coding-context-cli copilot-instructions.md

Coding Context CLI - GitHub Copilot Instructions

Project Overview

This is a Go-based command-line tool that assembles dynamic context for AI coding agents. The CLI collects context from rule files, task prompts, and parameters, then outputs a combined context for AI models.

Key Concepts

• Rules: Reusable context snippets (.md or .mdc files) with optional YAML frontmatter
• Tasks: Task-specific prompts identified by task_name in frontmatter
• Selectors: Frontmatter-based filtering (e.g., -s language=Go)
• Parameters: Variable substitution in task prompts (e.g., -p issue_number=123)
• Bootstrap Scripts: Executable scripts that run before rules are processed

Project Structure

.
├── main.go              # CLI entry point and main logic
├── markdown.go          # Frontmatter parsing
├── param_map.go         # Parameter map implementation
├── selector_map.go      # Selector filtering logic
├── token_counter.go     # Token estimation
├── frontmatter.go       # Frontmatter extraction
├── examples/            # Example rules and tasks
│   └── agents/
│       ├── rules/       # Example rule files
│       └── tasks/       # Example task files
└── docs/                # Documentation (Diataxis framework)

Go Coding Standards

Style and Formatting

• Follow standard Go formatting (gofmt)
• Use goimports for import organization
• Run make lint before committing
• Keep functions focused and small
• Use meaningful variable and function names

Error Handling

• Always check and handle errors explicitly
• Wrap errors with context using fmt.Errorf
• Return errors rather than panicking
• Use custom error types when appropriate

Testing

• Write unit tests for all new functions
• Use table-driven tests for multiple scenarios (this is the project standard)
• Aim for >80% test coverage
• Use meaningful test names that describe the scenario (e.g., TestSelectorMap_MatchesIncludes/single_include_-_match)
• Test edge cases and error conditions
• Run go test -v ./... to verify all tests pass