| name | ls-plan-w-quizmaster |
| description | Use when the user wants to plan something — load immediately when planning intent is expressed, don't ask first. Triggers on 'I want to plan', 'let's plan', 'plan something out', 'help me plan', 'plan it out', 'quiz me on the requirements', 'plan-w-quizmaster'. NOT a mandatory gate on implementation. For expert opinions use /consult-polymaths. |
Plan with Quizmaster
Plan with Ultimate Quizzer methodology - thoroughly understand requirements via structured questioning before generating any plan.
Variables
| Variable | Source | Description |
|---|
| USER_PROMPT | $1 | The user's request to plan |
| PLAN_OUTPUT | Docs/Plans/ | Output directory |
STEP 1: Variant Selection (MANDATORY FIRST ACTION)
Before doing ANYTHING else, you MUST ask the user which quizmaster prompt variant to use. This is always the very first question — no exceptions.
Use AskUserQuestion with this exact question:
Question: "Which quizmaster prompt variant should we use for this planning session?"
Options:
- "small" — Lightweight, faster questioning. Good for simple tasks.
- "full" — Elaborated 10-domain context sweep. Thorough but verbose.
- "v4" — Visual coverage maps, quality metrics, assumption gate, retrospective.
- "v5" — Self-evaluating, ever-evolving meta planner. Learns and adapts.
- "v6" — Adaptive quizmaster with fact dependencies, unified inversion, invisible ceremony. Built by the Polymathic Tribunal.
- "v7" — v6 + master plan/sub-plan decomposition, workstream discovery, phased DAG orchestration. Enables parallel agent dispatch.
- "v8 (Recommended)" — v7 + specification enrichment, self-contained sub-plans, buildability gate, golden-path scenarios, banned ambiguity. Fixes the decision-to-specification gap that caused placeholder code in v7.
Do NOT skip this step. Do NOT default to any variant. Always ask.
STEP 2: Load the Selected Prompt
After the user selects a variant, use the Read tool to load the corresponding prompt file from the skill directory:
| Variant | File to Read |
|---|
| v8 | ULTIMATE_QUIZZER_PROMPT_v8.md |
| v7 | ULTIMATE_QUIZZER_PROMPT_v7.md |
| v6 | ULTIMATE_QUIZZER_PROMPT_v6.md |
| v5 | ULTIMATE_QUIZZER_PROMPT_v5.md |
| v4 | ULTIMATE_QUIZZER_PROMPT_v4.md |
| full | ULTIMATE_QUIZZER PROMPT_full.md |
| small | ULTIMATE_QUIZZER_PROMPT_small.md |
Read the file from: ${CLAUDE_SKILL_DIR}/<filename>
Follow the loaded prompt's instructions exactly. The prompt file defines the quizmaster's personality, questioning methodology, state tracking format, and all behavioral rules. It is the authority — this SKILL.md only handles variant selection and plan output.
STEP 3: Quiz the User
Follow the loaded prompt file's methodology to interrogate the user's request. The prompt defines how to ask questions, track state, and determine when enough information has been gathered.
STEP 4: Generate the Plan
When the user says "plan it" / "ok plan" / "enough" / "go ahead":
- Validate assumptions - List all assumptions, confirm with user
- Summarize - Known/Open/Assumptions in 5-10 bullets
- Generate plan - Follow the Plan Format below
- Save - Write to
Docs/Plans/<filename>.md
Plan Format
After quizzing is complete, generate the plan. The plan format depends on the variant used:
- v7: Produces a folder with
master.md + sub-<workstream>.md files. The v7 prompt defines the exact master and sub-plan templates. See ULTIMATE_QUIZZER_PROMPT_v7.md for details.
- v6 and earlier: Produces a single plan file using the format below.
Single-File Plan Format (v6 and earlier)
# Plan: <descriptive task name>
## Task Description
<describe what will be accomplished>
## Objective
<clearly state the goal and success criteria>
## Problem Statement
<define the problem being solved>
## Solution Approach
<describe the technical approach>
## Relevant Files
<list files to be modified/created>
## Team Orchestration
You operate as the team lead and orchestrate the team to execute this plan.
You NEVER write code directly - you use Task and Task\* tools to deploy team members.
### Team Members
<list builders and validators>
## Step by Step Tasks
<structured task list with dependencies>
## Acceptance Criteria
<measurable criteria from quizzing>
## Validation Commands
<specific commands to verify completion>
## Assumptions Made
<list assumptions from quizzing session>
## Notes
<optional additional context>
Enforced Best Practices
When generating the plan, always include these engineering disciplines as part of the execution strategy. These are non-negotiable when going through quizmaster planning — the whole point of rigorous planning is rigorous execution.
| Practice | Skill | When |
|---|
| Isolated workspace | git worktrees | Create a worktree before touching code. Keep main clean. |
| Test-driven development | test-driven development | Write tests before implementation for each task in the plan. |
| Structured implementation plan | a structured execution doc | The quizmaster plan feeds directly into a writing-plans execution doc. |
| Systematic debugging | systematic debugging | When tests fail, follow the debugging skill — don't guess. |
| Verification before completion | verification before completion | Every task must pass verification commands before claiming done. |
| Code review | polymathic code review | Request review before merging back. |
| Branch completion | structured merge/PR/cleanup | Follow the structured merge/PR/cleanup flow at the end. |
Include a "Execution Workflow" section in every generated plan that applies these disciplines in order:
- Create worktree → 2. Write tests → 3. Implement → 4. Debug failures → 5. Verify → 6. Review → 7. Finish branch
Report
After saving the plan:
v7 (folder output):
Plan Created: Docs/Plans/<folder-name>/
Files:
- master.md
- sub-<workstream-1>.md
- sub-<workstream-2>.md
- ...
Topic: <brief description>
Quizzing Summary:
- Variant used: v7
- Workstreams identified: N
- Domains covered: X/10
- Questions asked: N
- Assumptions validated: Y
To execute:
# ultracode: open master.md and run it — the Workflow tool fans the sub-plans out across agents
v6 and earlier (single file):
Plan Created: Docs/Plans/<filename>.md
Topic: <brief description>
Quizzing Summary:
- Variant used: <variant>
- Domains covered: X/10
- Questions asked: N
- Assumptions validated: Y
To execute:
# ultracode: open the plan and run it — the Workflow tool fans the sub-plans out across agents
Self-Validation
The Stop hook validates:
- A new .md file exists in Docs/Plans/
- File was created within last 10 minutes
You are now in QUIZMASTER PLAN MODE. Start with STEP 1: ask which variant to use.