/stage

1. Task context

You are applying ADR002's micro-iteration methodology: add minimal components per specific failure, not rigid stage jumps.

Micro-Iteration Principle:

• Test → Specific failure → Add minimal component to fix that failure
• Stage labels (1.1, 1.2, 1.3) are valid and expected
• When components match Stage N definition, call it Stage N (emergent, not target)

2. Tone context

Write prompts that are clear, systematic, and follow ADR004 standards. Use proper Anthropic template components for the requested stage.

7. Immediate task description or request

Generate/update $2 command to Stage $1 following anthropic-prompt-template.md and ADR002 micro-iteration approach.

Formatting Requirements: Always format the description field as "[YYYY-MM-DD] [Stage X] Command purpose" using the current date to enable version tracking.

8. Thinking step by step (MANDATORY)

You MUST use the sequential_thinking tool with exactly 2 thoughts before generating the staged prompt:

First Thought: Identify specific failure (if iterating)

• What specific failure occurred in testing?
• What minimal component fixes this specific failure?
• Does this component addition match a stage definition?

Second Thought: Plan minimal transformation

• What's the smallest change that fixes the failure?
• Which Anthropic component addresses this specific issue?
• Keep existing working components unchanged

Only after completing both thoughts, proceed to generate the staged prompt.

<prompt_template> @~/.claude/templates/anthropic-prompt-template.md </prompt_template>

Symbol Reference Guide:

• @ - Includes file contents when command runs (provides additional context)
• ~ - Cross-system home directory path (works on both local and remote systems)

IMPORTANT: Never add @ references unless explicitly instructed by the user. These references are design decisions that require human judgment about what context is appropriate for each command.

Available Paths (for human reference):

• @~/.claude/templates/ - Complete document templates
• @~/.claude/patterns/ - Reusable behavioral patterns
• ~/.claude/commands/ - Command definitions

When user provides @ references in their command design, include them exactly as specified.

Apply Anthropic's 4 Best Practices:

1. Be explicit with your instructions - Provide clear, specific details about desired output

   • ❌ Bad: "Create an analytics dashboard"
   • ✅ Good: "Create an analytics dashboard. Include as many relevant features and interactions as possible. Go beyond the basics to create a fully-featured implementation."

2. Add context to improve performance - Explain motivation behind instructions

   • ❌ Bad: "Never use ellipses"
   • ✅ Good: "Your response will be read aloud by a text-to-speech engine, so never use ellipses since the text-to-speech engine will not know how to pronounce them."

3. Be vigilant with examples - Ensure examples align with desired behaviors

4. Respect system boundaries - Clearly separate AI verification from human testing

Use the Anthropic template components appropriate for the requested stage per ADR002 mapping.

Save this command as: ~/.claude/commands/$2.md (requires .md extension for Claude Code recognition)