Skip to content

AGENTS.md

Latest commit

 

History

History
67 lines (44 loc) · 2.12 KB

AGENTS.md

File metadata and controls

67 lines (44 loc) · 2.12 KB

Project AGENTS Guide for OpenAI Codex

This AGENTS.md file provides guidance for OpenAI Codex and other AI agents contributing to the DisMAL codebase.

DisMAL (Discovery-Mod-And-Lookup Toolkit) is a CLI-based Python tool for interacting with BMC Discovery appliances via API or CLI.

Project Structure

  • /core: Primary source code for data access, command dispatch, and output formatting
  • /bin: CLI entry points and scripts
  • /docs: Documentation files and markdown guides
  • /tests: Pytest-based unit and integration tests
  • /output_*: Example directories for generated output (used for testing or demonstration)
  • /tools: Optional utilities for development, automation, or Codex agent runners

Agents must avoid modifying:

  • *.csv or generated data under output_*
  • Any binary log or appliance exports

Coding Conventions

General Python

  • Use Python 3.8+ compatible syntax
  • Follow PEP8 style guide unless local conventions differ
  • Prefer built-in libraries (argparse, subprocess, requests) unless justified
  • Use meaningful function and variable names
  • Include inline comments for any logic branches, string parsing, or regex
  • Write code that a junior developer can easily read and understand

CLI-Specific Guidance

  • All CLI commands are defined and dispatched in dismal.py using argparse
  • Preserve subcommand structure and --access_method logic
  • Add new subcommands only if tested and scoped appropriately

Output and File Writing

  • Generated output should go into a new subdirectory named output_<appliance> or a temporary path during tests
  • Use --stdout only for human inspection, not structured parsing

Branch Policy

  • Perform code changes on the codex branch. If the branch is unavailable or you are on a different branch, warn the maintainer before proceeding.

Testing Requirements

Tests should be created and maintained under /tests and follow these conventions:

# Run all tests
python3 -m pytest

# Run specific test file
python3 -m pytest tests/test_access.py

# Generate coverage report
coverage run -m pytest && coverage report