chat-history`](skills/dump-chat-history/) | exporting original conversations from selected harnesses and dates as a faithful transcript or a topic-titled Typst and PDF field note. |
  30. | [`create-advisor`](skills/create-advisor/) | creating or updating a named advisor profile from a researcher's conversation history. |
  31. Export conversation history with `/dump-chat-history`: choose one or more harnesses and a start date, then get original text with source provenance. A requested PDF uses a topic title and the bundled Typst field-note template. Research classification is a separate `/how-to-analyze-dialog` step; `/create-advisor` chains the two when needed.
  32. A typical path: `/survey` a field → `/brainstorm-ideas` with an optional [advisor](advisors/) → `/autoresearch` to run validator-scored attempts → `/write-paper` and `/write-slides`.
  33. Autoresearch supports an optional [cycle completion hook](skills/autoresearch/references/cycle-end-hook.md)
  34. to play a generated chime or run a notification command after each cycle.
  35. Set `cycle_end_hook: sound` in `research/STATE.md` to use an installed system player;
  36. notifications are disabled by default.
  37. Slide templates are maintained in [GiggleLiu/sci-brain-slides](https://github.com/GiggleLiu/sci-brain-slides). The `write-slides` skill uses the published Typst package `@preview/sci-brain-slides:0.1.0`; no template checkout is needed.
  38. > **One-time setup** for PDF rendering: `python3 -m pip install --user pymupdf4llm`.
  39. ### Agentic trigger — `how-to-*`, invoked automatically
  40. | Skill | Use when |
  41. |---|---|
  42. | [`how-to-build-kb`](skills/how-to-build-kb/) | turning a list of picked papers into verified knowledge-base entries — references.bib, INDEX.md, and NOTES.md. |
  43. | [`how-to-download-ref`](skills/how-to-download-ref/) | adding arXiv IDs or DOIs to a knowledge base — fetches metadata, PDFs, and full text, then updates references.bib and INDEX.md. |
  44. | [`how-to-write-ideas-report`](skills/how-to-write-ideas-report/) | writing a proposal-style ideas report from a finished brainstorm-ideas session or a chosen research direction. |
  45. | [`how-to-technical-writing`](skills/how-to-technical-writing/) | writing or polishing scientific prose at the sentence and paragraph level, in a manuscript, report, or review. |
  46. | [`how-to-review-figure`](skills/how-to-review-figure/) | judging the visual design of a figure, plot, or diagram against a scientific-plot rubric. |
  47. | [`how-to-flow`](skills/how-to-flow/) | one hard, testable goal resists a direct solution and needs an autonomous decide–simulate–learn–backjump search. |
  48. | [`how-to-analyze-dialog`](skills/how-to-analyze-dialog/) | classifying exported research conversations by topic and analyzing their prompting patterns across six academic dimensions. |
  49. ## Where Things Are Saved
  50. Everything lands in one folder inside your project:
  51. ```
  52. <project>/.knowledge/
  53. references.bib verified BibTeX — point your LaTeX/Typst bibliography here
  54. INDEX.md auto-generated table of contents
  55. NOTES.md curated sub-themes, open problems, bottlenecks
  56. <id>_<slug>.md full-text papers rendered to markdown
  57. .raw/ original PDFs, metadata, and optional sources (gitignored)
  58. .figures/ extracted figures (gitignored)
  59. ```
  60. - **Project knowledge base** — `<project>/.knowledge/` (layout above). Populated by `/survey`, `/how-to-download-ref`, `/know-me-better`.
  61. - **Advisor knowledge bases** — `advisors/<slug>/.knowledge/` — each advisor's private literature cache, same layout.
  62. - **History exports** — `docs/chat-history/<date>-<topic>/` — original text, provenance, and optional Typst/PDF; research analysis goes to `docs/dialog/analysis/<run-slug>/`.
  63. - **Conversation logs** — `docs/discussion/` — timestamped per session; the next session picks up where you left off.
  64. - **Ideas reports** — `articles/` in your current directory, with a matching `.bib` file.
  65. ## Want to Become an Advisor?
  66. If you've used Claude Code or Codex for research conversations and want your thinking style captured as a reusable advisor profile, just tell your agent:
  67. ```
  68. clone https://github.com/QuantumBFS/sci-brain,
  69. invoke create-advisor skill in the cloned repo to create my profile,
  70. then submit a pr,
  71. include all relevant chat history, interview output and the generated profile.
  72. ```
  73. The whole process is interactive — you review everything before it's published, and you decide whether to include the raw conversation data (for research purposes) in the PR.
  74. ## Upgrading
  75. > ⚠️ **Breaking change in v0.5.** Skills are renamed by trigger kind. Update slash commands:
  76. | Previous command | Current entry point |
  77. |------------------|---------------------|
  78. | `/download-ref` | `/how-to-download-ref` |
  79. | `/conversation-dump` | `/dump-chat-history` → `/how-to-analyze-dialog` |
  80. | `/how-to-dump-dialog` | `/dump-chat-history` for export; `/how-to-analyze-dialog` for research classification |
  81. | `/figure-taste` | `/how-to-review-figure` |
  82. | `/flow` | `/how-to-flow` |
  83. | `/paper-writer` | `/write-paper` |
  84. | `/paper-reviewer` | `/review-paper` |
  85. | `/slide-writer` | `/write-slides` |
  86. | `/incarnate` | `/create-advisor` |
  87. | `/survey` (KB-building step) | `/how-to-build-kb`, invoked by `/survey` |
  88. | `/brainstorm-ideas` (report mode) | `/how-to-write-ideas-report`, invoked by `/brainstorm-ideas` |
  89. > ⚠️ **Breaking change in v0.3.** Knowledge bases moved from per-topic registries (`~/.claude/survey/<topic>/` with `summary.md` + `references.bib`) to one `<project>/.knowledge/` per project, with `references.bib` living *inside* the KB. The `fetch-papers` skill was folded into `download-ref --from-bib`. See [`docs/kb-migration.md`](./docs/kb-migration.md) for `mv` commands.
  90. Four internal stages became modes of their goal-level skill in v0.3:
  91. | Previous command | Current entry point |
  92. |------------------|---------------------|
  93. | `/survey-writer` | `/survey` in Survey Report mode |
  94. | `/idea-writer` | `/how-to-write-ideas-report` |
  95. | `/import-dialog` | `/create-advisor` with Markdown dialog input |
  96. | `/soul-extraction` | `/create-advisor` in analysis-only mode |
  97. ## Contributors
  98. **Initiators**: [Lei Wang](https://github.com/wangleiphy) and [Jin-Guo Liu](https://github.com/GiggleLiu)
  99. ## License
  100. MIT. Feel free to adapt from the current codebase, BUT please acknowledge this package properly, thank you.

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.