Getting it into your agent
This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.
/plugin marketplace add JeonKH81/claude-cove-reference-verifier/plugin install cove-reference-verifierWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/jeonkh81/claude-cove-reference-verifier/cove-reference-verifier)<a href="https://agentmods.dev/skills/jeonkh81/claude-cove-reference-verifier/cove-reference-verifier"><img src="https://agentmods.dev/badge/skills/jeonkh81/claude-cove-reference-verifier/cove-reference-verifier/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/jeonkh81/claude-cove-reference-verifier/cove-reference-verifier"><img src="https://agentmods.dev/badge/skills/jeonkh81/claude-cove-reference-verifier/cove-reference-verifier.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00237 | $0.04782 |
| Opus 5 | $0.00118 | $0.02391 |
| Sonnet 5 | $0.00047 | $0.00956 |
| Haiku 4.5 | $0.00024 | $0.00478 |
Grade A, and why
cove-reference-verifier scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 275 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CoVe Reference Verifier
본 skill은 LLM이 학술 원고에서 흔히 만들어내는 reference hallucination(존재하지 않는 PMID·잘못된 저자·헷갈린 저널·실제 abstract에 없는 주장)을 감지하고 수정하기 위한 도구다. 이론적 근거는 references/cove_method.md에 정리된 Chain-of-Verification (Dhuliawala et al., ACL Findings 2024) Factor+Revise 변형이다.
언제 이 skill이 트리거되어야 하나
- 사용자가 작성한 manuscript draft, review 원고, 학회 발표 자료 등의 reference 리스트 검증 요청.
- "이 인용 맞아?", "PMID 진짜 있어?", "이 논문이 실제로 그런 주장을 했어?" 같은 in-text citation 사실확인.
- 다른 LLM(혹은 본인의 이전 draft)이 만들어낸 reference list의 일괄 검증.
- 한국어/영어 모두 지원. Vancouver, AMA, APA 어떤 스타일이든 입력 가능.
이 skill은 새 문헌을 검색하지 않는다. 그 작업은 clinical-research-harness:lit-search 등 별도 skill의 영역이다.
핵심 원칙 (반드시 준수)
CoVe 논문 Section 3과 본 skill의 references/cove_method.md에서 도출된 4가지 비협상(non-negotiable) 원칙:
- Factored execution: 각 verification question(Q1–Q7)은 독립된 도구 호출로 답한다. 이전 답변이나 사용자 원본 reference 텍스트가 verification prompt context에 같이 들어가지 않게 하라. 함께 들어가면 LLM이 hallucination을 그대로 복제한다 (논문 Section 3.3, Wikidata Joint Prec 0.29 vs Factored 0.32). (단 Phase 4 cross-check은 원문 대조가 목적인 별도 단계이므로 이 규칙의 예외다 — Phase 4 참조.)
- Tool-grounded only: PMID·저자·제목 등은 PubMed MCP 도구가 직접 반환한 값만 ground truth로 사용한다. 도구가 반환하지 않은 정보는 LLM이 "기억해서" 채워 넣지 말 것. 도구 실패 시
unverifiable로 표시한다. - Open question: yes/no verification 사용 금지. "Is the author X?" 대신 "Who are the authors?" (논문 Table 4: open 0.22 > yes/no 0.19).
- Atomic 분해: 한 reference에 대해 최대 6개 atomic question으로 분해. Shortform이 longform보다 정확 (논문 Section 4.3: 17% → 70%).
이 원칙을 어기면 본 skill의 검증 결과는 신뢰할 수 없다.
Workflow (Factor+Revise 4 phases)
본 skill은 CoVe 논문의 4단계를 reference 검증 도메인에 맞게 구체화한 것이다.
Phase 1 — Baseline Parsing (논문 Step 1: Generate Baseline)
사용자의 입력에서 reference 항목을 추출하고 atomic field로 분해한다.
경로 주의:
scripts/·references/는 이 SKILL.md가 있는skills/cove-reference-verifier/가 아니라 플러그인 루트에 있다. 스킬 실행 시 작업 디렉터리가 플러그인 루트라는 보장이 없으므로, 아래 예시처럼 플러그인 루트를 가리키는${CLAUDE_PLUGIN_ROOT}로 절대경로화해 호출한다. (해당 환경변수가 없는 수동 실행 시에는 플러그인 루트 경로로 치환.)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 275 lines · 237 tokens per session scan A e022f8cf0c81
cove-reference-verifier is a skill published in the GitHub repository JeonKH81/claude-cove-reference-verifier (5 stars, last pushed 1mo ago), licensed MIT. It adds 237 tokens to every session and 4,782 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
instrument-data-to-allotrope
Convert laboratory instrument output files (PDF, CSV, Excel, TXT) to Allotrope Simple Model (ASM) JSON format or flattened 2D CSV. Use this skill when scientists need to standardize instrument data for LIMS systems, data lakes, or downstream analysis. Supports auto-detection of instrument types. Outputs include full…
matlab
Build, review, migrate, and safely plan MATLAB or GNU Octave numerical workflows, including arrays, tabular/time data, tests, projects, graphics, MAT files, and explicit Python interoperability.
exploratory-data-analysis
Perform bounded, local exploratory analysis of explicitly supported scientific files. Use for redacted CSV/TSV/JSON profiles; optional NumPy, HDF5, FASTA/FASTQ, and basic image metadata inspection; missingness/leakage audits; outlier and transformation sensitivity; and rigorous EDA report scaffolds. Other domain…
phylogenetics
Build and analyze phylogenetic trees using MAFFT (multiple alignment), IQ-TREE 2 (maximum likelihood), and FastTree (fast NJ/ML). Visualize with ETE3 or FigTree. For evolutionary analysis, microbial genomics, viral phylodynamics, protein family analysis, and molecular clock studies.
research-engineer
An uncompromising Academic Research Engineer. Operates with absolute scientific rigor, objective criticism, and zero flair. Focuses on theoretical correctness, formal verification, and optimal implementation across any required technology.
mapping-to-snomed
Maps clinical concept spans extracted by OpenMed to SNOMED CT concepts through a USER-SUPPLIED terminology server (the user's own Ontoserver, Snowstorm, or UMLS/UTS), never a bundled vocabulary. Use when the user wants to code findings, disorders, procedures, body structures, or substances to SNOMED CT, run an ECL…