cognitive-tools-mcp / gikendaasowin-aabajichiganan

# This image is our total flow ->  ___ ## STAGE 1  - Scientific Investigation follows a series of logic and orderly steps to formulate hypotheses. - 1. Identify question - 2. Form hypothesis - 3. Conduct experiment - 4. Analyze data - 5. Draw conclusions ___ ## STAGE 2  - O.O.Re.D. - Observe - Observe the input, needs, and situation. - Orient - Orient to the solution. - Reason - Why is the proposed solution the correct answer? What are the facts? - Decide - Decide which tools and actions are needed to implement the solution. ___ ## STAGE 3  - Steps in Critical Thinking: - 1. What is the purpose of my thinking? - 2. What precise question am I trying to answer? - 3. Within what context or framework am I operating? - 4. What information do I have and need to gather? - 5. How reliable and credible is this information? - 6. What concepts, algorithms, and facts are relevant to my thinking? - 7. What conclusions can I draw from this information? - 8. What am I taking for granted; what assumptions am I making? - 9. If I accept conclusions, what are the implications? - 10. What would be the consequences if I put this solution into action? - pre-Act action - prompting strategies - list tools we need upon 'Act' stage. - websearch - mcp servers - file and code tools - user feedback - knowledge base ___ ## STAGE 4  - Scientific Investigation follows a series of logic and orderly steps to formulate hypotheses. - 1. Identify question - 2. Form hypothesis - 3. Conduct experiment - 4. Analyze data - 5. Draw conclusions ___ ## STAGE 5  - O.O.Re.D. - Observe - Observe the input, needs, and situation. - Orient - Orient to the solution. - Reason - Why is the proposed solution the correct answer? What are the facts? - Decide - Decide which tools and actions are needed to implement the solution. ___ ## STAGE 6  - Fact based Action ___ ## TOOLING LOGIC ### INPUT - single tool call: `deliberate(input: string, context?: string)` #### MCP TOOL DOES THE FOLLOWING - The 'Orient stage': presents LLM calling the mcp tool each of the prompting strategies from [modern-prompting.mdc](modern-prompting.mdc) within and asked to apply these to the 'input' -> LLM determines which prompting strategy would provide the best solution the most efficiently by assigning every strategy a ```solution level: {0.00-0.99}``` and ```efficiency level: {0.00-0.99}``` rating which will be summed together to determine which tool to use based on which tools ≧1.53 - If more than one tool should come to the summation of ≥1.53 -> use a combination of the strategies - LLM is prompted with Stages 1-5 - Ask the LLM what tools are needed to accomplish the task? - Stage 6 is formatted output of mcp tool ___ ### OUTPUT The final output should be formatted markdown that follows this strucure: ___ ```DELIBERATION: [[though process through stages]] + [[selected cognitive technique/s output]]``` ```SELECTED TOOLS: [[list of available tools to accomplish task]]``` ___ ### ACCOMPLISHES - Single-shot self-prompt of cognitive strategies to for 'thinking' techniques to accomplish the task at hand. - Enhanced ability to adapt to varying input and task requirements. - Thorough evaluation of potential solutions and tools to required for the best course of action. - Additional LLM attention of all tools available for task. - Allows LLMs to take more cognitive time to ```deliberate``` the best strategy and tools for any task. ___ - (the 0.00 - 1.00+ system is CRITICAL) → (no percentages or ‘metrics’)