Skip to content

Latest commit

 

History

History
68 lines (54 loc) · 2.65 KB

File metadata and controls

68 lines (54 loc) · 2.65 KB

AGENTS.md - AI Assistant Advisory Notes

Purpose

This document provides AI-specific guidance for working with the SLIM CLI codebase. For project details, refer to the documentation files listed below.

Essential Documentation Order

  1. README.md - Project overview, installation, basic usage
  2. ARCHITECTURE.md - System design, components, extension points
  3. CONTRIBUTING.md - Development setup, how to add features, coding standards
  4. TESTING.md - Test architecture, running tests, writing new tests

AI Assistant Guidelines

When Reviewing Code

  • Python 3.9+ is required due to numpy 2.0.2 dependency
  • UV is the recommended package manager for development
  • pip is included as a dependency for subprocess operations

When Suggesting Changes

  • Follow existing Typer patterns with Rich integration for CLI output
  • Use console.print() for user output, logging.debug() for debugging
  • Maintain the established command structure in src/jpl/slim/commands/
  • New best practices require registration in practice_mapping.py
  • Consider semantic versioning when making changes (update src/jpl/slim/VERSION.txt)
  • Use semantic commit message format: <type>(<scope>): <description>

When Running Commands

  • Always use uv run prefix for development commands
  • Test with --dry-run flag first when making destructive changes
  • Use --logging DEBUG to see detailed execution information

Common Pitfalls to Avoid

  • Don't assume pip/venv setup - check if user is using UV
  • Don't create new files unless absolutely necessary
  • Don't modify git config or push without explicit user request
  • Don't be verbose in GitHub templates - keep issue/PR descriptions clear and concise

Testing Reminders

  • Run uv run pytest tests/ before suggesting PR submission
  • Check both unit tests and YAML integration tests
  • AI-related tests may fail without proper model configuration

Key Context

  • This is a NASA-AMMOS project for applying software best practices
  • Supports AI models via LiteLLM integration
  • Uses YAML-based configuration for prompts and tests
  • Best practices are fetched from a centralized SLIM registry

GitHub Templates

  • Issue templates: Located in .github/ISSUE_TEMPLATE/ directory
  • Pull request template: Located at .github/PULL_REQUEST_TEMPLATE.md
  • When creating issues or PRs: Read and follow these templates for proper formatting
  • Template content: Keep descriptions clear, concise, and accurate - avoid verbose explanations

Quick Command Reference

# Development setup
uv sync

# Run CLI
uv run slim --help

# Run all tests
uv run pytest tests/

# Debug mode
uv run slim --logging DEBUG [command]