EPC: My Production Workflow

This is how I code. Fast, efficient, production-ready. No fluff.

My philosophy: Ship code that works.

• Every decision ships to production
• Performance beats perfection
• Working code now > elegant code later
• Measure everything, guess nothing

Why I Built This

Standard AI assistants waste my time with essays. I need:

• Instant fixes for real problems
• Responses under 300 tokens
• Code that goes straight to production
• Zero explanations unless I ask

Commands

/epc optimize search          # Standard: analyze → plan → fix
/epc quick fix memory leak    # Quick: instant pattern match
/epc emergency payment crash  # Emergency: fix now, think later
/epc typescript add types     # TypeScript mode
/epc analyze slow checkout    # Deep dive (when needed)

My Agents

🟡 javascript-specialist - Algorithms & performance. O(n²)→O(n). Done.

🔵 react-virtuoso - React UI. <16ms renders. Accessible. Fast.

⚪ git-wizard - Git operations. Clean history. No merge hell.

Instant Fixes

Problem	Solution	Confidence
Slow list	React.memo	10/10
undefined error	Optional chaining	10/10
Memory leak	WeakMap	9/10
Prop drilling	Context	9/10
Stack overflow	Remove recursion	10/10

Non-Negotiable Rules

1. Code goes to files - Never in chat
2. Metrics only - 500ms→50ms, not "faster"
3. One solution - The right one
4. Ship confidence - Every fix has a score
5. Production first - Works > perfect

My Standards

Performance:

• Renders <16ms or it's broken
• Bundle <250KB or it's bloated
• O(n²) is unacceptable
• Memory leaks = emergency

Quality:

• Functions <50 lines
• Complexity <10
• Accessibility required
• Tests when critical

What I Don't Do

❌ Explain obvious things
❌ Show multiple options
❌ Justify simple fixes
❌ Write documentation first
❌ Optimize prematurely
❌ Debate patterns

Results

My workflow ships code that:

• Handles millions of users
• Renders in milliseconds
• Recovers from errors
• Scales without rewrites

This isn't about following trends. It's about shipping code that works, performs, and scales. Today.

License

MIT - Take it, adapt it, ship it.

---

Built by someone who measures performance in milliseconds and ships code that scales.