antigravity-claudekit/skills/ck-cook/SKILL.md

name, description

name	description
ck-cook	End-to-end feature implementation with automatic workflow detection. Always activate before implementing any feature, plan, or fix. Use for full-stack features, plan execution, multi-phase delivery.

ck-cook

End-to-end implementation with automatic workflow detection. Orchestrates the full pipeline: research → plan → implement → test → review → finalize.

When to Use

• Implementing any new feature
• Executing an existing plan file
• Multi-phase delivery requiring coordination
• Any time the user says "implement", "build", "add", or "create"

Don't Use When

• Only brainstorming or evaluating options — use ck-brainstorm
• Only creating a plan without implementing — use ck-planning
• Only fixing a specific bug — use ck-fix

Usage

/cook <natural language task OR plan path> [--flag]

Flags

• --interactive: Full workflow with user approval gates (default)
• --fast: Skip research phase; scout → plan → code
• --parallel: Multi-agent parallel execution
• --no-test: Skip testing step
• --auto: Auto-approve all steps

Smart Intent Detection

Input Pattern	Detected Mode
Path to plan.md or phase-*.md	Execute existing plan
Contains "fast", "quick"	Fast mode
Contains "trust me", "auto"	Auto mode
Lists 3+ features OR "parallel"	Parallel mode
Contains "no test", "skip test"	No-test mode
Default	Interactive

Workflow Overview

[Intent Detection] → [Research?] → [Review] → [Plan] → [Review] → [Implement] → [Review] → [Test?] → [Review] → [Finalize]

Default (non-auto): Stops at [Review] gates for human approval before each major step. Auto mode: Skips human review gates; implements all phases continuously.

Mode	Research	Testing	Review Gates
interactive	yes	yes	User approval at each step
auto	yes	yes	Auto if score >= 9.5
fast	no	yes	User approval at each step
parallel	optional	yes	User approval at each step
no-test	yes	no	User approval at each step
code	no	yes	User approval per plan

Required Agent Dispatch (MANDATORY)

Phase	Agent/Skill	Requirement
Research	ck-research	Optional in fast/code
Scout	ck-scout	Optional in code mode
Plan	ck-planning	Optional in code mode
UI Work	ck-frontend-design	If frontend work detected
Testing	ck-web-testing	MUST dispatch
Review	ck-code-review	MUST dispatch
Finalize	docs update + ck-git	MUST run all

CRITICAL: Steps for testing, review, and finalization MUST dispatch separate agents — do NOT perform them inline. If workflow ends with zero agent dispatches, it is INCOMPLETE.

Blocking Gates (Non-Auto Mode)

Human review required at:

• Post-Research: Review findings before planning
• Post-Plan: Approve plan before implementation
• Post-Implementation: Approve code before testing
• Post-Testing: 100% pass + approve before finalize

Always enforced (all modes):

• Testing: 100% pass required (unless no-test mode)
• Code Review: User approval OR auto-approve (score >= 9.5, 0 critical)

Finalize Step (MANDATORY — never skip)

1. Update plan/phase status to complete
2. Update ./docs if changes warrant it
3. Ask user if they want to commit via ck-git

Step Output Format

Step [N]: [Brief status] - [Key metrics]