CtrlK
BlogDocsLog inGet started
Tessl Logo

research-citation-check

当用户明确要求"核查/优化综述 `{主题}_review.tex` 的正文引用"、"运行 research-citation-check",或要求使用旧名 check-review-alignment skill 时使用。通过宿主 AI 的语义理解逐条核查引用是否与文献内容吻合,只在发现致命性引用错误时对"包含引用的句子"做最小化改写,并优先复用 `research-literature-review` 的渲染脚本输出 PDF/Word,过渡期允许 fallback 到旧名 systematic-literature-review。核心原则:不为了改而改,无法确定是否为致命性错误时保留原样并在报告中警告。⚠️ 不适用:用户只是想生成系统综述正文(应使用 research-literature-review);用户只是想新增/核对 BibTeX 条目(应使用专门的 bib 管理流程)。

67

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

65%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-organized, actionable skill body with a clear sequenced workflow and real executable scripts, weakened by minor redundancy, a missing post-edit verification feedback loop, and a referenced config.yaml that is not present in the bundle.

Suggestions

Add an explicit post-edit verification feedback loop in the workflow (e.g., after rewriting a cited sentence, re-render or re-check that the citation still compiles; revert if it breaks) to satisfy the destructive-change checkpoint expectation.

Provide the missing config.yaml referenced in '参考与验证' and loaded by run_ai_alignment.py, or remove the reference if configuration is optional, so navigated paths all resolve.

Trim redundancy between '修改边界' and the closing '核心原则' section, and consider whether the bensz-collect-bugs collaboration block belongs in this skill's core body.

DimensionReasoningScore

Conciseness

The body is mostly lean (bullet lists, no explanations of what LaTeX/BibTeX/PDF are), but the final '核心原则' section restates the 修改边界 levels and the description's core principle, and the bensz-collect-bugs collaboration block is tangential boilerplate — so it could be tightened rather than every token earning its place.

2 / 3

Actionability

Provides executable commands referencing the real script ('python3 scripts/run_ai_alignment.py --work-dir ... --prepare' / '--render'), concretely specifies the JSON input fields (句子、bibkey、DOI/URL、摘要), the citation category taxonomy (fake_citation, wrong_citation, ...), and the report section layout — copy-paste ready guidance.

3 / 3

Workflow Clarity

The six-step workflow is clearly sequenced with guards ('缺任何核心文件时立即停止', '无法确认时保持原样'), but the destructive sentence-rewriting lacks an explicit validate→fix→retry feedback loop after edits, which the scoring notes say should cap destructive-change workflows at 2.

2 / 3

Progressive Disclosure

Structure is well sectioned and the script entry point is signaled one level deep, but the body references 'config.yaml' (and the script loads it) while that file is absent from the bundle — a dangling reference that breaks clean navigation against the actual bundle structure.

2 / 3

Total

9

/

12

Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, well-crafted description that covers concrete actions, natural trigger terms, explicit use-when guidance, and clear disambiguation against neighboring skills. It is dense but free of fluff or over-claims.

DimensionReasoningScore

Specificity

Lists multiple concrete actions: '逐条核查引用是否与文献内容吻合', '对包含引用的句子做最小化改写', and '复用 research-literature-review 的渲染脚本输出 PDF/Word' — matching the anchor for multiple specific concrete actions rather than the vague single-action anchors.

3 / 3

Completeness

Explicitly answers both: WHAT (semantic citation checking with minimal rewriting and PDF/Word rendering) and WHEN ('当用户明确要求...时使用'), with an explicit '⚠️ 不适用' clause covering the negative cases — well above the cap-at-2 threshold for a missing Use-when clause.

3 / 3

Trigger Term Quality

Includes natural phrases a user would actually say ('核查/优化综述 {主题}_review.tex 的正文引用', '运行 research-citation-check') plus the legacy name 'check-review-alignment', giving good coverage of natural invocation terms.

3 / 3

Distinctiveness Conflict Risk

The explicit '⚠️ 不适用' disambiguation (generate review body → research-literature-review; BibTeX entry management → dedicated bib flow) carves out a clear niche unlikely to trigger the wrong skill.

3 / 3

Total

12

/

12

Passed

Validation

87%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

Total

14

/

16

Passed

Repository
huangwb8/ChineseResearchLaTeX
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.