hypothesis-formation-variable-identification - SKILL.md Agent Skill

name: hypothesis-formation-variable-identification description: 'SOP: identify variables and their roles within a hypothesis' version: 1.0.0 category: hypothesis-formation type: sop campaign: hypothesis-formulation input: Mechanism description (from mechanism-extraction output) output: 'Variable list + role annotation (IV/DV/mediator/moderator/control) + operationalizability assessment' dependencies: skills: - subagent-spawning

Variable Identification

Identify all variables from a mechanism description and annotate their roles within the hypothesis structure.

HARD-GATE

Preconditions (all must hold before starting): 1. At least 1 causal mechanism chain has been provided 2. The mechanism chain contains identifiable variable names

Not met → stop, return error: mechanism description is insufficient to identify variables.

Pipeline

Precondition check: verify mechanism chain completeness
Variable extraction: enumerate all variables from every mechanism chain (including implicit variables)
Role classification: assign a role to each variable (IV/DV/mediator/moderator/control/confound)
Deduplication and merging: when the same variable appears in different mechanisms, merge it and annotate multiple roles
Operationalizability assessment: evaluate whether each variable can be measured (high/medium/low/unclear)
Output a structured variable list

Output Format

[
  {
    "name": "Variable name",
    "role": "IV | DV | mediator | moderator | control | confound",
    "description": "What this variable represents",
    "source_mechanism": ["mechanism name(s) where this variable appears"],
    "operationalizable": "high | medium | low | unclear",
    "operationalization_notes": "How it might be measured or manipulated"
  }
]

Identify at least 2 variables (1 IV + 1 DV).