Test Kitchen

Test Kitchen

Parallel implementation framework with two gate skills:

| Skill | Gate | Trigger |

|-------|------|---------|

| test-kitchen:omakase-off | Entry | FIRST on any build/create/implement request |

| test-kitchen:cookoff | Exit | At design→implementation transition |

Flow

"Build X" / "Create Y" / "Implement Z"
    ↓
┌─────────────────────────────────────┐
│  OMAKASE-OFF (entry gate)           │
│  Wraps brainstorming                │
│                                     │
│  Choice:                            │
│  1. Brainstorm together             │
│  2. Omakase (3-5 parallel designs)  │
└─────────────────────────────────────┘
    ↓
[Brainstorming / Design phase]
    ↓
Design complete, "let's implement"
    ↓
┌─────────────────────────────────────┐
│  COOKOFF (exit gate)                │
│  Wraps implementation               │
│                                     │
│  Choice:                            │
│  1. Cookoff (2-5 parallel agents)   │
│  2. Single subagent                 │
│  3. Local implementation            │
└─────────────────────────────────────┘
    ↓
[Implementation]

Key Insight

Skills need aggressive triggers to work. They can't passively detect "uncertainty" or "readiness" - they must claim specific moments in the conversation flow.

• Omakase-off: Claims the BUILD/CREATE moment (before brainstorming)
• Cookoff: Claims the IMPLEMENT moment (after design)

When Each Triggers

Omakase-off (Three Triggers)

Trigger 1: BEFORE brainstorming

• "I want to build...", "Create a...", "Implement...", "Add a feature..."
• ANY signal to start building something
• Offers choice: Brainstorm together OR Omakase (parallel designs)

Trigger 2: DURING brainstorming (slot detection)

• 2+ uncertain responses on architectural decisions
• "not sure", "don't know", "either works", "you pick", "no preference"
• Offers to explore detected slots in parallel

Trigger 3: Explicitly requested

• "try both approaches", "explore both", "omakase"
• "implement both variants", "let's see which is better"

Cookoff

• "Let's implement"
• "Looks good, let's build"
• "Ready to code"
• Design doc just committed
• ANY signal to move from design to code

Omakase Mode (Skip Brainstorming)

If user picks "Omakase" at the entry gate:

• Quick context gathering (1-2 questions)
• Generate 3-5 best architectural approaches
• Implement ALL in parallel
• Tests pick the winner
• Skip detailed brainstorming entirely

Best for: "I'm flexible, show me options in working code"

Cookoff Mode (Parallel Implementation)

If user picks "Cookoff" at the exit gate:

• Each agent reads the same design doc
• Each agent creates their OWN implementation plan
• All implement in parallel
• Compare results, pick winner

Best for: "I want to see different implementation approaches"

Key Distinction

| | Omakase-off | Cookoff |

|-|-------------|---------|

| Gate | Entry (before/during brainstorming) | Exit (after design) |

| Question | HOW to explore? | HOW to implement? |

| Parallel on | Different DESIGNS | Same design, different PLANS |

| Triggers | Build request, indecision detection, explicit | "let's implement" signal |

| Skips | Brainstorming (optional via short-circuit) | Nothing - always after design |

Slot Detection (During Brainstorming)

When omakase-off delegates to brainstorming, it passively tracks architectural decisions where user shows uncertainty:

Detection signals:

• "not sure", "don't know", "either works", "both sound good"
• "you pick", "whatever you think", "no preference"
• User defers 2+ decisions in a row

Slot classification:

| Type | Examples | Worth exploring? |

|------|----------|------------------|

| Architectural | Storage engine, framework, auth method | Yes - different code paths |

| Trivial | File location, naming, config format | No - easy to change |

At end of brainstorming:

• If architectural slots exist → offer parallel exploration
• If no slots → hand off to cookoff for implementation