research-claim-checker

What This Skill Does

研究论断校验器 (Research Claim Checker) is a precision-oriented OpenClaw AI agent skill designed to audit research papers, analytical reports, and argumentative essays. It serves as an internal quality control mechanism, systematically analyzing the relationship between stated conclusions and the provided evidence. By decomposing complex documents into structured components, the skill identifies logical fallacies, missing data points, and gaps in evidence chains. It ensures that every claim made in a report is backed by empirical data or logical derivation, providing a clear credibility score and actionable suggestions for improvement.

Installation

To integrate this skill into your OpenClaw environment, execute the following command in your terminal: clawhub install openclaw/skills/skills/52yuanchangxing/research-claim-checker

Ensure your local environment includes the standard resource directory structure as defined in the source repository openclaw/skills, specifically the spec.json and template.md files, to enable structured output formatting.

Use Cases

• Academic & Professional Audits: Verifying the soundness of findings in research papers or market analysis reports before final submission.
• Evidence Chain Analysis: Identifying logical leaps or missing references that could weaken a business proposal or technical argument.
• Drafting & Reviewing: Utilizing the skill as a 'red team' partner to identify potential points of attack in a draft, allowing for proactive strengthening of arguments.
• Evidence Gap Detection: Generating a structured 'To-Do' list for additional research needed to substantiate bold or controversial claims.

Example Prompts

1. "请检查附件中的市场调研报告，提取核心结论并对照文中的证据链，列出所有证据不足的地方。"
2. "我写了一篇关于AI模型架构的分析稿，请使用研究论断校验器帮我识别逻辑断点，并给出提高可信度的改写建议。"
3. "请根据这篇研究摘要，输出一份结构化的论断校验报告，包括可信度判断和需要补充的证据清单。"

Tips & Limitations

• Not an Authority: This skill acts as a reasoning tool, not a final arbiter of objective truth. Always perform human verification for critical claims.
• Input Hygiene: For best results, provide clean, structured input. If the text is highly ambiguous or filled with jargon without definition, the skill may flag false positives regarding 'evidence gaps'.
• Avoid Misuse: Do not use this tool to fabricate citations or to replace formal peer review processes. It is a secondary verification aid.
• Workflow: Always start by letting the skill generate a 'Draft for Review' (可审阅草案) before requesting a final actionable list to ensure the interpretation of your input is aligned with your intent.