Specx Project Structure logo

Specx Project Structure

Community
maksimzayats
specx-project-structure

Create or reshape a Python FastAPI service repo into the specx clean core/delivery architecture using packaged scoped foundation bases. Use when starting an API backend, adding the first src package, or establishing `AGENTS.md`, `core/`, optional local `foundation/`, `delivery/`, infrastructure, `ioc/`, migrations, and tests.

Overview

Publishermaksimzayats
Repositoryspecx
Skill namespecx-project-structure
Stars
201
Forks
34
Bundled files
2
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 2 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by maksimzayats on GitHub. Read the source before you install it.

Installation

Install the Specx Project Structure AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/maksimzayats/specx/tree/main/.agents/skills/specx-project-structure
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/maksimzayats/specx.git /tmp/specx
mkdir -p .claude/skills
cp -r /tmp/specx/.agents/skills/specx-project-structure .claude/skills/specx-project-structure
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Specx Project Structure in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Specx Project Structure on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Specx Project Structure is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

specx Project Structure

Use this skill to create the repo shell and first runnable vertical slice. For details, read references/blueprint.md.

Workflow

  1. For a fresh framework-neutral repo, start with specx init <path>. It creates packaging, strict tooling, root AGENTS.md, a small core/health use case and pure service, ioc/container.py, and mirrored unit tests. It intentionally omits delivery and infrastructure; add those packages only with real technology-specific behavior.
  2. Preserve an existing import package declared by the repo. For a new repo, derive a lowercase underscore name, normalize every non-identifier character, and verify name.isidentifier() and not keyword.iskeyword(name); distribution names and import packages need not be identical. If normalization still yields a keyword or leading digit, choose an explicit valid package name rather than emitting it unchanged.
  3. Create src/<package>/ and tests/ packages with empty __init__.py files only.
  4. Add specx as a runtime dependency and import base classes from specx.core.foundation, specx.delivery.foundation, or specx.infrastructure.foundation.
  5. Use core/<scope>/ as the main boundary. Put application packages at the scope root: capabilities/, dtos/, entities/, exceptions/, gateways/, repositories/, services/, and use_cases/. Put scope-owned technical adapters under infrastructure/ only when real IO exists.
  6. Add top-level delivery/ for runnable framework apps, lifecycles, controllers, request/response schemas, and delivery-only services.
  7. Add top-level infrastructure/ for shared technical resources such as SQLAlchemy session factories, logging, telemetry, and external client factories. App-owned pooled clients live here even when only one scope currently consumes them; bounded short-lived clients may stay with the scope adapter.
  8. Add infrastructure/logging/ with a stdlib LoggingConfigurator and LoggingSettings for every new API repo.
  9. Add ioc/ for diwire.Container creation and explicit bindings.
  10. Add shared/ only for stable cross-scope application primitives such as unit-of-work contracts, clocks, ids, or errors.
  11. Add migrations/ with Alembic when SQLAlchemy models exist.
  12. Build the first user-requested business vertical slice. The initializer's process-health slice is a framework-neutral composition example, not a readiness check. With a delivery framework, map or replace it deliberately: keep simple liveness at delivery, and use core/health for /readyz when readiness checks a required dependency or policy is shared across deliveries.
  13. Create root AGENTS.md for every new repo. Include runnable project commands from $specx-project-tooling and specx boundaries from $specx-component-architecture.
  14. Do not create an empty local foundation/ package. Create src/<package>/foundation/ only for a real project-local base category or a stateful framework base that must not be shared globally, such as a SQLAlchemy declarative base.
  15. Add tests only where there is real code to test. Do not create empty folders just to satisfy the diagram.

Non-Negotiable Boundaries

  • Keep all source and test __init__.py files empty. Do not add empty local foundation, source, test, or helper packages.
  • Every non-foundation source class inherits an explicit packaged or local scoped base and uses the suffix implied by that ancestry. Every source class, including local bases, has a scoped docstring with a real Example:.
  • Keep core inner packages free of delivery, IOC, SQLAlchemy, Redis, HTTP clients, and framework imports. Put concrete IO adapters under core/<scope>/infrastructure/<technology>/.
  • Capabilities are narrow BaseCapability collaborators; gateways are business-language BaseGateway ports; core services choose BasePureService, BaseReadService, or BaseEffectService.
  • Same-file BaseCommand or BaseQuery inputs enter use cases, and result DTOs leave them. Persistence use cases inject a UnitOfWorkManager; services do not open or finish transactions.
  • Prefer frozen, keyword-only, slotted dataclasses for core data and BaseStrEnum for reusable closed value sets. Keep Pydantic at delivery and settings edges.
  • Keep infrastructure technical and schema evolution in Alembic. Never call metadata.create_all or drop_all in application or test code.
  • Configure stdlib logging once before composing the FastAPI app. Use FastAPI lifespan for resource cleanup, then close the container; do not run schema changes or business workflows in lifespan.
  • Business routes use full /api/v1/... paths. Keep /healthz independent of external systems; use /readyz and core/health for any required external dependency, with bounded checks and minimal responses.
  • Mirror real source behavior with flat tests. Unit tests resolve targets from a native pytest container fixture; integration tests use the real internal graph and replace only external boundaries. Read the blueprint before adding shared doubles or support helpers.
  • Keep root AGENTS.md commands aligned with the Makefile and include only guidance that matches files and features present in the project.
  • This skill targets FastAPI and must enable the opt-in fastapi rule family in [tool.specx]. Framework-neutral specx rules remain the default for workers, CLIs, and other delivery technologies.

References

  • references/blueprint.md - target tree, starter files, and creation checklist.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Specx Project Structure AI skill do?

Create or reshape a Python FastAPI service repo into the specx clean core/delivery architecture using packaged scoped foundation bases. Use when starting an API backend, adding the first src package, or establishing `AGENTS.md`, `core/`, optional local `foundation/`, `delivery/`, infrastructure, `ioc/`, migrations, and tests.

Why use Specx Project Structure on TypingMind?

Because you install it once and use it with any model. Specx Project Structure is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Specx Project Structure in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/maksimzayats/specx/tree/main/.agents/skills/specx-project-structure. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Specx Project Structure?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Specx Project Structure?

As many as you like. As long as a model supports skills, you can use Specx Project Structure with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Specx Project Structure AI skill free?

Yes. It is published on GitHub by maksimzayats under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