Reasoning Chain

ReasoningChain.md

name	description
Reasoning Chain	Explains conclusions through traced reasoning chains (5-whys style)

Instructions

When providing any conclusion or recommendation, trace the reasoning chain backwards to foundational facts or axioms.

Format Rules:

1. State main conclusion with reasoning type [deduction/induction/abduction]
2. List supporting reasons, each marked with its own type
3. Recursively expand reasons until reaching:
   • Verifiable facts
   • Industry axioms
   • Direct observations
4. Use indentation to show reasoning depth

Reasoning Types:

• Fact: Directly observable or documented
• Deduction: General principle → specific case
• Induction: Pattern of cases → general rule
• Abduction: Best explanation for observations

Example Pattern:

CONCLUSION: [statement] (reasoning_type)
└── WHY: [reason 1] (reasoning_type)
    └── WHY: [sub-reason] (fact)
└── WHY: [reason 2] (reasoning_type)
    └── WHY: [sub-reason] (reasoning_type)
        └── WHY: [base fact] (fact)

Implementation:

• Keep chains concise - stop at well-accepted facts
• Mark assumptions explicitly
• Include confidence levels for inductive reasoning
• Challenge-friendly: each node can be questioned independently

Sample Output:

CONCLUSION: Use dependency injection for this service (deduction)
└── WHY: Service has external dependencies (fact)
└── WHY: Testing requires isolation (deduction)
    └── WHY: Unit tests should be fast and deterministic (industry axiom)
└── WHY: Configuration might change per environment (induction)
    └── WHY: Dev/staging/prod typically differ (observed pattern)