research-citation-check

Original：🇨🇳 Chinese

Translated

4 scripts

This skill is to be used when users explicitly request to 'verify/optimize in-text citations in the review `{topic}_review.tex`', 'run research-citation-check', or request to use the old name check-review-alignment skill. It uses the host AI's semantic understanding to verify each citation against the corresponding literature content one by one. Only when a fatal citation error is detected will the 'sentence containing the citation' be minimally rewritten. It prioritizes reusing the rendering script of `research-literature-review` to output PDF/Word, and allows fallback to the old name systematic-literature-review during the transition period. Core principle: Do not make changes for the sake of changing; when it is impossible to determine whether it is a fatal error, keep the original content and issue a warning in the report. ⚠️ Not applicable: When users only want to generate the main body of a systematic review (should use research-literature-review); when users only want to add/verify BibTeX entries (should use the dedicated bib management process).

2installs

Sourcehuangwb8/chineseresearchlatex

Added on2026-09-05

NPX Install

npx skill4agent add huangwb8/chineseresearchlatex research-citation-check

SKILL.md Content (Chinese)

View Translation Comparison →

Research Citation Check

Used to verify whether the in-text citations in the existing
```
{topic}_review.tex
```
are consistent with the corresponding paper content.
Only minimally rewrite the "sentence containing the citation" when a fatal error is confirmed.
Rendering PDF/Word prioritizes relying on
```
research-literature-review
```
; preparing structured input does not depend on this skill.
Compatible with prompt triggers using the old name
```
check-review-alignment
```
;
```
.check-review-alignment/
```
remains the stable historical workspace name.

Input

```
work_dir
```
: Contains
```
*_review.tex
```
and
```
.bib
```
files
Optional
```
--tex
```
: Specify the tex file name

Core Principles

Do not make changes for the sake of changing
Do not modify when uncertain
Only rewrite necessary sentences
Preserve all LaTeX command structures

Process

Input

Provide necessary input according to user requests and configuration files; if information is missing, list it clearly and stop the steps that depend on this input.

Execution Steps

When a bug caused by the design flaw of this skill occurs in the user's environment, prioritize using
```
bensz-collect-bugs
```
to record it in
```
~/.bensz-skills/bugs/
```
according to specifications. It is strictly prohibited to directly modify the source code of the skill installed in the user's local Claude Code / Codex.
If the AI can still complete the user's task via a workaround, record the bug first, then continue to complete the current task.
When the user explicitly requests actions such as "report bensz skills bugs", call the local
```
gh
```
and
```
bensz-collect-bugs
```
, and only upload new bugs to
```
huangwb8/bensz-bugs
```
; do not pull / clone the entire bug repository.

Dependency Check

Only enforce the check for
```
research-literature-review
```
when performing rendering; fallback to
```
systematic-literature-review
```
is allowed during the transition period
If only using
```
--prepare
```
, rendering dependencies are not required to be available

Pre-check and Localization

Locate
```
*_review.tex
```
and the corresponding
```
.bib
```
file
Stop immediately if any core file is missing

Structured Context Extraction

bash

cd /path/to/research-citation-check
python3 scripts/run_ai_alignment.py --work-dir "/path/to/work_dir" --prepare

Generate
```
ai_alignment_input.json
```
The input should include at least: sentences, bibkey, literature metadata, DOI/URL, PDF abstract segments or BibTeX abstract

AI Semantic Verification

Evidence priority: PDF abstract segments > BibTeX abstract/title > inference only from sentences

Each citation must be judged as one of the following:

```
fake_citation
```
```
wrong_citation
```
```
contradictory_citation
```
```
weak_support
```
```
overclaim
```
```
style_issue
```

Keep the original content when uncertain, and record it in Warnings

Report

The report must include at least:
- Summary
- Specific details
- Critical Fixes (P0)
- Warnings (P1)
- Rendering Result
The details of each citation must include at least: title, DOI, original sentence, actual content of the literature, rationality assessment, issue level

Rendering

bash

cd /path/to/research-citation-check
python3 scripts/run_ai_alignment.py --work-dir "/path/to/work_dir" --render

Output

{work_dir}/.check-review-alignment/ai_alignment_report.md

{work_dir}/.check-review-alignment/ai_alignment_input.json

Modified
```
{topic}_review.tex
```
Newly generated
```
{topic}_review.pdf
```
Newly generated
```
{topic}_review.docx
```

Output Management

Temporary files for new tasks of this Skill are uniformly written to

./.bensz-api/task-{yyyymmdd-hhmm}-{short-description}/{skill-name}/input|output|log/

. Reuse the same task root directory for the same task; create

shared/

only when multiple Skills collaborate. Formal deliverables are not written to this directory; historical hidden directories only allow explicit compatible reading, migration or cleanup.

Verification

Configuration is in
```
config.yaml
```
Script entry:
```
scripts/run_ai_alignment.py
```

Rendering dependency:

research-literature-review

(fallback:

systematic-literature-review

)

Failure and Recovery

Retain error evidence and completed products; retry from the most recent failed step only after input, environment or external dependencies are restored.

Constraints

P0: Must fix, allow minimal rewriting or correction of incorrect bibkey
P1: Only warn, no rewriting
P2: Skip entirely
Prohibited actions:
- Rewrite sentences that do not contain citations
- Rewrite entire paragraphs
- Introduce new bibkeys (unless fixing incorrect keys)
- Fabricate paper content

Common Hard Constraints

When tasks require persistent storage, use a unique
```
./.bensz-api/task-{yyyymmdd-hhmm}-{short-description}/
```
root directory; place shared materials in
```
shared/
```
, and Skill-specific materials in the Skill's
```
input/
```
,
```
output/
```
,
```
log/
```
.
Formal deliverables, source code and formal plans are saved according to project agreements and not written to the task workspace; do not overwrite, delete, migrate or remotely write without authorization.
For project maintenance changes, check BAC availability and record requirements, AI outputs, tool results, file changes and verification summaries; BAC only performs process audits and does not replace signature, responsibility or compliance judgments.
Do not record API Keys, access tokens, passwords, Cookies, environment/credential files, private Prompts, identity information, local usernames, hostnames or unnecessary large-volume raw data.
File paths must be standardized and restricted within authorized project scopes; external URLs, subprocesses and network access follow the principle of least privilege to prevent path traversal, SSRF and command injection.
Skill versions are uniquely recorded in its own
```
config.yaml:skill_info.version
```
; changes to public APIs, protocols, directories or configurations are synchronized with documents and
```
CHANGELOG.md
```
.
Only submit design flaws of the Skill or Bensz infrastructure itself to
```
bensz-collect-bugs
```
; first desensitize and write to
```
~/.bensz-skills/bugs/
```
, do not interrupt the current task, and only report publicly when explicitly requested by the user. It is strictly prohibited to directly modify the Skill source code installed by the user.

Skill-specific Constraints

Do not exceed the scope specified in this Skill description and the above process; do not disguise unverified information as definitive conclusions.

research-citation-check

NPX Install

Tags

SKILL.md Content (Chinese)

Research Citation Check

Input

Core Principles

Process

Input

Execution Steps

Dependency Check

Pre-check and Localization

Structured Context Extraction

AI Semantic Verification

Report

Rendering

Output

Output Management

Verification

Failure and Recovery

Constraints

Common Hard Constraints

Skill-specific Constraints