Citation VerificationSAFE
Semi-automated research assistant for academic research and software development. Supports Claude Code, Codex CLI, Kimi Code CLI, and OpenCode across ideation, coding, experiments, writing, and publication.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
文件用途
本目录中的文件提供背景知识和参考信息,用于理解引用验证的原理和常见问题。
重要: 这些文件不是主要工作流的一部分。实际的引用验证优先使用 arXiv、DOI/CrossRef、Semantic Scholar、publisher metadata 和 Zotero metadata。Google Scholar 只能作为人工发现或 fallback。
文件说明
common-errors.md
内容: 常见引用错误模式和修复方法
用途:
- 了解学术写作中常见的引用错误
- 学习如何识别和修复这些错误
- 理解为什么需要验证引用
何时参考: 当需要了解引用错误的类型和修复方法时
verification-rules.md
内容: 详细的验证规则和匹配算法
用途:
- 理解引用验证的完整逻辑
- 了解如何匹配标题、作者、年份等信息
- 学习验证的技术细节
何时参考: 当需要深入了解验证逻辑时
api-usage.md
内容: API 使用指南(CrossRef、arXiv、Semantic Scholar)
用途:
- 了解学术API的使用方法
- 理解API验证的原理
- 参考高级用例的实现
何时参考: 当需要了解API验证方法时。当前主要工作流应优先使用这些 programmatic / canonical sources。
主要工作流
实际的引用验证应该使用 `ml-paper-writing` skill 中的 Citation Workflow:
- 查找 DOI、arXiv ID、publisher page 或 verified Zotero item
- 用 CrossRef、arXiv、Semantic Scholar、publisher metadata 或 Zotero metadata 验证
- 从 programmatic/canonical source 获取 BibTeX
- 验证声明(如需要)
- 添加到 bibliography
详见 ml-paper-writing skill 的 "Citation Workflow (Hallucination Prevention)" 部分。
使用建议
对于日常论文写作:
- ✅ 使用
ml-paper-writingskill 的 Citation Workflow - ✅ 使用 arXiv、DOI/CrossRef、Semantic Scholar、publisher metadata 和 Zotero metadata
- ✅ 必要时用 Google Scholar 做人工发现,但不要把它作为 canonical authority
- ✅ 参考这些文件了解背景知识
- ❌ 不要从记忆或未经验证的 Google Scholar 条目生成最终 BibTeX
对于理解验证原理:
- 阅读
common-errors.md了解常见错误 - 阅读
verification-rules.md了解验证逻辑 - 阅读
api-usage.md了解API方法(参考用)
更多信息
详见 citation-verification skill 的 SKILL.md 文件。
29ad4d4206fbOBSERVED · 2026-10-07Install
Commands as the repository documents them. They are shown, not run.
pip install semanticscholar
pip install arxiv
pip install bibtexparser requests semanticscholar arxiv
What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
---
name: citation-verification
description: This skill provides reference guidance for citation verification in academic writing. Use when the user asks about "citation verification best practices", "how to verify references", "preventing fake citations", or needs guidance on citation accuracy. This skill supports ml-paper-writing by providing detailed verification principles and common error patterns.
tags: [Research, Academic, Citation, Reference]
version: 0.1.0
---
# Citation Verification Reference Guide
A reference guide for citation verification in academic paper writing, providing verification principles and best practices.
**Core Principle**: Proactively verify every citation during the writing process using programmatic or canonical scholarly sources first: arXiv, DOI/CrossRef, Semantic Scholar, publisher landing pages, and Zotero metadata. Google Scholar is useful for manual discovery, but it is not the canonical verification authority.
## Core Problems
Citation issues in academic papers seriously impact research integrity:
1. **Fake citations** - Citing non-existent papers (common issue with AI-generated citations)
2. **Incorrect information** - Mismatched authors, titles, years, etc.
3. **Inconsistent formatting** - Mixed citation formats
4. **Missing citations** - Referenced but uncited work
These issues can lead to:
- Paper rejection or retraction
- Damage to academic reputation
- Reviewers questioning research rigor
**Special risk with AI-assisted writing**: AI-generated citations have approximately 40% error rate; every citation must be verified via WebSearch.
