reasoning

name	reasoning
description	Inspect or set OmA reasoning effort profile (global + per-teammate overrides) for depth/cost control.

Run OmA reasoning-effort manager.

Input: $ARGUMENTS

Protocol:

1. Parse requested global effort value and optional teammate overrides.
   • example: global=high reviewer=xhigh executor=medium researcher=low
2. Supported values:
   • low
   • medium
   • high
   • xhigh
3. If no value is provided, show current profile from .omg/state/reasoning.json when available.
4. Map global effort to OmA posture:
   • planning depth
   • verification strictness
   • expected latency/token cost
5. If teammate overrides are provided, apply them only to listed roles and keep the global default for others.
6. Read .omg/state/session-lock.json before mutating shared reasoning state.
7. If filesystem tools are available:
   • if the current orchestration session owns the lock, write/update .omg/state/reasoning.json
   • otherwise write .omg/state/sessions/[session-slug]/reasoning.json instead and report the ownership conflict
8. Clarify boundary: this profile controls OmA orchestration behavior and recommendations, not Gemini runtime flags directly.

Suggested reasoning.json shape: { "effort": "low|medium|high|xhigh", "planning_depth": "minimal|balanced|deep|maximum", "verification_strictness": "light|normal|high|very-high", "team_overrides": { "oma-reviewer": "xhigh", "oma-executor": "medium" }, "updated_at": "", "source": "oma:reasoning" }

Response:

• Keep it concise and operator-facing.
• Include Reasoning Status, Effort Matrix, Team Overrides, and Recommended Route.