Reasoning Levels

Reasoning levels control how deeply the AI analyzes code.

Levels

Level	Description	Best For
thorough	Deep analysis with extended thinking	Code reviews (default)
standard	Balanced analysis	Refine command (default)
fast	Quick responses	Rapid feedback

Setting Reasoning Levels

Per-Command (CLI Flags)

roborev review --reasoning fast      # Quick review
roborev refine --reasoning thorough  # Careful fixes
roborev prompt --reasoning standard "Explain this code"

Per-Repository

.roborev.toml

review_reasoning = "thorough"  # For code reviews (default: thorough)
refine_reasoning = "standard"  # For refine command (default: standard)

Defaults

• Reviews: Default to thorough for comprehensive analysis
• Refine: Defaults to standard for balanced fix generation

Agent Support

The reasoning level maps to agent-specific features:

Agent	Implementation
Codex	model_reasoning_effort (high/medium/low)
Claude Code	Not yet applied (value stored only)
Other agents	Not yet applied (value stored only)

When to Use Each Level

Thorough

• Pre-merge code reviews
• Security-sensitive code
• Complex algorithms

Standard

• Iterative fixing with refine
• General code changes
• Balanced speed/quality

Fast

• Quick syntax checks
• Exploratory reviews
• High-volume commits

See Also

• Per-Repository Configuration - Set per-project defaults
• Auto-Fixing - Refine command usage