# Colin Development Guidelines

> Colin (Context Lineage) is a context engine for the AI era. It takes interconnected source documents, resolves dependencies, applies transformations (including LLM calls), and produces outputs your agents can use.

- Skill: `tools-only/colin-development-guidelines` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add tools-only/colin-development-guidelines`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tools-only/colin-development-guidelines/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: tools-only (https://skillmd.com/u/tools-only)
- Updated: 2026-09-29
- Page: https://skillmd.com/skills/tools-only/colin-development-guidelines

---

# Colin Development Guidelines

Colin (**Co**ntext **Lin**eage) is a context engine for the AI era. It takes interconnected source documents, resolves dependencies, applies transformations (including LLM calls), and produces outputs your agents can use.

## Required Development Workflow

**CRITICAL**: Always run these commands before committing:

```bash
uv sync                    # Install dependencies
uv run prek run --all-files  # Lint (ruff) + type check (ty)
uv run pytest              # Run tests
```

**All must pass** - this is enforced by CI.

## Documentation Requirements

**CRITICAL**: These docs must be kept up to date with all work:

### User-Facing Documentation (`docs/`)

**MANDATORY** for any user-facing changes:

- New features, configuration options, CLI changes, template syntax
- Documentation must be holistically integrated, not appended
- Restructure existing documentation if needed for clarity
- Consider the complete user journey—how will users discover and learn this?
- Pro forma documentation is unacceptable; every word must serve the user

**When updating user docs:**

- Read related docs first to understand structure and voice
- Place new content where users would naturally look for it
- Update related sections that reference affected functionality
- Remove or update any content that becomes incorrect
- Consider whether new pages are needed or existing pages should be restructured

### `docs/architecture.md`

- Live updated architectural overview
- Must reflect current system design
- Update when architecture changes

### `docs/decisions/`

- Architecture Decision Records (ADRs)
- Format: `NNN-short-name.md` (e.g., `001-mvp-scope.md`)
- Record *why* core decisions were made
- New architectural decisions require new files
- Not every small choice - only significant architectural decisions

**When to create an ADR:**

- Choosing between multiple valid approaches
- Scope decisions (what's in/out)
- Technology choices
- Pattern decisions that affect multiple components

## Testing Standards

- Every test: atomic, self-contained, single functionality
- Use parameterization for multiple examples of same functionality
- Use separate tests for different functionality pieces
- **ALWAYS** put imports at the top of the file, not in the test body
- **NEVER** add `@pytest.mark.asyncio` to tests - `asyncio_mode = "auto"` is set globally
- **ALWAYS** run pytest after significant changes

## Code Standards

- Python ≥ 3.10 with full type annotations
- Use `from __future__ import annotations` only where needed (files with forward references, TYPE_CHECKING blocks, or complex type annotations)
- **ALWAYS** put imports at module level (no local imports inside functions)
- **NEVER** export everything in `__init__.py` - `__all__` is a DX surface for framework USERS, not internal convenience. Ruthlessly curate to maximize user capability with minimal surface area. The framework itself imports directly from modules.
- Minimize `# type: ignore` - fix types properly instead
- Follow existing patterns and maintain consistency
- Prioritize readable, understandable code - clarity over cleverness
- Never use bare `except` - be specific with exception types
- Each feature needs corresponding tests

## Git & Commits

- Never force-push on collaborative repos
- Keep commit messages brief - just headlines
- Run checks before committing
- Never amend commits

## Writing Style

- Be brief and to the point
- **NEVER** use "This isn't..." or "not just..." constructions
- State what something IS directly