# Human Alignment & Preferences
> How to steer Boring's AI to match your personal or team's "Human Needs".
---
## 🧠 The Alignment Engine
Boring doesn't just work in a vacuum. It is designed to adapt to **how you work**. This is achieved through two main systems: **Rubrics** (Explicit Rules) and **Learned Memory** (Implicit Habits).
---
## 📋 1. Explicit Steering: Rubrics
Rubrics are JSON files located in `.boring/brain/rubrics/`. They tell the AI exactly how to evaluate its own work before presenting it to you.
### How to use them:
1. **Create a Rubric**: Define a new `.json` file (e.g., `mobile_standards.json`).
2. **Define Criteria**: Add weights and levels for things like "Performance", "Accessibility", or "Variable Naming".
3. **AI Self-Correction**: When Boring runs a task, it uses these rubrics to grade itself. If it gets a low score, it will **automatically refactor** the code until it meets your "Human Standards".
---
## 💾 2. Implicit Learning: Memory
Every time you correct Boring (e.g., "I prefer using `async/await` here"), Boring uses the `boring_learn` tool to save that preference.
### The Learning Loop:
- **Interaction**: You provide feedback or fix an error.
- **Pattern Extraction**: Boring identifies the "Fix Pattern" and stores it in `.boring/brain/learned_patterns/`.
- **Precognition**: Next time a similar situation occurs, Boring will proactively use your preferred style.
---
## 🔄 3. Adaptive Workflows (SpecKit)
The `.agent/workflows/` files are not static. The **Workflow Evolver** analyzes your project's unique requirements and modifies these checklists automatically.
### Making it "More Human":
- If your team requires a security audit for every DB change, just mention it in your `PROMPT.md`.
- Boring will **evolve** your `speckit-tasks.md` to include a mandatory `[ ] Security Check` step.
---
## 💡 Best Practices for Alignment
1. **Be Explicit**: Use `boring profile learn` to manually teach the AI your favorite libraries or linting styles.
2. **Review the Brain**: Occasionally check `.boring/brain/learned_patterns/` to see what the AI has learned about you. You can manually delete any patterns that are no longer relevant.
3. **Team Sharing**: Commit the `.boring/brain/rubrics/` folder to Git so your entire team shares the same "Human Quality Gates".
