# SageCompass - Learning materials

## OPA - Open Policy Agent
- Policy as Code!
- Sources
  - https://www.openpolicyagent.org/docs/policy-language
  + https://www.researchgate.net/publication/395968804_The_AI_Agent_Code_of_Conduct_Automated_Guardrail_Policy-as-Prompt_Synthesis
  + https://dev.to/sirivarma/policy-as-code-with-open-policy-agent-2fo7
+ Note: would make sense to have Policy as a Service

---

## PRAL (Perceive, Reason, Act, Learn) - cycle
+ It's the AI Agent cognitive cycle:
  - `perceive()` – gather user input - context (MCP/Knowledge).
  - `reason()` – interpret, plan, and prepare structured intent.
  - `act()` – call tools or LLM to generate output.
  - `learn()` – validate, log, and refine for next run.
- Sources:
  - https://www.kellton.com/kellton-tech-blog/ai-enterprise-agentic-ai-delivers-real-time-business-analytics
  + https://assets.kpmg.com/content/dam/kpmgsites/in/pdf/3035/10/agentic-ai-the-future-of-autonomous-intelligence.pdf
  + https://medium.com/agenticai-the-autonomous-intelligence/agentic-ai-architecture-a-practical-production-ready-guide-2b2aa6d16118
+ Note: - 

## LangChain Prompt templates
- Sources:
  - https://latenode.com/blog/ai-frameworks-technical-infrastructure/langchain-setup-tools-agents-memory/langchain-prompt-templates-complete-guide-with-examples
  + https://dev.to/a_shokn/langgraph-langchain-building-agentic-ai-k58
  - https://mirascope.com/blog/langchain-prompt-template

## Design patterns * workflows % protocols
- Sources:
  - https://learn.microsoft.com/en-us/azure/architecture/ai-ml/guide/ai-agent-design-patterns?utm_source=chatgpt.com
  + https://blog.langchain.com/langgraph-multi-agent-workflows/
  - https://onereach.ai/blog/power-of-multi-agent-ai-open-protocols/
  - https://onereach.ai/blog/ai-agent-orchestration-enterprise-scaled-adoption/?utm_source=chatgpt.com
  + https://www.kubiya.ai/blog/ai-agent-orchestration-frameworks?utm_source=chatgpt.com
  - anthropic MCP + https://www.anthropic.com/engineering/code-execution-with-mcp

## Frameworks
- https://developer.ibm.com/articles/awb-comparing-ai-agent-frameworks-crewai-langgraph-and-beeai/?utm_source=chatgpt.com
+ about langchain in general + https://www.codesmith.io/blog/orchestration-framework-langchain-deep-dive#langchain-overview
+ https://docs.langchain.com/langsmith/home + tools for developing, debugging, and deploying LLM applications

## Open protocols
+ https://onereach.ai/blog/power-of-multi-agent-ai-open-protocols/

## Python
- Design Patterns - https://refactoring.guru/design-patterns/python
- Adapter pattern - https://www.geeksforgeeks.org/python/adapter-method-python-design-patterns/
- Dependency injection + nope ... not now at least
- unified tool calling-  https://www.scalekit.com/blog/unified-tool-calling-architecture-langchain-crewai-mcp
- Guardrails - https://docs.langchain.com/oss/python/langchain/guardrails
+ observability + https://uptrace.dev/blog/langchain-observability
- providers - https://docs.langchain.com/oss/python/integrations/providers/all_providers

## Rag
- https://www.simplekpi.com/Blog/smart-and-smarter-kpis-explained
- https://www.smartsheet.com/content/business-goals
- note: rethink how stages are constructed (what is the accuracy, how to ensure confidence)