evennia/AGENTS.md

AGENTS.md

This file provides guidance to AI coding agents working in this repository.

Evennia is a Python (>= 3.12) framework for building text-based multiplayer online games (MUD/MUX/MUSH/MOO). It is a library, not a game.

Quick Reference

Use uv run to execute commands in the project venv. Prefer uv over pip.

uv pip install -e .                          # dev install
make format                                  # black + isort
make lint                                    # black --check
make cleanrot                                # check agent context for rot
uv run pytest .agents/tools/tests/ -v        # agent tooling tests (pytest)

Running Evennia tests: make test requires evennia on PATH. With uv run, you must init a test game dir first and run from inside it — see Testing.

Key Rules

• TDD: write tests first. Prefer no-DB tests > mocks > full DB-backed tests.
• Don't manually format code. Run make format after editing.
• All code must have Google-style docstrings. See Code Style.
• After editing agent context files (anything under .agents/ or AGENTS.md), run make cleanrot and fix all warnings before reporting done.

Skills

Vendor-agnostic skills live in .agents/skills/. Vendor directories (.claude/skills/, .codex/skills/) symlink there.

Docs

• Core Beliefs — design principles that guide tradeoffs (toolkit not game, think in Python not SQL, hooks not patches, compose don't branch)
• Architecture — two-process model, typeclass system, command system, subsystems, flat API, settings
• Testing — running tests, test base classes, DB setup, CI matrix
• Code Style — docstring conventions, command docstring format
• Development Commands — install, game lifecycle, test/format commands, PR conventions
• CI/CD — GitHub Actions workflows, test matrix, database configs, Docker, secrets
• GitHub Issues & PRs — listing, searching, and reviewing issues/PRs with gh CLI