mike-diff /orient — Codebase Proficiency
Build comprehensive understanding of the codebase before starting work.
Workflow
Phase 2 - Structure Mapping:
- Explore directory structure
- Identify key modules and their responsibilities
- Map entry points and main code paths
Phase 3 - Pattern Recognition:
- Note coding conventions and patterns used
- Identify abstractions and how they're applied
- Understand testing approach and file organisation
Phase 4 - Synthesise:
- Build mental model of how components connect
- Identify the most important files to understand
- Present orientation summary
Process
Step 1: Read Project Documentation
cat AGENTS.md README.md 2>/dev/null
cat package.json pyproject.toml Cargo.toml go.mod 2>/dev/null | head -50
Step 2: Explore Structure
ls -la
ls -R src/ app/ lib/ 2>/dev/null | head -100
find . -maxdepth 3 -name "*.md" -o -name "main.*" -o -name "index.*" 2>/dev/null | head -20
Step 3: Understand Patterns
Read 2-3 representative files to understand code style, module structure, import patterns, and error handling.
Step 4: Deep Dive Key Areas
Use codebase search to understand the main entry point, core abstractions, and how components communicate.
Output
Tech Stack
- Language: [primary language and version]
- Framework: [main framework]
- Build: [build tool and key scripts]
- Test: [test framework]
Architecture
[2-3 sentences describing the overall architecture]
Key Components
[path]- [responsibility][path]- [responsibility]
Patterns & Conventions
- [Notable pattern 1]
- [Notable pattern 2]
Entry Points
- Main:
[file] - Tests:
[command] - Dev:
[command]
Understanding Checklist
- Core data flow understood
- Main abstractions identified
- Test patterns recognised
- Build process clear
</output_format>
## Constraints
<constraints>
FOCUS:
- Understanding over speed
- Accuracy over completeness
- Key paths over exhaustive coverage
AVOID:
- Superficial scanning without comprehension
- Suggesting changes or improvements
- Making assumptions without verification
</constraints>