paper-review
Skill442 starsChanged 21 days agoMITView on GitHub
Conduct reviewer-style, evidence-safe reviews of academic manuscripts across disciplines using progressive method, domain, review-round, and optional reviewer overlays. Use when the user asks for a critique, scientific review, priority comments, top-to-bottom review, revision-round assessment, prior-comment regression check, or review memo rather than ordinary drafting. Infer the smallest applicable module set from the conversation and manuscript; ask only when unresolved domain or review-profile ambiguity would materially change the review. Includes optional psychometrics and SEM, computational and LLM, water and CNHS, flood and hydrodynamics, and explicit Ethan-style modules.
29 other projects use a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- unravel11/research-paper-skillstheir own version, first seen Nov 16, 2026
- adithyanraj03/dsh-skills-packtheir own version, first seen Nov 15, 2026
- angadhn/botferencetheir own version, first seen Nov 13, 2026
- Agentic-Assets/corbis-mcptheir own version, first seen Nov 11, 2026
- bas1l/claude-skillsedited, first seen Nov 11, 2026
- kfangw/research-opstheir own version, first seen Nov 8, 2026
Used with
Projects with it also use paper-writing.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
MLNLP-WorldSkillpaper-writing
学术论文写作检查与优化助手。基于 MLNLP-World 社区整理的论文写作技巧,帮助检查和优化学术论文。Use when: (1) 检查论文 LaTeX 格式和排版, (2) 优化公式符号使用, (3) 改进图表设计, (4) 润色英文学术表达, (5) 检查参考文献格式, (6) 投稿前终稿检查, (7) 用户询问论文写作技巧或规范。
Also in 6 projects that use this
Other versions29 projects use a version of this.
Word for word · 0
None found.
Edited · 1
- bas1l/claude-skills★ 0100% same
Wrote their own version · 28
- EvoScientist/EvoSkills★ 486same 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★ 45same name
- bimajanuri/paper-review-skill★ 24same name
- TakaGoto/rag-learning-academy★ 24same name
- angadhn/botference★ 20same name
- borgr/paper-sharpener★ 19same name
- guruvamsi-policharla/paper-review-skill★ 17same name
18 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 repository6 other files it gives its agents.
- AGENTS.mdAGENTS.md
- academic-writing-skillsSkill
- academic-writing-skills-zh-cnSkill
- academic-writing-skills-zh-twSkill
- paper-review-zh-cnSkill
- paper-review-zh-twSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
15 rules. 0 are common, 13 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 20 edited versions we read keep these.
13
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 10
Use academic-writing-skills as the manuscript-integrity base for lifecycle, evidence, prose, change-impact, and release checks.
- Distinctive: line 12
Do not edit the manuscript unless the user authorizes revision.
- Very specific: line 26
Prefer an explicit review round.
- Distinctive: line 28
Read overlay-contract.md before applying optional modules.
- Distinctive: line 52
Keep reusable domain knowledge in a direct reference file rather than expanding this routing file.
- Distinctive: line 61
Add the module to the routing list and architecture tests.
- Distinctive: line 67
Framing and contribution: verify that the literature supports the gap, the task matches the methods, and the contribution does not exceed …
- Distinctive: line 68
Design and methods: verify sampling or source selection, measurement, model logic, estimand, assumptions, uncertainty, comparison design, …
- Distinctive: line 69
Results and claim scope: verify question-to-result coverage, statistical or computational evidence, alternative explanations, mechanisms, …
- Distinctive: line 71
Displays and notation: verify figures, tables, equations, captions, units, symbols, panels, text–display agreement, and the provenance of …
- Distinctive: line 72
Writing and delivery: verify terminology, repetition, observable stock phrasing, flow, summaries, citations, companion files, metadata, and …
- Distinctive: line 78
Do not invent mechanisms, citations, facts, thresholds, analyses, prior comments, or reviewer preferences.
- Distinctive: line 80
Do not infer AI authorship from prose.
- Distinctive: line 110
Follow the manuscript's actual reader functions and structure.
- Distinctive: line 122
Use the base retrospective.
“Common” is measured: kept by at least 6 in 10 of the 20 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
Use the Manuscript-Integrity Base
Use academic-writing-skills as the manuscript-integrity base for lifecycle, evidence, prose, change-impact, and release checks. Apply this skill as a reviewer orchestrator that selects only the method, domain, review-round, venue, or reviewer modules needed for the manuscript.
Do not edit the manuscript unless the user authorizes revision. A request to review, critique, comment, or assess a round is non-mutating. Use the appropriate document or PDF skill for extraction, comments, tracked changes, and visual rendering.
Establish the Review Basis
Identify:
- manuscript and companion artifacts in scope
- review-only versus authorized revision
- manuscript archetype, study design, lifecycle stage, and intended venue
- main questions, outcomes, methods, evidence, and claimed contribution
- prior draft, prior comments, responses, and round label when supplied
- candidate method, domain, venue, reviewer, and project overlays
- unavailable sources or ambiguities that could change the review standard
Prefer an explicit review round. If none is supplied, use developmental, substantive, integration, or submission; do not infer R1–R4 from prose quality, filename, or comment count.
Read overlay-contract.md before applying optional modules. Read round-calibration.md only when review maturity or cross-round progress is in scope.
For a supplied rubric, coded check report, or version-comparison review, use review-checklist-and-cases.md to record applicable checks and evidence-backed repairs. It adds traceability and worked cases, not another mandatory full-paper pass.
Select Modules Progressively
Infer candidate modules from the current conversation, title, Abstract, questions, Methods, tables, figures, equations, and supplement. Select the smallest set that covers the actual study. State the selected modules in the review header.
Ask one targeted question only when the available evidence leaves two materially different review standards plausible—for example, whether a latent-variable model is confirmatory or exploratory, or whether the user wants a neutral review versus a named reviewer style. Do not ask when the manuscript itself resolves the choice.
Load:
- display-notation-provenance.md for equations, formal notation, derived display quantities, normalization or aggregation rules, uncertainty displays, or cross-artifact symbol consistency
- quantitative-psychometrics-sem.md for surveys, scales, psychometrics, factor analysis, latent variables, SEM, mediation, or multi-group comparison
- ai-llm-computational.md for simulations, ABM, coupled computational models, ML, GenAI, LLMs, synthetic respondents, or agent evaluation
- water-cnhs-uncertainty.md for water resources, coupled natural–human systems, water policy, frameworks, reviews, uncertainty, or equifinality
- flood-hydrodynamics-catastrophe.md for flood risk, catastrophe models, inundation, drainage, hydrodynamics, hazards, exposure, vulnerability, or loss
- ethan-style-overlay.md only when the user explicitly requests Ethan-style review or establishes that lab review context
- project-precedents.md only after the exact named project is established and only together with an explicit relevant reviewer or lab context
Load multiple modules for genuinely hybrid studies. A citation, keyword, or study-area mention alone does not establish a module. Treat named methods, thresholds, citations, and reporting practices inside modules as checks to evaluate, not assumptions that the manuscript must follow them.
Support New Domain Modules
Keep reusable domain knowledge in a direct reference file rather than expanding this routing file. A contributed module should contain:
- explicit trigger and exclusion cues
- method and evidence checks
- claim-scope and alternative-explanation checks
- figure, table, equation, and reproducibility checks where relevant
- terminology cautions without project facts
- at least one routing or boundary eval
Add the module to the routing list and architecture tests. Keep professor, laboratory, journal, and project rules in separate conditional overlays.
Run the Review
Apply the base skill's four passes, then use the selected modules to deepen—not duplicate—the review:
- Framing and contribution: verify that the literature supports the gap, the task matches the methods, and the contribution does not exceed the evidence.
- Design and methods: verify sampling or source selection, measurement, model logic, estimand, assumptions, uncertainty, comparison design, and reproducibility as applicable.
- Results and claim scope: verify question-to-result coverage, statistical or computational evidence, alternative explanations, mechanisms, and formal versus descriptive comparisons.
- Discussion and synthesis: distinguish direct findings, literature-supported interpretation, plausible alternatives, implications, limitations, and speculation.
- Displays and notation: verify figures, tables, equations, captions, units, symbols, panels, text–display agreement, and the provenance of derived quantities. When the display-and-notation module applies, use its symbol ledger and provenance trace rather than checking each artifact in isolation.
- Writing and delivery: verify terminology, repetition, observable stock phrasing, flow, summaries, citations, companion files, metadata, and named-stage readiness.
For every material issue, inherit the base skill's scope, S0–S4 severity, evidence status, issue status, authority needed, affected artifacts, and blocking stage. If an optional reviewer overlay uses MUST, SHOULD, QUERY, or PREFERENCE, keep those labels separate from severity.
Protect Evidence and Reviewer Boundaries
Do not invent mechanisms, citations, facts, thresholds, analyses, prior comments, or reviewer preferences. A request to explain “why” does not authorize a causal story unsupported by analysis or literature.
Do not infer AI authorship from prose. Report observable features such as vague subjects, stock transitions, empty intensifiers, repetitive cadence, duplicated synthesis, or unsupported mechanisms.
Do not call a cross-round issue resolved without the prior instruction and current evidence. A response letter or resolved comment thread is evidence of an author claim, not evidence that the current artifact contains the requested change. Do not make a preference blocking unless a verified venue, reporting, author, or reviewer requirement makes it so.
Output Format
Adapt the memo to the available artifact and stage. Do not invent comments for absent sections or target a fixed number.
Header
REVIEW PROFILE: [general scientific / named conditional overlay]
PAPER: [title]
MODE: [review only / authorized revision]
STAGE OR ROUND: [maturity label or supplied round]
MODULES: [selected references and why]
SOURCE BASIS: [current files, prior material, unavailable sources]
READINESS: [named next-stage readiness and primary bottleneck]Cross-Round Status
Include only with prior evidence. Classify issues as resolved, partial, open, regressed, waived, new, or not verifiable.
Priority-Ranked Action Items
For each material issue, give its location, classification, severity and evidence status, specific problem, bounded action or decision, affected artifacts, and whether it blocks the named next stage.
Section-Linked Comments
Follow the manuscript's actual reader functions and structure. Anchor comments to short fragments. Distinguish validity, reporting, domain convention, venue rule, reviewer preference, and local prose improvement.
Writing and Presentation Flags
Include representative cases and a revision direction. Do not rewrite unless authorized.
Next-Step Instruction
State which issues, decisions, sources, and artifacts must be addressed before the next review.
Functional-Completeness Retrospective
Use the base retrospective. Name the lifecycle stage, modules loaded, selection evidence, unresolved ambiguity, prior-round evidence, artifacts covered, change propagation, checks performed, and readiness limits.
Safety and Restraint
Be direct, technically precise, and respectful. Do not impersonate a named reviewer, express frustration, or use humiliation. Never suppress a validity, ethics, evidence, or metadata issue because the manuscript is in a late round.
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.

