MCPcopy Create free account

hub / github.com/HKUDS/Vibe-Trading / functions

Functions26,887 in github.com/HKUDS/Vibe-Trading

Functiontest_tamper_editing_one_record_content_is_pinpointed
The core negative case: hand-edit ONE record's payload, leave its stored seq/prev_record_hash/record_hash untouched, and assert verify_chain n
agent/tests/test_governance.py:303
Functiontest_tamper_that_also_fixes_its_own_hash_is_still_caught_downstream
A more sophisticated tamper: edit a record's payload AND recompute that record's own record_hash so it is internally self-consistent. The chai
agent/tests/test_governance.py:353
Functiontest_tampered_callable_key_is_corrected_to_spec
(tmp_path: Path)
agent/tests/scheduling/test_system_jobs.py:324
Functiontest_tampered_callable_name_takes_dispatch_failure_path
Break-probe as a test: a job file tampered to name a real import path (a stdlib process-spawning function) is rejected by the allowlist; the p
agent/tests/scheduling/test_internal_dispatch.py:296
Functiontest_tampered_executor_kind_prompt_is_corrected
QA (b): user hand-edits executor_kind to 'prompt' → corrected to 'internal' (and the callable key restored), so the row can never be prompt-di
agent/tests/scheduling/test_system_jobs.py:300
Functiontest_tampered_manifest_absolute_path_refused
(run_dir: Path, tmp_path: Path)
agent/tests/test_run_artifact_tool.py:284
Functiontest_tampered_manifest_escape_entries_refused
( run_dir: Path, tmp_path: Path, tampered: str )
agent/tests/test_run_artifact_tool.py:270
Functiontest_tampered_manifest_is_invalid_artifact
( harness_case: HarnessCase, artifact_dir: Path, )
agent/tests/harness_eval/test_runner.py:142
Functiontest_tampered_pickle_with_unkeyed_sha256_is_rejected
(tmp_path, no_api_key)
agent/tests/test_alpha_bench_cache.py:52
Functiontest_tampered_poll_url_is_rejected_and_never_contacted
(tap_env, monkeypatch, poll_url)
agent/tests/test_tap_forward.py:203
Functiontest_tap_credential_name_is_overridable
(monkeypatch)
agent/tests/test_alpaca_tap_routing.py:126
Functiontest_tap_disabled_does_not_route_through_tap
(monkeypatch)
agent/tests/test_alpaca_tap_routing.py:141
Functiontest_target_day_all_market_zero_bar_classifies_as_not_ready
()
agent/tests/test_paper_trade_readiness.py:260
Functiontest_target_day_all_market_zero_bar_returns_not_ready_material
()
agent/tests/test_paper_trade_rejections.py:409
Functiontest_target_day_without_a_bar_returns_in_band_not_ready
()
agent/tests/test_paper_trade_tool.py:174
Functiontest_target_eps_basis_must_match_peers
()
agent/tests/quantlib/valuation/test_comps.py:425
Functiontest_target_negative_metric_produces_negative_implied_value_without_raising
A target with negative LTM EPS produces a negative implied equity value via P/E -- economically odd, but the module does not fabricate an excl
agent/tests/quantlib/valuation/test_comps.py:517
Functiontest_targets_opencode
(provider: str, base_url: str | None, expected: bool)
agent/tests/test_opencode_provider.py:134
Functiontest_task_failed_event_does_not_leak_abs_path
task_failed event payload (surfaced via read_events/SSE/get_swarm_status) must redact the internal absolute path. Pre-fix: raw result.error → FAIL
agent/tests/test_error_path_redaction.py:87
Functiontest_task_name_is_injectable
()
agent/tests/scheduling/test_kernel.py:80
Functiontest_task_property_on_the_facade_is_read_only
``_task`` was a writable attribute before the kernel adoption; as a property it must reject assignment so owners cannot bypass the kernel's gu
agent/tests/test_runtime_scheduler_kernel.py:188
