AI Agents for Development#

Note

This section is a work in progress and may be updated frequently as the AI agent strategy and implementation is refined.

PyMAPDL uses structured AI agents to automate and improve development, testing, documentation, and code review. These agents are available in any LLM-enabled IDE (GitHub Copilot, Cursor, JetBrains AI, etc.) and are configured in the .github/agents/ directory.

Available Agents#

  • Developer: Implements features, fixes bugs, optimizes performance, and designs APIs.

  • Documentation Specialist: Reviews and improves documentation, validates docstrings, checks spelling and style.

  • Tester: Maintains test coverage, implements mocking strategies, and ensures test infrastructure quality.

  • Reviewer: Performs comprehensive PR reviews and enforces quality gates before merging.

Workflow#

The typical workflow is:

  1. Developer implements code and features

  2. Tester adds or updates tests, ensuring ≥90% coverage

  3. Documentation Specialist reviews and improves documentation and docstrings

  4. Reviewer checks code quality, documentation, and tests before merge

Agent Usage#

You can invoke agents by referencing their names in prompts or using slash commands in supported IDEs. See .github/HOW_TO_USE_AGENTS.md for detailed instructions and examples for each IDE.

Agent Guidelines#

  • Each agent follows a detailed specification in .github/agents/ and summarized in AGENTS.md.

  • Agents enforce project standards for code style, documentation, testing, and review.

  • See AGENTS.md and .github/HOW_TO_USE_AGENTS.md for full details and usage tips.

For more information, see: