Skip to content

momo-land/PHD-management-skills / skills/phd-review

phd-review

Verified

对实际论文稿件或研究计划进行有证据的批判审查,找出论证、方法、基线、可复现性和局限方面的问题并形成修订建议。适用于模拟审稿或投稿前检查;不伪装成真实审稿人意见。

FileKindSize
SKILL.en.mdmarkdown2.8 KB
SKILL.mdskill md2.3 KB
agents/openai.yamlconfig281 B

SKILL.en.md

View on GitHub
---
name: phd-review
description: Critically review an actual manuscript or research plan using evidence, identifying issues in arguments, methods, baselines, reproducibility and limitations. Use for simulated peer review or a pre-submission check, clearly distinguishing AI suggestions from genuine reviewer feedback.
---

# Critical paper review

Read the [local workflow](../phd-manager/references/local-workflow.en.md). Identify the actual manuscript/plan and version, then read associated paper, experiment and literature records. Without a manuscript, explain that the review covers only the plan or metadata. Reply in English to English requests and preserve original research content unless revision or translation is requested.

## Review

1. State the central claim and chain of validation, checking alignment between the research question, contribution and evidence.
2. Where applicable, assess whether related work supports the claimed gap; methods and data support reproduction; baselines are comparable; splitting and analysis avoid leakage/bias; conclusions stay within results; and limitations address important assumptions.
3. For each finding, give a specific location or evidence, its effect on the conclusion, and a feasible correction or validation step. Prioritize by research validity and clarity, not a target number of findings.
4. Distinguish demonstrated problems, missing information and questions requiring further experiments. Do not declare an unread reference incorrect or treat an unreported parameter as proof an experiment never ran.
5. Verify external facts against actual sources when needed and state any unresolved checks. Assess novelty within the literature examined; do not promise exhaustive coverage or an acceptance probability.

## Save and deliver

- Clearly label the output **AI review suggestions**. When saving to the workspace, use `reviews` with `focus`, `assessment`, `recommendations`, `evidence` and `origin: "ai"`. Notes may include the manuscript path and associated paper title/ID; do not invent a `paperId` field on reviews.
- Preserve the manuscript unless revision is also requested. Continue with [phd-paper](../phd-paper/SKILL.en.md) when asked to implement changes. Register associated tasks when planning is requested, without marking them completed.
- Check the write result and read it back. Deliver the key issues, evidence locations and revision order. If there are no material findings supported by the evidence, say so and state the review's coverage.

## Imported upstream resources

When the task benefits from the original methods or bundled tools, use the [GitHub integration](../github-peer-review/SKILL.en.md) and read its complete upstream instructions. Check its requirements and scope; tool output does not replace the local record workflow.
receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementationStars: 323kUpdated Claude CodeCodexVerifiedInstall
requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirementsStars: 323kUpdated Claude CodeCodexVerifiedInstall
using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallbackStars: 323kUpdated Claude CodeCodexVerifiedInstall
verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions alwaysStars: 323kUpdated Claude CodeCodexVerifiedInstall