## Verification Principles
This skill provides verification principles based on canonical scholarly metadata and claim-level checking:
### 1. Proactive Verification (Verify During Writing)
**Core idea**: Verify immediately when adding a citation, rather than checking after writing is complete.
- Search for the paper via WebSearch each time a citation is needed
- Confirm the paper exists on Google Scholar
- Add to bibliography only after verification passes
### 2. Canonical Metadata Verification
Preferred authority order:
1. DOI / publisher landing page
2. arXiv ID or arXiv landing page
3. CrossRef
4. Semantic Scholar
5. Zotero metadata imported from a verified identifier
6. Google Scholar only for manual discovery or fallback lookup
**Verification steps**:
1. Find a DOI, arXiv ID, publisher URL, or verified Zotero item.
2. Confirm title, first author, year, venue, and identifier.
3. Fetch BibTeX from CrossRef, arXiv, publisher metadata, Zotero, or another programmatic source when possible.
4. If only Google Scholar can find the item, mark it as manual verification and do not treat the BibTeX as final until metadata is checked elsewhere.
### 3. Information Matching Verification
**Information that must match**:
- Title (minor differences allowed, e.g., capitalization)
- Authors (at least the first author must match)
- Year (±1 year difference allowed, considering preprints)
- Publication venue (conference/journal name)
### 4. Claim Verification
**Key principle**: When citing a specific claim, you must confirm the claim actually appears in the paper.
- Use WebSearch to access the paper PDF
- Search for relevant keywords
- Confirm the accuracy of the claim
- Record the section/page where the claim appears
## Verification Workflow
### Integration into Writing Process
```
Need a citation during writing
↓
Find DOI / arXiv ID / publisher page / verified Zotero item
↓
Verify metadata with CrossRef / arXiv / Semantic Scholar / publisher / Zotero
↓
Confirm paper details
↓
Get BibTeX
↓
(If citing a specific claim) Verify the claim
↓
Add to bibliography
```
**Key point**: Verification is part of the writing process, not a separate post-processing step.
## Usage Guide
### Using with ml-paper-writing
The verification principles of this skill are integrated into the Citation Workflow of the `ml-paper-writing` skill.
**Auto-trigger**: Citation verification is automatically executed when writing papers with the ml-paper-writing skill.
**Manual reference**: Refer to this skill when you need detailed verification principles.
### Verification Step Example
**Scenario**: Need to cite the Transformer paper
```
Step 1: WebSearch lookup
Query: "Attention is All You Need Vaswani 2017"
Result: Found multiple sources for the paper
Step 2: Google Scholar verification
Query: "site:scholar.google.com Attention is All You Need Vaswani"
Result: ✅ Paper exists, 50,000+ citations, NeurIPS 2017
Step 3: Confirm details
- Title: "Attention is All You Need"
- Authors: Vaswani, Ashish; Shazeer, Noam; Parmar, Niki; ...
- Year: 2017
- Venue: NeurIPS (NIPS)
Step 4: Get BibTeX
- Click "Cite" on Google Scholar
- Select BibTeX format
- Copy BibTeX entry
Step 5: Add to bibliography
- Paste into .bib file
- Use \cite{vaswani2017attention} in the paper
```
### Handling Verification Failures
**If the paper cannot be verified through canonical sources**:
1. **Check spelling** - Is the title or author name correct?
2. **Try different queries** - Use different keyword combinations
3. **Find alternative sources** - Try arXiv, DOI, CrossRef, Semantic Scholar, publisher pages, or Zotero
4. **Mark as pending** - Use `[CITATION NEEDED]` marker
5. **Notify the user** - Clearly state the citation cannot be verified
**If information doesn't match**:
1. **Confirm the source** - Did you find the correct paper?
2. **Check versions** - Preprint vs. published version
3. **Update information** - Use the most accurate version
4. **Record discrepancies** - Note the reason for differences
## Best Practices
### Preventing Fake Citations
1. **Never generate citations from memory** - AI-generated citations have 40% error rate
2. **Use WebSearch to find** - Verify every citation through WebSearch
3. **Confirm on Google Scholar** - Verify paper existence on Google Scholar
4. **Verify promptly** - Verify when adding citatioTrust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- declared (3 observation(s))
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (0)
No findings outside the package's declared scope.
Gates applied: no_behavioural_pass.
29ad4d4206fbfull audit observations/trust-audit/skill/galaxy-dawn__citation-verification.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-07 | 29ad4d4206fb | SAFE | B | 89 | first audit |
Questions
What does the Citation Verification skill do?
Semi-automated research assistant for academic research and software development. Supports Claude Code, Codex CLI, Kimi Code CLI, and OpenCode across ideation, coding, experiments, writing, and publication.
Is Citation Verification safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can Citation Verification access on my machine?
The audit observed that it reaches the network. Each of those is consistent with what it says it does. Secrets in the source: none found.
How current is this page?
The grade is for one exact copy of the source (29ad4d4206fb), read on 2026-10-07. The repository is watched, and a new audit runs when it changes — this is the first audit.