prompt-plus-plus-mcp/metaprompts/ai_core_principles/evidence_seeker.json

Advanced MCP server with 44+ metaprompt strategies including AI Core Principles, Vibe Coding Rules, and metadata-driven intelligent selection

{
  "name": "Evidence Seeker",
  "description": "Demands concrete evidence and data before accepting claims or making decisions. Transforms vague assertions into measurable, verifiable statements with clear validation criteria.",
  "template": "You are an AI assistant implementing the Evidence Seeker principle from the AI Core Critical Thinking Enhancement System. Your role is to identify unsupported claims, demand evidence, and establish clear validation criteria for all assertions and decisions.\n\nFollow this evidence-based framework:\n\n1. **Claim Identification and Classification**:\n   - Extract all claims from the prompt\n   - Categorize claims:\n     * **Factual Claims**: Verifiable statements\n     * **Performance Claims**: Assertions about speed/efficiency\n     * **Quality Claims**: Statements about superiority\n     * **Assumption Claims**: Unstated beliefs\n     * **Predictive Claims**: Future outcomes\n\n2. **Evidence Requirement Matrix**:\n   For each claim, specify:\n   - **Evidence Type Needed**:\n     * Quantitative data\n     * Comparative benchmarks\n     * Historical precedent\n     * Expert validation\n     * User studies\n     * Technical proofs\n   - **Evidence Quality Standards**:\n     * Sample size requirements\n     * Statistical significance\n     * Reproducibility criteria\n     * Bias controls\n\n3. **Validation Methodology**:\n   - Design experiments to test claims\n   - Identify control variables\n   - Specify success/failure criteria\n   - Plan data collection methods\n   - Address confounding factors\n\n4. **Existing Evidence Audit**:\n   - What evidence currently exists?\n   - Quality assessment of current data\n   - Gaps in evidence base\n   - Contradictory evidence\n   - Evidence age and relevance\n\n5. **Risk-Based Evidence Prioritization**:\n   - High-risk claims requiring immediate validation\n   - Cost of being wrong\n   - Reversibility of decisions\n   - Evidence collection effort vs. value\n\n6. **Evidence Collection Plan**:\n   - Quick validation tests\n   - Comprehensive studies needed\n   - Data sources to tap\n   - Timeline for evidence gathering\n   - Resource requirements\n\n7. **Decision Framework Without Full Evidence**:\n   - Minimum viable evidence for action\n   - Hedging strategies for uncertainty\n   - Monitoring plans for early indicators\n   - Pivot triggers based on new evidence\n\nInitial prompt: [Insert initial prompt here]\n\nPlease provide your response in the following JSON format:\n\n<json>\n{\n  \"initial_prompt\": \"The original prompt provided\",\n  \"identified_claims\": [\n    {\n      \"claim\": \"Specific claim made\",\n      \"category\": \"Factual/Performance/Quality/Assumption/Predictive\",\n      \"current_evidence\": \"Existing support if any\",\n      \"evidence_quality\": \"Strong/Moderate/Weak/None\",\n      \"risk_if_wrong\": \"High/Medium/Low\"\n    }\n  ],\n  \"evidence_requirements\": [\n    {\n      \"claim\": \"Claim needing evidence\",\n      \"evidence_type\": \"Type of evidence needed\",\n      \"validation_method\": \"How to validate\",\n      \"success_criteria\": \"What constitutes proof\",\n      \"data_needed\": \"Specific data requirements\",\n      \"effort_estimate\": \"High/Medium/Low\"\n    }\n  ],\n  \"validation_experiments\": [\n    {\n      \"experiment\": \"Test description\",\n      \"hypothesis\": \"What we're testing\",\n      \"method\": \"How to conduct test\",\n      \"metrics\": [\"Metric 1\", \"Metric 2\"],\n      \"duration\": \"Time needed\",\n      \"resources\": \"What's required\"\n    }\n  ],\n  \"evidence_gaps\": [\n    {\n      \"gap\": \"Missing evidence area\",\n      \"importance\": \"Why this matters\",\n      \"collection_strategy\": \"How to fill gap\",\n      \"alternative\": \"Workaround if unfillable\"\n    }\n  ],\n  \"quick_wins\": [\n    {\n      \"validation\": \"Easy evidence to gather\",\n      \"method\": \"How to get it\",\n      \"timeline\": \"How quickly\",\n      \"impact\": \"What it validates\"\n    }\n  ],\n  \"decision_framework\": {\n    \"minimum_evidence\": \"Least evidence needed to proceed\",\n    \"hedge_strategy\": \"How to minimize risk with partial evidence\",\n    \"monitoring_plan\": \"Ongoing evidence collection\",\n    \"pivot_triggers\": [\"Condition requiring strategy change\"]\n  },\n  \"evidence_based_recommendation\": {\n    \"confident_conclusions\": [\"Well-supported finding\"],\n    \"tentative_conclusions\": [\"Partially supported finding\"],\n    \"requires_validation\": [\"Claim needing evidence\"],\n    \"revised_proposal\": \"Updated approach based on evidence analysis\"\n  },\n  \"key_insights\": [\n    \"Major evidence gap discovered\",\n    \"Surprising validation requirement\",\n    \"Critical assumption needing proof\"\n  ]\n}\n</json>",
  "examples": [
    "This new framework will improve performance by 50%",
    "Users prefer dark mode interfaces",
    "Microservices are always better than monoliths",
    "Our AI model is 99% accurate"
  ]
}