mirror of
https://github.com/OpenHands/OpenHands.git
synced 2026-03-22 13:47:19 +08:00
77f868081c
Create OpenHandsAgentSettings(AgentSettings) in the OpenHands codebase that extends the SDK's AgentSettings with a 'security' section containing confirmation_mode (critical) and security_analyzer (major). The SDK's export_schema() picks these up automatically via its metadata conventions. Backend: - SecuritySettings pydantic model with SDK metadata annotations - OpenHandsAgentSettings subclass used by _get_sdk_settings_schema() - _SDK_TO_FLAT_SETTINGS bridges dotted SDK keys to flat Settings attrs so existing consumers (session init, security-analyzer setup) work - _extract_sdk_settings_values seeds from flat fields for UI display Frontend: - /settings/security route renders the security schema section - Nav: LLM -> Security -> Condenser -> Critic (both SAAS and OSS) - Removed empty General page (no schema section exists for it yet) Tests: - New test_get_sdk_settings_schema_includes_security_section - All 119 backend + 10 frontend tests pass Co-authored-by: openhands <openhands@all-hands.dev>
Introduction
This folder contains unit tests that could be run locally.
Run all test:
poetry run pytest ./tests/unit
Run specific test file:
poetry run pytest ./tests/unit/test_llm_fncall_converter.py
Run specific unit test
poetry run pytest ./tests/unit/test_llm_fncall_converter.py::test_convert_tool_call_to_string
For a more verbose output, to above calls the -v flag can be used (even more verbose: -vv and -vvv):
poetry run pytest -v ./tests/unit/test_llm_fncall_converter.py
More details see pytest doc