paper-review
Skill14 starsChanged 4 months agoMITView on GitHub
Use when the user wants reviewer feedback on a draft paper before submission, mentions paper review, reviewer simulation, PC review, or rebuttal preparation
27 other projects use a version of this.
Adopted by
Most recently taken up by
- kfangw/research-opstheir own version, first seen Nov 8, 2026
- 70asunflower/70asunflower-skillstheir own version, first seen Nov 8, 2026
- Go-MinSeong/paper-review-servicetheir own version, first seen Nov 8, 2026
- bas1l/claude-skillstheir own version, first seen Nov 3, 2026
- angadhn/botferencetheir own version, first seen Oct 30, 2026
- TakaGoto/rag-learning-academytheir own version, first seen Oct 28, 2026
Other versions27 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 27
- EvoScientist/EvoSkills★ 481same name
- WenyuChiou/academic-writing-skills★ 405same name
- aspi6246/Claude-Code-Skills-for-Academics★ 153same name
- voidful/academic-skills★ 126same name
- AlexWortega/ai-peer-review-skill★ 63same name
- drpwchen/paper-review-and-digest★ 39same name
- bimajanuri/paper-review-skill★ 24same name
- TakaGoto/rag-learning-academy★ 24same name
- angadhn/botference★ 20same name
- borgr/paper-sharpener★ 19same name
17 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- writing-skillsSkill
- academic-pptx-skillSkill
- analyze-plotSkill
- paper-extractSkill
- paper-plotSkill
- paper-writingSkill
- systematic-researchSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 2 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 19 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 58
Before Step 2: all 4 domains covered, expertise 1-5 spread (min one 1-2, one 5), distinct tastes, user approved.
- Distinctive: line 70
Before Step 3: all 5 reviews complete, each follows template structure, each persona-consistent.
“Common” is measured: kept by at least 6 in 10 of the 19 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Paper Review
Core principle: Realistic review requires DIVERSE PERSONAS. Five subagents embody distinct reviewers (unique domain, expertise 1-5, taste) and review in parallel. Each supports interactive QA with persistent memory.
<HARD-GATE>
Step 1 MUST complete before Step 2. Step 2 MUST complete before Step 3.
Main agent MUST NOT review — only coordinate: profiles -> dispatch -> collect -> QA.
</HARD-GATE>Process Flow
digraph paper_review {
rankdir=TB;
"Start" [shape=oval];
"Existing reviews?" [shape=diamond];
"Ask: re-review, QA, exit?" [shape=diamond];
"Step 1: Generate 5 profiles\n[reviewer-profiles.md]" [shape=box];
"User approves?" [shape=diamond];
"Step 2: Discover template\nfrom templates/, dispatch 5 subagents" [shape=box];
"All 5 done?" [shape=diamond];
"Step 3: QA with reviewer?\n[qa-protocol.md]" [shape=diamond];
"Load profile+history,\nrespond in character\n(use scripts/qa-entry)" [shape=box];
"More QA?" [shape=diamond];
"Done — offer aggregate\n(scripts/aggregate-report)" [shape=oval];
"Start" -> "Existing reviews?";
"Existing reviews?" -> "Ask: re-review, QA, exit?" [label="yes"];
"Existing reviews?" -> "Step 1: Generate 5 profiles\n[reviewer-profiles.md]" [label="no"];
"Ask: re-review, QA, exit?" -> "Step 1: Generate 5 profiles\n[reviewer-profiles.md]" [label="re-review"];
"Ask: re-review, QA, exit?" -> "Step 3: QA with reviewer?\n[qa-protocol.md]" [label="QA"];
"Ask: re-review, QA, exit?" -> "Done — offer aggregate\n(scripts/aggregate-report)" [label="exit"];
"Step 1: Generate 5 profiles\n[reviewer-profiles.md]" -> "User approves?";
"User approves?" -> "Step 1: Generate 5 profiles\n[reviewer-profiles.md]" [label="no"];
"User approves?" -> "Step 2: Discover template\nfrom templates/, dispatch 5 subagents" [label="yes"];
"Step 2: Discover template\nfrom templates/, dispatch 5 subagents" -> "All 5 done?";
"All 5 done?" -> "Step 2: Discover template\nfrom templates/, dispatch 5 subagents" [label="no"];
"All 5 done?" -> "Step 3: QA with reviewer?\n[qa-protocol.md]" [label="yes"];
"Step 3: QA with reviewer?\n[qa-protocol.md]" -> "Load profile+history,\nrespond in character\n(use scripts/qa-entry)" [label="yes"];
"Step 3: QA with reviewer?\n[qa-protocol.md]" -> "Done — offer aggregate\n(scripts/aggregate-report)" [label="no"];
"Load profile+history,\nrespond in character\n(use scripts/qa-entry)" -> "More QA?";
"More QA?" -> "Load profile+history,\nrespond in character\n(use scripts/qa-entry)" [label="yes"];
"More QA?" -> "Done — offer aggregate\n(scripts/aggregate-report)" [label="no"];
}Step 1: Profiles
Generate 5 reviewer profiles. Reference: reviewer-profiles.md. Output: review/reviewers/profiles.md.
<HARD-GATE-STEP1>
Before Step 2: all 4 domains covered, expertise 1-5 spread (min one 1-2, one 5), distinct tastes, user approved.
</HARD-GATE-STEP1>Step 2: Parallel Review
List templates/ recursively to discover available review templates. Match by target venue. Load the matching review-guid.md.
Dispatch 5 subagents simultaneously. Each receives: profile + template + full paper + output path (review/reviews/reviewer<N>/review.md).
Task prompt: "You ARE Reviewer N. Use template at <path>. Reflect your domain, expertise, and taste. Include ALL template sections. Do NOT write as generic AI."
<HARD-GATE-STEP2>
Before Step 3: all 5 reviews complete, each follows template structure, each persona-consistent.
</HARD-GATE-STEP2>Step 3: Interactive QA
Reference: qa-protocol.md. Use scripts/qa-entry.cmd (Win) or scripts/qa-entry.sh (Linux) to append entries. Load profile + full QA history before every response. Answer in character.
<HARD-GATE-STEP3>
Review MUST exist before QA. Append with script, not manually. Record Internal Reasoning for consistency.
</HARD-GATE-STEP3>Completion
Offer scripts/aggregate-report.cmd (Win) or scripts/aggregate-report.sh (Linux). Output: review/aggregate-report.md.
Guardrails
| Red flag | Reality |
| "I'll review without profiles" | No profiles = generic reviews. Step 1 first. |
| "All reviewers should be experts" | Non-experts reveal exposition gaps. Spread 1-5. |
| "I'll summarize all 5 into one" | Collapsing personas = missing the point. |
| "Same review criteria for all papers" | List templates/, discover types, match by venue. |
| "I remember the profile, no need to reload" | Load profile + QA history before EVERY response. |
| "Reviews look similar = consensus" | Similar reviews = persona collapse, not validation. |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.