ChineseResearchLaTeX is a collection of LaTeX templates and an AI-assisted workflow for preparing Chinese research documents such as grant proposals, papers, theses, and academic CVs. Researchers use it to plan, format, review, compile, and revise these documents with human oversight. The catalogue skills and instructions support its agent-based research-writing workflow.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add huangwb8/ChineseResearchLaTeX --skill research-literature-interpretationgit clone --depth 1 https://github.com/huangwb8/ChineseResearchLaTeXWrote 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/huangwb8/chineseresearchlatex/research-literature-interpretation)<a href="https://agentmods.dev/skills/huangwb8/chineseresearchlatex/research-literature-interpretation"><img src="https://agentmods.dev/badge/skills/huangwb8/chineseresearchlatex/research-literature-interpretation/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/huangwb8/chineseresearchlatex/research-literature-interpretation"><img src="https://agentmods.dev/badge/skills/huangwb8/chineseresearchlatex/research-literature-interpretation.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00089 | $0.02572 |
| Opus 5.5 | $0.00036 | $0.01029 |
| Sonnet 5.5 | $0.00018 | $0.00514 |
| Haiku 4.5 | $0.00009 | $0.00257 |
Grade A, and why
research-literature-interpretation 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 1mo 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 — 138 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Research Literature Interpretation
把单篇论文压缩成读者能复述、质疑和迁移的解释链：旧方法为何受限，作者改变了什么，机制为何可能有效，证据支持到何种强度，代价和失效条件是什么。
主动取舍比篇幅重要。页码、章节和图表只服务复核，不支配叙事。
范围与文件边界
- 仅处理单篇论文；发现、筛选和去重交给
research-literature-radar，多论文综合交给综述技能。 - 确认标题、作者、年份、venue、版本、DOI/arXiv ID、一手 URL、访问日期和可读范围。版本冲突并列记录，不擅自裁决。
- 优先阅读摘要、引言、结论、方法、关键图表/图注、消融、失败案例、附录和代码说明，再深挖影响结论的部分。只有摘要或图表时，交付证据地图并收缩结论。
- 只用合法可访问的一手材料和可靠跟进来源；不绕过访问限制，不执行未知代码。区分静态阅读、实际运行和作者自报结果。
- 默认写入
docs/papers/<friendly-id>/<friendly-id>.md。已有笔记默认修订或追加，不覆盖用户内容；不改raw/或.bensz-api/research-literature-radar/catalog.jsonl。中间材料放本轮.bensz-api/。
流程
输入
按用户请求和配置文件提供必要输入；缺失信息应明确列出并停止依赖该输入的步骤。
执行步骤
- 仅将本 Skill 的设计缺陷（流程漏判、输入契约不完整或环境假设错误）视为可上报 bug；用户数据错误、第三方服务抖动、用户主动改源码和模型偶发波动不属于此范围。
- 发现设计缺陷时先脱敏记录到
~/.bensz-skills/bugs/，当前任务继续；只有用户明确要求公开上报时，才使用本机gh api直传，不 clone 仓库。 - 不收集用户名、主机名、工作目录、密钥、令牌、Cookie 或其它无关隐私；不得直接修改用户本地已安装 Skill 的源代码来“顺手修 bug”。
版本唯一来源为同目录 config.yaml:skill_info.version；本文件只描述稳定工作契约，不重复易变配置。
确定读者要做的判断
明确读者水平、目的（理解、选型、复现、批判或迁移）和论文类型。从标题、摘要、引言、结论、核心图表与相关工作提出主线，再回查方法；不要逐节抄目录。
提炼不超过三个核心命题
只保留“若为真会改变判断”的论文特有命题。内部按以下链条核查，正文不机械显示字段：
Claim → Why → Mechanism → Evidence → Alternative → Boundary → Verdict
链条缺口意味着继续取证或缩小结论，不能用背景填补。
重建机制
按“输入/状态/输出 → 信息如何保留、丢弃、读取 → 计算如何实现”解释。公式或证明只保留关键前提、构造、结论和直觉；细节放可选核查层。类比、事后重建须标为“教学类比”“我的解释”或“基于文本的重建”。
按论文类型调整镜头：
- 方法/系统：瓶颈、表示/协议改变、端到端收益；拆分模块、规模、训练配方、实现和部署成本。
- 理论/证明：前提、关键构造、结论；检查依赖、适用域、反例和最脆弱假设。
- 实证/因果：比较对象、识别策略、效应量与不确定性；检查混杂、功效、数据和外推。
- 分析/数据集：测量对象、数据/标注假设、发现；检查泄漏、代表性和指标有效性。
- 综述性单篇：范围、组织原则、综合结论；检查纳入标准、遗漏和反方证据。
建立证据链并压力测试
优先针对性任务/定理、公平对照、辨识力强的消融和论文内失败结果。数字只有改变判断时才保留，并附指标、比较对象、规模、预算、硬件或误差；未报告就明说。
每个命题都要有最近的替代解释和削弱条件。设计最小反事实：去关键部件、匹配参数/数据/计算、换分布/评价或匹配实现效率。不得把相关性、单一 benchmark、作者自述或混合系统收益升级为因果机制。
写成一个分层版本
交付笔记必须以 interpretation_contract: research-literature-interpretation-evidence-v1 开头，并在 frontmatter 保存 source_id、evidence_depth（normative/fulltext/abstract/web/source）、read_scope、stable_citation、question、method、key_results、limitations 和 supports_or_refutes。这些字段是证据索引，不替代正文；fulltext 不得与仅摘要读取范围同时声明。旧笔记只能用 --allow-legacy 做只读检查，不能作为新运行的完整解读证据。
What ships with it
8 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago Changed · +4 lines ee05f58b7200
- 1mo ago Changed · +24 lines c71a4dcea1ad
- 1mo ago First seen · 110 lines · 89 tokens per session scan A 1276a31fc881
research-literature-interpretation is a skill published in the GitHub repository huangwb8/ChineseResearchLaTeX (2,980 stars, last pushed 10d ago), licensed MIT. It adds 89 tokens to every session and 2,572 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-05.
Other skills, from other repositories
semantic-scholar-deep
Deep research over the Semantic Scholar Graph API. Covers endpoints missing from allenai's lookup skill — paper references (backward citations), recommendations, batch paper lookup (up to 500 IDs), snippet search, and multi-hop citation graph traversal (BFS forward/backward). Use when the user asks to build a citation…
adaptyv
Cloud laboratory platform for automated protein testing and validation. Use when designing proteins and needing experimental validation including binding assays, expression testing, thermostability measurements, enzyme activity assays, or protein sequence optimization. Also use for submitting experiments via API…
openevidence
Query OpenEvidence (clinical evidence assistant) without an MCP server. Closely mirrors the openevidence-mcp tool surface - check auth, list question history, fetch an article, or ask a new question (with optional follow-up thread). Use when the user asks medical / evidence-based clinical questions, references…
thesis-control
Use when AI-assisted thesis or manuscript edits risk claim drift, scope creep, loss of intended use, experiment-role promotion, or repeated revisions that fail to converge; provides author-intent control, lightweight or strict contracts, drift audits, revision escalation, and human gates.
manuscript-reframe
Reframe report-like academic drafts into paper-form scientific arguments while preserving or explicitly renegotiating author intent; requires an approved old-versus-proposed spine, evidence and argument baselines, analysis-role control, and post-edit drift review.
onescience-primitives
A resource-search tool for finding reusable OneScience building blocks. It searches a catalogue for the most relevant scientific workflow components and returns their supporting details and usage contracts.