# Mode Selection

> For understanding what each mode does, which mode is best (most appropriate mode) for a given situation, and establishing a pattern for choosing when and how a mode delegates or passes operation to another mode. This skill provides `mode selection strategy`. Load with read_file on .kilocode/skills/mode-selection/SKILL.md (ignore the absolute path in the location tag).

- Skill: `scottermonk/mode-selection` (Agent Skill)
- Install (CLI): `npx skillmds@latest add scottermonk/mode-selection`
- Raw SKILL.md: https://api.skillmd.com/api/skills/scottermonk/mode-selection/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: ScotterMonk (https://skillmd.com/u/scottermonk)
- Updated: 2026-09-10
- Page: https://skillmd.com/skills/scottermonk/mode-selection

---


# Mode selection

## Mode descriptions

**Location**
- Read skill-local guidance only when `.kilocode/skills/mode-selection/AGENTS.md` is confirmed to exist; otherwise use root `AGENTS.md` for project-specific mode-rule locations.

**For Planning & Orchestration**
- `architect`: All-in-one planning. Create phases and tasks -> QA -> User Approval -> Switch to `dispatcher`.
- `planner-a`: Complex Planing Stage 1. Create phases -> Brainstorm -> Switch to `planner-b`.
- `planner-b`: Complex Planning Stage 2. Create detailed tasks -> User Approval -> Switch to `planner-c`.
- `planner-c`: Complex Planning Stage 3. QA -> Finalize -> User Approval -> Switch to `dispatcher`.
- `dispatcher`: Manage execution. Coordinate implementation modes. Log. Fullfill the plan.
**For Implementation & Ops**
- `coder-sr`: Complex engineering, analysis, debugging, test creation, test execution, deep error investigation and diagnosis. Front-end and back-end.
- `coder-jr`: Routine coding, test creation, test execution, strict instruction adherence. Front-end and back-end.
- `githubber`: Git/GitHub operations.
- `tasky`: Small, isolated work, including simple file operations.
- `ask`: General inquiries.

## Mode selection strategy
**Evaluate** the current `task`. If another mode is more appropriate, **pass** the `task` and payload to that mode.
**Payload Requirements**:
- **Context**: Include relevant bug/issue details. If a `plan` is active, include pertinent sections.
- **Instructions**: Provide specific, actionable implementation steps.
- **Completion Trigger**: Specify the exact command to return when the task is finished.
**Response Protocol**:
- Mandate a reply via the `result` parameter containing a concise summary of the outcome.

**Prioritize** budget-friendly modes in this order (Low to High):
1.  **Low Budget** (Renaming, moving files, simple text replacement, DB column copying)
    - Use `tasky`.
2.  **Medium Budget** (Refactoring, simple function creation, writing, or if Low fails)
    - Use `coder-jr`.
3.  **High Budget** (Complex modification, test creation and use, debugging, or if Medium fails)
    - Use `coder-sr`.