Tests: ragleap-agents

Tests for ragleap_agents.agent. A scripted fake model replaces the LLM and the tools are plain in-process Tool objects, so nothing here touches a network.

• test_no_tools_or_empty_task_makes_no_model_call passed
  No tools or empty task makes no model call (no description in the source; shown from the name)
• test_model_says_done_immediately passed
  Model says done immediately (no description in the source; shown from the name)
• test_done_with_an_answer_is_used_when_no_step_was_taken passed
  Done with an answer is used when no step was taken (no description in the source; shown from the name)
• test_prose_instead_of_json_ends_the_run_as_done passed
  Prose instead of json ends the run as done (no description in the source; shown from the name)
• test_single_step_then_done_feeds_the_result_back_and_summarises passed
  Single step then done feeds the result back and summarises (no description in the source; shown from the name)
• test_summarize_off_and_summary_failure_use_the_fallback passed
  Summarize off and summary failure use the fallback (no description in the source; shown from the name)
• test_observation_cannot_close_its_own_fence passed
  Observation cannot close its own fence (no description in the source; shown from the name)
• test_task_cannot_close_the_request_fence passed
  Task cannot close the request fence (no description in the source; shown from the name)
• test_step_cap_and_hard_cap passed
  Step cap and hard cap (no description in the source; shown from the name)
• test_the_same_action_twice_stops_the_run passed
  The same action twice stops the run (no description in the source; shown from the name)
• test_invalid_proposals_stop_the_run_without_running_a_tool passed
  Invalid proposals stop the run without running a tool (no description in the source; shown from the name)
• test_observations_are_capped_and_omitted_when_the_prompt_budget_is_spent passed
  Observations are capped and omitted when the prompt budget is spent (no description in the source; shown from the name)
• test_deadline_stops_before_any_model_call passed
  Deadline stops before any model call (no description in the source; shown from the name)
• test_tool_exception_text_never_reaches_the_model_or_the_result passed
  Tool exception text never reaches the model or the result (no description in the source; shown from the name)
• test_model_failure_is_reported_without_its_text passed
  Model failure is reported without its text (no description in the source; shown from the name)
• test_taint_forces_approval_for_outbound_tools passed
  Taint forces approval for outbound tools (no description in the source; shown from the name)
• test_without_taint_an_outbound_tool_runs_normally passed
  Without taint an outbound tool runs normally (no description in the source; shown from the name)
• test_undeclared_tools_are_treated_as_tainting_and_outbound passed
  Undeclared tools are treated as tainting and outbound (no description in the source; shown from the name)
• test_requires_approval_pauses_and_an_approval_runs_the_real_call passed
  Requires approval pauses and an approval runs the real call (no description in the source; shown from the name)
• test_rejection_ends_the_run_without_calling_the_model_or_the_tool passed
  Rejection ends the run without calling the model or the tool (no description in the source; shown from the name)
• test_an_approved_tainting_tool_taints_the_resumed_run passed
  An approved tainting tool taints the resumed run (no description in the source; shown from the name)
• test_resume_misuse_is_refused passed
  Resume misuse is refused (no description in the source; shown from the name)
• test_resume_claims_the_run_so_a_second_resume_during_execution_is_refused passed
  Resume claims the run so a second resume during execution is refused (no description in the source; shown from the name)
• test_a_paused_run_can_be_resumed_by_a_new_agent_sharing_the_store passed
  A paused run can be resumed by a new agent sharing the store (no description in the source; shown from the name)
• test_construction_is_validated passed
  Construction is validated (no description in the source; shown from the name)
• test_validate_arguments passed
  Validate arguments (no description in the source; shown from the name)
• test_parse_plan_tolerates_fences_and_prose_but_not_garbage passed
  Parse plan tolerates fences and prose but not garbage (no description in the source; shown from the name)

No file-level description in the source.

• test_version_matches_pyproject passed
  Version matches pyproject (no description in the source; shown from the name)
• test_only_one_import_package_is_shipped passed
  ragleap-rag installs the import package `ragleap`; a second package of that