Coding Context Adapter¶
Transforms codebase analysis into AI agent coding context.
Purpose¶
Generate context files that help AI agents understand and work with the codebase:
- AGENTS.md: Instructions for AI agents working on this codebase
- CONTEXT.md: Project context, patterns, and conventions
Output¶
Or in specified location:
Required Model Sections¶
| Section | Required | Used For |
|---|---|---|
meta |
Yes | Project identification |
technologies |
Yes | Tech stack context |
architecture |
Yes | Code organization |
interfaces |
Optional | API patterns |
data |
Optional | Data model patterns |
quality |
Optional | Code quality notes |
Output Structure¶
AGENTS.md¶
# AI Agent Instructions: {Project Name}
## Project Overview
{From meta.description}
## Technology Stack
{From technologies - what's used and why}
## Code Organization
{From architecture.structure and architecture.layers}
## Key Patterns
{From architecture.patterns - how things are done here}
## Coding Conventions
{Inferred from code analysis}
- Naming: {camelCase, snake_case, etc.}
- File structure: {patterns observed}
- Error handling: {standard approach}
- Testing: {patterns and locations}
## Important Files
{Key entry points, configs, frequently modified}
## Build & Run
{From technologies.build_tools}
## When Making Changes
- {Pattern to follow}
- {Files to update together}
- {Tests to run}
## What to Avoid
{Anti-patterns or deprecated approaches}
CONTEXT.md¶
# Project Context: {Project Name}
## What This Project Does
{From meta.description, interfaces overview}
## Architecture
{From architecture.components and architecture.layers}
### Components
{Component list with responsibilities}
### Dependencies Between Components
{Component dependency graph}
## Data Model
{From data.entities - key entities and relationships}
## External Integrations
{From interfaces.integrations}
## API Surface
{From interfaces.apis - key endpoints}
## Development Setup
{From technologies.build_tools}
## Common Tasks
- Adding a new endpoint: {pattern}
- Adding a new entity: {pattern}
- Running tests: {command}
Model to Output Mapping¶
Project Overview¶
| Model Path | AGENTS.md Section |
|---|---|
meta.project_name |
Title |
meta.description |
Overview |
meta.repository |
Links |
Technology Context¶
| Model Path | Output |
|---|---|
technologies.languages[] |
Primary language, version |
technologies.frameworks[] |
Framework context |
technologies.build_tools[] |
Build commands |
Code Organization¶
| Model Path | Output |
|---|---|
architecture.structure.entry_points |
Main files |
architecture.structure.config_locations |
Config files |
architecture.layers[] |
Layer description |
architecture.patterns[] |
Patterns to follow |
Conventions (Inferred)¶
Analyze code to infer: - Naming conventions (from identifier patterns) - File naming (from directory structure) - Import patterns (from import statements) - Error handling style (from error_handling patterns) - Test naming and location
Invocation¶
"Generate coding context"
"Create AGENTS.md"
"Generate AI context files"
"Create context for AI coding"
Or as part of multi-output:
Configuration¶
Output Location¶
Where should I create the context files?
1. Project root (./AGENTS.md, ./CONTEXT.md)
2. Documentation directory ({docs}/AGENTS.md, {docs}/CONTEXT.md)
Default: Project root
Detail Level¶
What level of detail?
1. Minimal - Key patterns only
2. Standard - Patterns + conventions
3. Comprehensive - Full context with examples
Default: Standard
Use Cases¶
New AI Agent Onboarding¶
When AI starts working on codebase: 1. Reads AGENTS.md for instructions 2. Reads CONTEXT.md for understanding 3. Has immediate context for effective assistance
Documentation for AI Tools¶
Works with: - Claude Code (via AGENTS.md) - GitHub Copilot (via CONTEXT.md) - Cursor (via context files) - Other AI coding assistants
Human Onboarding¶
Context files also help human developers: - Quick project overview - Pattern reference - Where things are located
See Also¶
- Templates - Output templates
- codebase-analysis - Analysis engine
- analysis-outputs - Adapter family