linux-mcp-server AGENTS.md

Linux MCP Server

Read-only MCP server for Linux system diagnostics.

Commands

uv sync                      # Install dependencies
uv run pytest                # Run tests
uv run ruff check src tests  # Lint
uv run pyright               # Type check
make verify                  # All checks (required before commit)

Project Layout

• src/linux_mcp_server/tools/ - MCP tools (logs, network, processes, services, storage, system_info)
• src/linux_mcp_server/commands.py - Command definitions
• src/linux_mcp_server/formatters.py / parsers.py - Output formatting and parsing
• tests/ - Mirrors src structure

Rules

Code: PEP 8, type hints required, async/await for I/O, 120 char max, prefer Pydantic over dataclasses

Testing:

• Run make verify before committing
• Use parameterized tests and fixtures (shared fixtures go in conftest.py)
• Use pytest-mock (mocker fixture) for mocking instead of unittest.mock imports
• Use autospec=True when patching; spec=<object> with Mock
• 100% patch coverage for new code
• Do not use @pytest.mark.asyncio for tests. It's not necessary because the project uses asyncio_mode = "auto"

Security (Critical):

• All tools must be read-only with readOnlyHint=True
• Validate all input, use allowlists for file paths, sanitize shell params

Adding Tools

1. Create tool in src/linux_mcp_server/tools/ using @mcp.tool(), @log_tool_call, @disallow_local_execution_in_containers decorators
2. Register command in commands.py
3. Write tests in tests/tools/

See src/linux_mcp_server/tools/processes.py for reference.

Commits & PRs

Use Conventional Commits: <type>(<scope>): <subject>

Types: feat, fix, docs, test, refactor, perf, chore

PRs must be small and focused - one logical change per PR. Split large changes into incremental PRs.

Docs

Full details: docs/contributing.md | Architecture: docs/architecture.md