Agent skill

citecheck

Use when the user asks to "verify citations", "check references", "validate paper citations", or "evaluate reference relevance". It extracts references from LaTeX/PDF papers, checks formatting rules, verifies existence via Crossref / Semantic Scholar / OpenAlex / PubMed / arXiv / dblp / Google Scholar / WebSearch, and scores thematic/semantic relevance.

color4-altgithub.com/color4-altGitHub ↗
claude-codeMIT
Install
npx skills add color4-alt/CiteCheck --skill citecheck --agent claude-code

Same command for any agent — swap --agent for codex, cursor, copilot.

Facts
Files in the skill folder: 7
SKILL.md size: 5 KB
Bundled scripts: none
Version: 0.1.0
Path: skills/citecheck/SKILL.md
Open the folder on GitHub →
Where it comes from
Stars: 55
Language: Python

Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.

From the SKILL.md

# CiteCheck — Paper Citation Verification ## Overview CiteCheck verifies academic paper citations by combining structured parsing with agent-native LLM evaluation. It supports LaTeX source files (preferred) and PDF fallback. ## Workflow 1. **Parse paper**: Call `citecheck` CLI to read LaTeX (preferred) or PDF, extract references and body text 2. **Format check**: Call `citecheck` CLI to validate bibliography entries 3. **Queryability verification**: Call `citecheck` CLI to verify existence via Crossref / Semantic Scholar 4. **Evaluate thematic relevance**: Evaluate directly using the agent's reasoning capabilities — compare cited paper title/abstract/venue against the citing paper 5. **Evaluate semantic accuracy**: Evaluate directly using the agent's reasoning capabilities — compare in-text citation context against cited source content 6. **Generate report**: Aggregate all results into a Markdown report > **Why matching steps are not done by the CLI** > The `citecheck` CLI can run standalone with optional `--api-key` for external LLM-powered matching. When used as a Skill, the host agent itself possesses LLM reasoning capabilities. Direct evaluation is faster, more consistent, and

What's inside
Steps it walks through
  1. Overview
  2. Workflow
  3. 1. Parse Paper (CLI)
  4. LaTeX source (preferred)
  5. PDF (fallback)
  6. 2. Format Check (CLI)
  7. 3. Queryability Verification (CLI + Agent Supplement)
  8. 4. Evaluate Thematic Relevance (Agent Direct)
  9. 5. Evaluate Semantic Accuracy (Agent Direct)
  10. 6. Output Report
  11. Summary
  12. Detailed Results Table
  13. Issue Summary
  14. Dependencies
Ships with 6 files
  • examples/example-latex-verification.md
  • examples/example-pdf-verification.md
  • references/api-reference.md
  • references/format-check-rules.md
  • references/semantic-matching-prompt.md
  • references/thematic-scoring-prompt.md
Commands it runs
citecheck path/to/latex_project/ --skip-verification --skip-semantic -o parsed_report.md
citecheck main.tex --skip-verification --skip-semantic -o parsed_report.md
citecheck paper.pdf --skip-verification --skip-semantic -o parsed_report.md
pip install CiteCheck
pip install CiteCheck[pdf]
About this skill
What does the citecheck skill do?

Use when the user asks to "verify citations", "check references", "validate paper citations", or "evaluate reference relevance". It extracts references from LaTeX/PDF papers, checks formatting rules, verifies existence via Crossref / Semantic Scholar / OpenAlex / PubMed / arXiv / dblp / Google Scholar / WebSearch, and scores thematic/semantic relevance.

How do I install it?

Run `npx skills add color4-alt/CiteCheck --skill citecheck --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.

Where does this skill come from?

From color4-alt/CiteCheck, a repository with 55 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.

Is a popular skill a good skill?

Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.

Keep going