test(multi_agents): comprehensive test suite for orchestrator, workflow agent, and data types (+73 tests)

[manus-use]

Summary

Comprehensive test suite for the manus_agent.multi_agents package — 73 fully-mocked tests covering all public types, the WorkflowAgent class, and the Orchestrator class.

What's tested

Component	Tests	Coverage
AgentType enum	10	Values, str subclass, lookup, iteration
ComplexityLevel enum	7	Values, str subclass, lookup, invalid
TaskPlan dataclass	14	Construction, defaults, immutability (frozen), equality, mutable-field unhashability
OrchestratorResult dataclass	8	Construction, defaults, mutability, equality
WorkflowAgent	10	Init (default/custom model), system_prompt content, Agent delegation, exception propagation
Orchestrator	14	Init (config, model_name), run success/failure paths, str conversion, multiple calls
Module exports	4	__all__ completeness, importability
Integration scenarios	3	End-to-end Orchestrator → WorkflowAgent → Agent mock chain
main() function	2	Banner output, handle_request invocation

Why this matters

The multi_agents package is the core orchestration layer used by the CLI's multi-agent mode (Orchestrator is imported directly in cli.py). It had zero dedicated test coverage — any refactor or enhancement risked silent regressions.

Test approach

• 100% mocked — no real LLM/API calls
• Tests the public contract (construction, defaults, delegation, error handling)
• Verifies frozen dataclass semantics (immutability, equality)
• Tests the Orchestrator's exception-catching resilience
• Confirms module __all__ exports match actual definitions

Checked against existing PRs (no overlap)

• test(tools): comprehensive test suite for manus_workflow module (+58 tests) #150 — tests manus_workflow.py (different module: tools/manus_workflow.py)
• test(tools): comprehensive test suite for workflow_tool module (+60 tests) #148 — tests workflow_tool.py (different module: tools/workflow_tool.py)
• test(tools): comprehensive test suite for code_execute module (+45 tests) #141 — tests code_execute (unrelated)
• test(sandbox): comprehensive test suite for docker_sandbox module (+69 tests) #143, test(sandbox): comprehensive test suite for exploit_sandbox module (+76 tests) #144 — sandbox tests (unrelated)

No existing open or merged PR covers the multi_agents/__init__.py orchestrator types or the multi_agents/workflow_agent.py WorkflowAgent class.

Test results

73 passed in 1.35s (test_multi_agents.py)
1231 passed, 3 deselected, 3 warnings in 25.60s (full suite)