Anti-Bullshit MCP Server

Instructions

Implementation Reference

• src/index.ts:106-173 (handler)

  The main execution logic for the 'analyze_claim' tool. It determines the validation framework, performs validation using helper functions, generates framework-specific cross-reference prompts, and formats the response content.

  if (name === "analyze_claim") {
    const framework: keyof typeof VALIDATION_FRAMEWORKS = 
      (typeof args.framework === 'string' && args.framework in VALIDATION_FRAMEWORKS) 
        ? args.framework as keyof typeof VALIDATION_FRAMEWORKS 
        : VALIDATION_FRAMEWORK;
    const validation = validateWithFramework(args.text, framework, {
      hasEmpirical: /evidence|study|research|data/i.test(args.text),
      servesWellbeing: /benefit|improve|help|support/i.test(args.text),
      maintainsHarmony: /balance|harmony|integrate/i.test(args.text)
    });
  
    const suggestions = getValidationSuggestions(args.text, framework);
    
    // Generate cross-referencing prompts
    const crossRefPrompts = [
      `- Use Exa MCP server to search for general information: "${args.text}"`,
      `- Use Brave Search for independent web sources: "${args.text}"`,
      `- Search ArXiv for preprints and technical papers: "${args.text}"`,
      `- Use Google Scholar MCP server to find peer-reviewed research: "${args.text}"`,
      `- Cross-reference findings between academic and general sources to identify consensus or conflicts`
    ];
  
    // Framework-specific cross-references
    if (framework === "empirical" || framework === "pluralistic") {
      crossRefPrompts.push(
        `- Compare methodologies between ArXiv papers and peer-reviewed research`,
        `- Analyze replication status across different studies`,
        `- Cross-validate findings between academic databases`
      );
    }
  
    if (framework === "responsible" || framework === "pluralistic") {
      crossRefPrompts.push(
        `- Use Exa MCP server to search for community impact studies: "${args.text}"`,
        `- Cross-reference academic findings with community experiences`,
        `- Compare traditional knowledge with modern research findings`
      );
    }
  
    if (framework === "harmonic" || framework === "pluralistic") {
      crossRefPrompts.push(
        `- Use Exa MCP server to search for alternative perspectives: "${args.text}"`,
        `- Compare Eastern and Western research approaches`,
        `- Synthesize findings across different knowledge systems`
      );
    }
  
    return {
      content: [
        {
          type: "text",
          text: `Analysis using ${framework} framework:\n\n` +
               `Requirements:\n${suggestions.join("\n")}\n\n` +
               `Confidence level: ${validation.confidence}\n\n` +
               `Suggested cross-references:\n${crossRefPrompts.join("\n")}`
        },
        {
          type: "text",
          text: JSON.stringify({
            framework,
            validation,
            suggestions,
            crossRefPrompts
          })
        }
      ],
    };
  }

• src/index.ts:35-53 (schema)

  Input schema definition for the 'analyze_claim' tool, specifying the required 'text' parameter and optional 'framework' with allowed values.

  {
    name: "analyze_claim",
    description: "Analyze a claim using multiple epistemological frameworks and suggest validation steps",
    inputSchema: {
      type: "object",
      properties: {
        text: {
          type: "string",
          description: "Claim to analyze",
        },
        framework: {
          type: "string",
          description: "Validation framework to use (empirical, responsible, harmonic, or pluralistic)",
          enum: ["empirical", "responsible", "harmonic", "pluralistic"],
        }
      },
      required: ["text"],
    },
  },

• src/index.ts:19-23 (registration)

  Server capabilities registration declaring 'analyze_claim' as an available tool.

  tools: {
    analyze_claim: true,
    validate_sources: true,
    check_manipulation: true
  },

• src/validation.ts:81-91 (helper)

  Helper function that applies the selected framework's validation logic to compute requirements and confidence level, called by the analyze_claim handler.

  export function validateWithFramework(
    claim: string,
    framework: keyof typeof VALIDATION_FRAMEWORKS,
    evidence: any
  ) {
    const validator = VALIDATION_FRAMEWORKS[framework];
    return {
      requirements: validator.validateClaim(claim),
      confidence: validator.confidenceLevel(evidence)
    };
  }

• src/validation.ts:93-99 (helper)

  Helper function that generates specific validation step suggestions from the framework's requirements, used in the analyze_claim response.

  export function getValidationSuggestions(
    claim: string,
    framework: keyof typeof VALIDATION_FRAMEWORKS
  ): string[] {
    const requirements = VALIDATION_FRAMEWORKS[framework].validateClaim(claim).requirements;
    return requirements.map(req => `- Verify if claim "${claim}" meets requirement: ${req}`);
  }