agentrigs
+ submit rig
~ / rigs / chenweichiang / research-writing-kit

chenweichiang/research-writing-kit

A method for writing research papers and grant proposals with an AI, plus a self-installing toolkit for Claude Code: argument skeleton first, verified citations, retraction and uncited-claim scans, document regression, submission checks. Bi

↗ GitHub ★ 0 MIT updated today project Claude Code
share on X
ARCHETYPE
Orchestrator
A bench of specialised subagents. The main agent mostly delegates.
CONTEXT TAX · EVERY TURN
~5.5k tokens
Moderate · median rig: 2.3k · breakdown
GUARDRAILS
1/5
No YOLO mode · details

Copy this rig

# review before running: this installs third-party code
$ npx degit chenweichiang/research-writing-kit/.claude ./rig-research-writing-kit  # inspect, then merge into .claude/

MCP servers are added to Claude Code at local scope; env vars are shown as YOUR_… placeholders — we never store values. Files are fetched with degit into a separate folder so you can review before merging.

$ npx degit chenweichiang/research-writing-kit/skills/build-pdf .claude/skills/build-pdf
$ npx degit chenweichiang/research-writing-kit/skills/co-author .claude/skills/co-author
$ npx degit chenweichiang/research-writing-kit/skills/doc-regress .claude/skills/doc-regress
$ npx degit chenweichiang/research-writing-kit/skills/fetch-refs .claude/skills/fetch-refs
$ npx degit chenweichiang/research-writing-kit/skills/paper-review .claude/skills/paper-review
$ npx degit chenweichiang/research-writing-kit/skills/rebuttal .claude/skills/rebuttal
$ npx degit chenweichiang/research-writing-kit/skills/verify-citations .claude/skills/verify-citations
$ curl -fsSL --create-dirs -o .claude/agents/citation-skeptic.md https://raw.githubusercontent.com/chenweichiang/research-writing-kit/main/agents/citation-skeptic.md
$ curl -fsSL --create-dirs -o .claude/agents/clean-reviewer.md https://raw.githubusercontent.com/chenweichiang/research-writing-kit/main/agents/clean-reviewer.md
$ curl -fsSL --create-dirs -o .claude/agents/de-cadencing-scholar.md https://raw.githubusercontent.com/chenweichiang/research-writing-kit/main/agents/de-cadencing-scholar.md
$ curl -fsSL --create-dirs -o .claude/agents/en-native-editor.md https://raw.githubusercontent.com/chenweichiang/research-writing-kit/main/agents/en-native-editor.md
$ curl -fsSL --create-dirs -o .claude/agents/zh-tw-native-editor.md https://raw.githubusercontent.com/chenweichiang/research-writing-kit/main/agents/zh-tw-native-editor.md

This rig commits no guardrails. Here is the community baseline instead — the deny/ask rules most often found across all 7,204 rigs:

{
  "permissions": {
    "deny": [
      "Read(./.env)",
      "Read(**/.env)",
      "Read(~/.ssh/**)",
      "Bash(rm -rf *)",
      "Read(**/*.pem)",
      "Bash(rm -rf /)",
      "Bash(git push --force:*)",
      "Bash(sudo *)",
      "Read(.env)",
      "Bash(rm -rf /*)",
      "Read(./.env.*)",
      "Read(~/.aws/**)",
      "Bash(git push --force*)",
      "Bash(rm -rf:*)",
      "Read(**/*.key)",
      "Read(**/.env.*)",
      "Bash(sudo:*)",
      "Bash(git reset --hard*)",
      "Bash(git reset --hard:*)",
      "Read(.env.*)"
    ],
    "ask": [
      "Bash(git push:*)",
      "Bash(git push *)",
      "Bash(git commit:*)",
      "Bash(rm *)",
      "Bash(rm:*)",
      "Bash(git rebase *)",
      "Bash(wget *)",
      "Bash(npm publish:*)",
      "Bash(git commit *)",
      "Bash(gh pr merge *)"
    ]
  }
}

Skills (7)

build-pdfco-authordoc-regressfetch-refspaper-reviewrebuttalverify-citations

Subagents (5)

citation-skepticCalibrated-skeptic second review of a flagged citation. It re-examines a single "this citation may be wrong" verdict, presuming the citation is correct and upholding the accusation only on verbatim co
clean-reviewerClean-context final review and adjudication. It carries none of the drafting history and reads the work as a reviewer would, to find the real weaknesses in a manuscript, in a reviewer-comment verdict
de-cadencing-scholarNative-English-scholar de-cadencing pass. Before an English draft ships, it finds and rewrites the rhythm patterns that statistical fingerprint tools stay green on but a human eye instantly reads as "
en-native-editorNative-English academic editor for a paper, proposal or response letter. The yardstick is the register of published papers in the target venue (the register_profile deviation list, Biber grammar featu
zh-tw-native-editorPart of the zh-TW addon. Native Taiwan-Mandarin academic editor for a Chinese paper, proposal or application. The yardstick is the register of published Taiwan journal papers in the same field (the re

Similar rigs

nicehiro/dotfiles
—
★ 7 setup Claude CodeCodex 1 mcp 17 skills 2 hooks 9d ago
Skill Collector 6.5k tok ·
ajbarea/techne
A Claude Code plugin marketplace: skills for code repos (techne) and documents (graphe), opt-in git guards (phylax), and a spoken Claude Code (keryx).
★ 1 project Claude Code 22 skills today
Skill Collector 1.8k tok ·
angadhn/botference
One room, three minds: a multi-LLM planning council for your terminal — you, Claude Code, and Codex debate and ship an implementation plan.
★ 19 project Claude Code 6 skills 19 agents 5d ago
Orchestrator 700 tok ·
TakaGoto/rag-learning-academy
A structured, multi-agent Claude Code learning environment for mastering Retrieval-Augmented Generation (RAG)
★ 7 project Claude CodeCodex 21 skills 20 agents 4 hooks 6mo ago
Fort Knox 3.8k tok ·
copied ✓