Python code quality linter with plugin-based validators and LLM-powered analysis workflows.

• Plugin-based validators - Extensible architecture for custom code quality rules
• AI-powered analysis - LLM integration for semantic validation and architectural analysis
• Justification workflow - Analyzes project structure and provides architectural recommendations
• Multiple output formats - JSON, Markdown, SARIF, natural language
• Configuration management - TOML-based config with fail-fast validation

pip install -e .

# Run validators on current directory
vibelint check .

# Run with automatic fixing
vibelint check --fix .

# Run justification workflow for architectural analysis
vibelint justification .

# Show available validators
vibelint check --help

Add to pyproject.toml:

[tool.vibelint]
# Validator severity overrides
rules = { EMOJI-USAGE = "BLOCK", DICT-GET-FALLBACK = "WARN" }

# LLM configuration (optional)
[tool.vibelint.llm]
fast_api_url = "http://localhost:1234/v1"
orchestrator_api_url = "http://localhost:1235/v1"

# Embedding configuration (optional)
[tool.vibelint.embedding_analysis]
enabled = true
model = "google/embeddinggemma-300m"

See AGENTS.instructions.md for development guidelines.

# Run tests
tox -e py311

# Run formatters and linters
isort src/ tests/
black src/ tests/
ruff check --fix src/ tests/
pyright src/

• Development Instructions
• Claude Code Config

See LICENSE file.