AgentToolHub
Skill

setting-up-logging

setting-up-logging — a Claude Code skill from quick-brown-foxxx/coding_rules_python

Features

  • ✓Colored logging setup
  • ✓Enhanced output readability
  • ✓Customizable logging configurations
  • ✓Shared utilities for linting
  • ✓Visit GitHub for more details

Installation Guide

npx -y skills add quick-brown-foxxx/coding_rules_python

Note: skills updates detection has bugs, more reliable is to do force add from time to time. Eg for all skills in current project for specified agents:

npx -y skills add quick-brown-foxxx/coding_rules_python -s "*" -a claude-code universal kilo codex opencode -y

Reusable code (shadcn-ui style)

The reusable/ folder contains copy-paste components — not an installable library. Copy what you need into your project's shared/ directory and update import paths. Each module is self-contained and documented.

Starting a new project

  • Read PHILOSOPHY.md for the mindset

  • Decide: single script or full project? (see writing-python-scripts or setting-up-python-projects skills)

  • Copy relevant files from templates/ into your project

  • Copy rules/coding_rules.md (or _short) into docs/

  • Copy PHILOSOPHY.md into docs/ (referenced by AGENTS.md)

  • Create CLAUDE.md symlink → AGENTS.md (Claude Code reads CLAUDE.md; the @docs/PHILOSOPHY.md import in AGENTS.md auto-loads philosophy into context)

  • Fill in TODO sections in AGENTS.md and pyproject.toml

  • Run uv sync && uv run pre-commit install

    • For AI agents

    Deploy skills from skills/ to ~/.claude/skills/. The top-level writing-python-code skill covers core standards; other skills cover specific domains.

    Quick reference

  • What tools? uv, basedpyright (strict), ruff, pytest, poethepoet, msgspec

  • Error handling? Result[T, E] from rusty-results. Exceptions = bugs only.

  • Data validation? msgspec.Struct for external data (JSON, configs, APIs). Validates at decode time.

  • CLI? typer (argparse only for stdlib-only scripts)

  • GUI? PySide6 + qasync (not QtAsyncio — still in technical preview)

  • Text output? Jinja2

  • Tests? E2e > unit. Real > mocked. Trustworthiness > coverage.

    • Reviews