🧩

Reasoning / Planning Paradigm

NIST AI RMFOWASP KC3AISVS C7NIST 800-53

Paradigms that enable AI agents to solve complex problems by breaking down tasks, making decisions, and forming plans. (NIST: Map, Measure; OWASP: KC3; AISVS: C7)

Subcomponents
  • Structured Planning / Execution
    Stepwise or chain-of-thought reasoning.
    →
  • ReAct (Reason + Act)
    Alternating between reasoning and action.
    →
Security Implications
Reasoning paradigms can be manipulated for misaligned behaviors or goal manipulation. Validate reasoning chains and outputs.
Implementation Considerations
Apply multi-stage validation, fact-checking, and output filtering. Reference: OWASP KC3, AISVS C7.
Related Components
🧠Language Models (LLMs)🎛️Orchestration (Control Flow)
NIST 800-53 Rev 5 Controls
AC-2AC-3SI-4SI-7AU-6
🗂️ Framework References
NIST AI Risk Management Framework
NIST AI RMF Official DocumentationNIST AI RMF 1.0 (PDF)
OWASP AISVS
Interactive AISVS ToolOWASP AISVS GitHub Repository