A collection of Claude Code skills for AI-driven development workflows.
| Skill | Description |
|---|---|
| write-spec | Writes and refines specs for AI-driven development. Guides discovery with targeted questions, then produces unambiguous requirements with testable verification criteria. |
| review-spec | Reviews a spec across ten quality dimensions (goal clarity, actors, inputs/outputs, failure modes, etc.) and produces a prioritized list of issues, warnings, and suggestions. |
| write-spec-plan | Generates an iterative implementation plan from a spec file and writes it to a companion -plan.md file. Requires the spec to have passed review-spec first. |
| Skill | Description |
|---|---|
| implement-spec-plan | Executes a plan file one iteration at a time, verifying acceptance criteria and running full regression tests before advancing. |
| verify-spec | Confirms the codebase fully satisfies a spec by walking through each iteration's acceptance criteria and running the full test suite. |
| freeze-spec | Distills a completed spec into a compact, immutable "context capsule" — preserving purpose, public interfaces, key decisions, and known limits while dropping internal mechanics — for use as background context in future work. |
| Skill | Description |
|---|---|
| ddd-analysis | Produces a complete Domain-Driven Design document (ubiquitous language, bounded contexts, aggregates, domain events, etc.) from requirements or a product description, one section at a time with user confirmation. |
| Skill | Description |
|---|---|
| python-coding | Standards and conventions for Python projects using uv, ruff, ty, strict typing, FastAPI, Strawberry, SQLAlchemy, Alembic, and Pydantic. |
| using-pytest | Guidelines for writing pytest tests following Given/When/Then/Clean patterns, with business-readable docstrings and organized test layers (unit, functional, interface, integration, evals). |
| Skill | Description |
|---|---|
| write-system-prompt | Authors production-quality system prompts for LLM agents and assistants. Asks clarifying questions before drafting and structures output across role, instructions, constraints, output format, examples, and tools. |
write-spec → review-spec → write-spec-plan → implement-spec-plan → verify-spec