mediar-ai/packages/skills/skills/stream-coding

Stream Coding v3.4

Documentation-first development methodology. When docs are clear enough, code generation becomes automatic.

Prerequisites

• Understanding of documentation practices
• AI coding assistant

Instructions

Core Insight

"If your docs are good enough, AI writes the code. The hard work IS the documentation. Code is just the printout."

The Stream Coding Truth

Messy Docs → Vague Specs → AI Guesses → Rework Cycles → 2-3x Velocity
Clear Docs → Clear Specs → AI Executes → Minimal Rework → 10-20x Velocity

Why Most AI-Assisted Development Fails

• People feed AI messy docs
• AI generates code based on assumptions
• Code doesn't match intent
• Endless revision cycles

Why Stream Coding Achieves 10-20x

• Documentation is clarified FIRST
• AI has zero ambiguity
• Code matches intent on first pass
• Minimal revision

13-Item Clarity Gate

Before generating code, documentation must pass:

1. Clear problem statement
2. Defined scope boundaries
3. Input/output specifications
4. Error handling requirements
5. Edge cases documented
6. Dependencies listed
7. Performance requirements
8. Security considerations
9. Testing criteria
10. Integration points
11. Data structures defined
12. API contracts specified
13. User stories/acceptance criteria

Time Allocation

• 40% Documentation & Clarification
• 40% Clarity Gate & Review
• 20% Code Generation & Testing

Guidelines

1. Documentation clarity is the bottleneck, not coding speed
2. Pass all 13 Clarity Gate items before generating code
3. Self-assess documentation quality with scoring rubric
4. Use mandatory section templates for consistency
5. Include Documentation Audit in Phase 1

Notes

• Skip the Clarity Gate and you're back to vibe coding
• Documentation is the hardest part - code is automatic
• v3.4 adds complete scoring rubric with weights

Source: frmoretto/stream-coding