plugin
Python backend
python-backend
A Python backend orchestrator with skills for FastAPI, TDD, Alembic migrations, CI/CD and performance. Runs pytest before it lets the agent stop.
When to use it
You build or maintain a production Python API.
Includes: 10 subagents
- Architect engineerDesigns module boundaries and API and database contracts, and records each structural decision as a short ADR.
- ChallengerAttacks the implementation with fuzzing, extreme inputs, failure paths and timeouts to find what breaks.
- Code implementerWrites the production code with TDD inside the files it was assigned, and returns the test command and its exit code.
- Compliance verifierRuns static analysis, security checks and architecture rules, and approves or rejects the change.
- CriticReviews the code without changing it: complexity, duplication, SOLID and maintainability.
- E2E black-box testerStarts the app, runs the user journeys in a real browser and saves screenshots. Reports blocked when it cannot run, never a fake pass.
- Product analystClarifies the business rules and writes the product section of the spec with user stories and acceptance criteria.
- Python backend orchestratorPlans and governs Python backend work (FastAPI, Django, Flask) with TDD, migrations and security gates.
- QA testerRuns the affected tests and the full suite once per change, extends the test matrix and closes with done or not done.
- UI/UX designerMaintains DESIGN.md and writes the design notes for a feature: layouts, states and accessibility.
Includes: 22 skills
- Adversarial challengerStress tests, fuzzing, worst-case inputs, timeouts and concurrency probes that look for the place where the code breaks.
- Black-box testerRuns real user journeys in a browser with Playwright, checks console and network errors, and saves screenshots at four screen widths.
- Build and CI gatesSets up linters, formatters, Docker builds and GitHub Actions so broken code never reaches the main branch.
- Code criticAn independent review of complexity, duplication, SOLID, anti-patterns and how easy the code is to change.
- Code implementerWrites production code with test-driven development: a failing test first, then the smallest clean change that passes.
- Code smells expertFinds and classifies code smells and technical debt with the Fowler and Beck catalog, with the reason each one hurts.
- Compliance verifierThe final check before release: non-functional requirements, static analysis, coverage and security, ending in an approved or rejected verdict.
- Context captureSaves what was done, why and how into context/memory/ so another agent or session can pick up the work.
- Context restoreReads the shared memory at the start of a session and summarizes earlier decisions before any work begins.
- Database migration expertSchema migrations, seed scripts and table versioning in Python backends with Alembic and SQL.
- Design spec expertWrites a short API or software design spec before coding new endpoints, schemas or services.
- Loop engineeringThe three-tier process: sizes each task, runs only the specialists it needs, isolates parallel workers and caps review loops at three.
- Performance and scalabilityFinds and fixes bottlenecks, memory leaks and scaling limits with caching, async workers and load profiling.
- Product analystTurns a vague request into clear scope, user stories and testable acceptance criteria in a 50-line spec.
- Pull request expertWrites short, honest pull requests with commit checks, a link to the test matrix and any skipped gate stated openly.
- Python expertModern Python: PEP 8, type hints and Protocols, generators, slots and dataclasses, asyncio TaskGroup and real error handling.
- QA testerDesigns and runs unit, integration and E2E tests and writes a test matrix backed by a real exit code 0.
- Refactoring expertImproves structure, readability and extensibility without changing behavior, with design patterns where they help.
- Security auditReviews code, infrastructure as code and designs for injection, broken access control and exposed secrets.
- Senior architecturePragmatic system design with short decision records (ADRs) and an architecture section based on KISS, YAGNI and SOLID.
- Test-driven developmentRed, green, refactor: failing tests first, solid mocks at the boundaries and designs that are easy to test.
- UI/UX designerDesign systems in a DESIGN.md with tokens, responsive states and WCAG AA accessibility, plus small per-feature design notes.
Includes: 3 hooks
Technical description
Production Python backend engineering skills: FastAPI, TDD, migrations, CI/CD, and performance.