Agent Fact Check Verify
A rigorous AI skill for multi-source information verification and credibility assessment.
Install & Use
Copy this prompt and send it to your AI assistant (Claude / Cursor / TRAE / Codex / WorkBuddy etc.) to auto-install:
Help me install this AI Skill: Agent Fact Check Verify. It is used for: A rigorous AI skill for multi-source information verification and credibility assessment. Full Skill content: https://321skill.com/skills/agent-fact-check-verify/raw/index.md Read that page and install it.
The prompt includes a link to the full Skill content. You can also view the full content.
This skill addresses the pain point of users struggling to quickly and objectively verify the authenticity of online information in the era of information overload. Whether it's rumors on social media, details in news reports, or data needed for work or study, users often face a mix of sources and difficulty distinguishing truth from falsehood. Manual verification is time-consuming, labor-intensive, and prone to subjective bias.
Usage is simple: you just need to make a request like "Verify/Check if this claim is true" to the AI assistant integrated with this skill. The AI will automatically deconstruct your query, integrate information from multiple channels such as government sources, official media, fact-checking websites, and social platforms, and analyze it based on internal rigorous scoring rules. It ultimately provides an integrated verification conclusion in a standardized four-part format (Correctness, Actual Situation, Conclusion, Relevant Links).
It is well-suited for content creators, academic researchers, media professionals, legal personnel, and any individual with high demands for information authenticity. It is especially useful for professionals who need to write reports, conduct market analysis, handle customer inquiries, or manage public opinion, and must ensure the accuracy of the information they rely on.
It is recommended to consistently use this tool when factual basis is needed for important decisions or public content. Note that it primarily verifies based on publicly available information and cannot cover non-public or paywalled content. Its conclusions are judgments based on current information and do not guarantee absolute correctness.
Key Features
Unlike ordinary search engines or single-source fact-checking tools, it enforces the integration of multi-source information from Tavily, Twitter, Reddit, etc. It employs a rigorous internal 100-point scoring system (not publicly disclosed) and a "Claim Core First" judgment logic, prioritizing the assessment of core facts to avoid misjudgment due to non-critical details. The output format is also strictly standardized.
Limitations
This skill relies on the availability and quotas of external APIs (e.g., Tavily, Twitter CLI, Reddit CLI), and its verification depth is limited by the accessibility of public information.
FAQ
Will this skill tell me the internal score?
No. According to its core principle, external responses do not reveal the scoring mechanism or specific scores, only providing the integrated conclusive judgment (Correct/Incorrect/Partially Correct/Insufficient Evidence).
Can it verify all types of information?
No. For opinions, satire, or predictive information, it does not make truthfulness rulings. For claims with insufficient publicly available information, it will judge as "Insufficient Evidence." In areas like public safety and medical risks, the judgment rules are more stringent.
Installation guide for AI assistants
If your AI coding assistant (Claude Code, Cursor, TRAE etc.) can see this page, send it this message to auto-install:
Visit https://321skill.com/skills/agent-fact-check-verify/raw/index.md to read the original Skill definition (Markdown format) for Agent Fact Check Verify, and install it according to the instructions.
Raw Markdown URL for AI: /skills/agent-fact-check-verify/raw/index.md