Skip to main content
Glama

cognitive-tools-mcp / gikendaasowin-aabajichiganan

by nbiish
PRD.md5.91 kB
# PRODUCT REQUIREMENTS DOCUMENT: GIKENDAASOWIN AABAJICHIGANAN MCP SERVER **Author:** ᓂᐲᔥ ᐙᐸᓂᒥᑮ-ᑭᓇᐙᐸᑭᓯ (Nbiish Waabanimikii-Kinawaabakizi) | **Date:** September 2, 2025 | **Version:** 8.9.6 ## 1. OBJECTIVE **Purpose:** Enable LLMs to leverage advanced cognitive deliberation frameworks for enhanced problem-solving through a structured 6-stage process that prompts LLMs to evaluate and select optimal cognitive techniques dynamically rather than providing pre-calculated solutions. ## 2. SCOPE **In-Scope Features:** - LLM-guided cognitive technique evaluation using 0.00-0.99 scoring system - 6-stage deliberation framework (Scientific Investigation → OOReD → Critical Thinking → Reviews → Action) - 15 modern prompting strategies for dynamic LLM selection - ≥1.53 threshold rule for strategy combination and selection - Iterative re-deliberation encouragement with tool usage estimation **Out-of-Scope:** - Hardcoded cognitive evaluations or pre-calculated strategy selections - Complex UI interfaces or visual components - Real-time collaborative deliberation features ## 3. USER EXPERIENCE **User Flow:** 1. LLM calls `deliberate(input: string, context?: string)` 2. Tool provides structured 6-stage cognitive framework 3. LLM evaluates 15 cognitive techniques with solution/efficiency ratings 4. LLM selects techniques scoring ≥1.53 for implementation 5. LLM receives guidance to return to deliberate after using recommended tools ## 4. STRUCTURE → 'camel' workflow → QAMMML (Quanta Atoms Molecule Matter Matter-Phase Lifeform) **Quanta:** Individual cognitive technique evaluations (0.00-0.99 scores) **Atoms:** Single-stage deliberation components (Scientific Investigation, OOReD, Critical Thinking, etc.) **Molecules:** 6-stage deliberation cycles combining multiple cognitive processes **Matter:** Complete deliberation framework with technique selection and tool recommendations **Matter Phases:** LLM-guided implementation cycles using selected cognitive techniques **Lifeforms:** Enhanced AI reasoning capabilities through systematic cognitive deliberation ## 5. FUNCTIONAL REQUIREMENTS ### Core Features - **LLM-Guided Evaluation:** Tool prompts LLM to evaluate cognitive techniques rather than providing hardcoded scores - **Dynamic Strategy Selection:** 15 modern prompting strategies available for LLM assessment - **Structured Framework:** 6-stage deliberation process with critical thinking questions - **Threshold-Based Selection:** ≥1.53 scoring rule for technique combination - **Tool Integration Planning:** Recommendations for tool usage and re-deliberation timing ## 6. NON-FUNCTIONAL REQUIREMENTS **Performance:** Fast framework delivery (<2s), scalable to multiple concurrent deliberations **Usability:** Single-parameter simplicity (input + optional context), clear structured prompts **Security:** No data persistence, stateless operation, input sanitization **Compatibility:** MCP protocol compliance, TypeScript/Node.js environment, npm package distribution ## 7. ASSUMPTIONS & CONSTRAINTS **Assumptions:** - LLMs can effectively evaluate cognitive techniques using numerical scoring - 0.00-0.99 scoring system provides sufficient granularity for technique selection - Users prefer LLM-guided evaluation over pre-calculated recommendations **Constraints:** - Technology: TypeScript, Node.js, MCP protocol specification - Architecture: Single-function tool interface with structured prompt output - Cognitive Framework: Must adhere to 6-stage deliberation process from new-mcp-flow.md ## 8. SUCCESS METRICS **Key Performance Indicators:** - Tool adoption rate: Target >80% preference over hardcoded approaches - LLM technique selection accuracy: Target >90% appropriate technique selection - Re-deliberation engagement: Target >60% users return to deliberate with tools as recommended - Problem-solving improvement: Target >40% better solution quality versus baseline ## 9. Knowledge Base **Knowledge Base:** - [new-mcp-flow.md](new-flow/new-mcp-flow.md): Complete flow specification and images - [modern-prompting.mdc](modern-prompting.mdc): 15 cognitive techniques for LLM evaluation - [latest.md](latest.md): Integration guidelines and implementation notes - [REFACTOR_SUMMARY.md](REFACTOR_SUMMARY.md): Migration from hardcoded to LLM-guided approach ## 10. ACCEPTANCE CRITERIA **Core Functionality:** - [ ] Tool prompts LLM for technique evaluation instead of providing pre-calculated scores - [ ] All 15 cognitive techniques from modern-prompting.mdc are presented for evaluation - [ ] 0.00-0.99 scoring system with ≥1.53 threshold rule implemented correctly - [ ] 6-stage deliberation framework follows new-mcp-flow.md specification exactly - [ ] Tool encourages iterative re-deliberation with estimated tool usage counts **Quality Standards:** - [ ] Performance meets sub-2-second response targets - [ ] No hardcoded cognitive evaluations remain in codebase - [ ] MCP protocol compliance verified through testing - [ ] TypeScript compilation successful with no errors ## 11. OPEN QUESTIONS - How to measure long-term cognitive improvement in LLM problem-solving? - Should we add analytics to track which cognitive techniques are most frequently selected? - What is the optimal tool usage count recommendation for complex problems? --- ## PRD BEST PRACTICES CHECKLIST - [x] Use clear, unambiguous language - [x] Include specific, measurable requirements - [x] Define success criteria objectively - [x] Balance detail with conciseness - [x] Treat PRD and Knowledge Base as living documents - [x] Reference PRD and Knowledge Base throughout development lifecycle --- *This PRD documents the successful implementation of LLM-guided cognitive deliberation as specified in new-mcp-flow.md. The tool now functions as a cognitive framework enhancer rather than a prescriptive system.*

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/nbiish/gikendaasowin-aabajichiganan-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server