Functiontest_task_property_tracks_the_kernel_task_lifecycle
``live_routes._runner_scheduler_task`` reads ``scheduler._task`` and treats None/done as "nothing to await" — the property must keep that cont
agent/tests/test_runtime_scheduler_kernel.py:52
Functiontest_task_property_tracks_the_loop_lifecycle
()
agent/tests/scheduling/test_kernel_sleep_hint.py:147
Functiontest_technical_indicators_opt_into_readonly_compaction_replay
The real technical-indicator tool can restore an identical lost result. Technical indicators are read-only but repeatable, so the generic replay
agent/tests/test_technical_indicator_compaction_replay.py:15
Functiontest_telegram_bot_token_in_request_url_is_redacted
()
agent/tests/test_redaction_channel_and_provider_tokens.py:11
Functiontest_telegram_status_suggests_numeric_authorized_private_chats
()
agent/tests/test_channels_runtime.py:207
Functiontest_telegram_token_does_not_assume_an_eight_to_ten_digit_id
(bot_id, in_url)
agent/tests/test_redaction_channel_and_provider_tokens.py:47
Methodtest_temperature_is_omitted_when_effort_is_sent
Effort enables adaptive thinking, which rejects temperature != 1. The platform's default is 0.0, so sending both fails the request with
agent/tests/test_llm_reasoning_effort.py:473
Methodtest_temperature_is_preserved_without_effort
No effort means no thinking, so the deterministic default stands.
agent/tests/test_llm_reasoning_effort.py:485
Methodtest_template_file_exists
(self)
agent/tests/test_vnpy_export.py:85
Methodtest_template_is_valid_python
(self)
agent/tests/test_vnpy_export.py:88
Functiontest_tencent_a_share_basket_warns_against_multiplicative_sources
The label collision was the bug: with both sources on `split_dividend` the warning returned None for a basket that mixes them.
agent/tests/test_price_caliber.py:95
Functiontest_tencent_a_share_caliber_is_additive_not_multiplicative
A-share qfq subtracts cash dividends from the level, not by a ratio.
agent/tests/test_price_caliber.py:80
Functiontest_tencent_and_eastmoney_a_share_basket_is_not_a_mix
Both additive with the same offsets: warning them apart was a false positive.
agent/tests/test_price_caliber.py:118
Methodtest_tencent_converts_and_stamps
(self, monkeypatch)
agent/tests/test_additive_conversion.py:169
Functiontest_tencent_execution_adapter_reads_only_the_raw_endpoint
(monkeypatch)
agent/tests/test_paper_trade_production_caliber.py:55
Functiontest_tencent_hk_series_is_unadjusted
Tencent serves no adjusted HK series at all: its fqkline reply carries only "day" for HK symbols (never "qfqday"/"hfqday"), so the loader's ``qfqd
agent/tests/test_price_caliber.py:85
Methodtest_tencent_keeps_additive_when_raw_missing
(self, monkeypatch)
agent/tests/test_additive_conversion.py:179
Functiontest_tencent_only_basket_stays_silent
One additive source is not a mix, so nothing to warn about.
agent/tests/test_price_caliber.py:128
Methodtest_tencent_raw_failure_serves_additive
(self, monkeypatch)
agent/tests/test_additive_conversion.py:268
Methodtest_tencent_skips_etf_codes
(self, monkeypatch)
agent/tests/test_additive_conversion.py:215
Functiontest_terminal_close_can_sell_acquired_shares_before_settlement
(tmp_path)
agent/tests/test_argentina_equity_engine.py:260
Functiontest_terminal_close_costs_reach_final_equity
( target_weight: float, expected_exit: float, )
agent/tests/test_terminal_close_accounting.py:43
Functiontest_terminal_close_preserves_the_cedear_credit_window
(tmp_path)
agent/tests/test_argentina_equity_engine.py:293
Functiontest_terminal_growth_at_or_above_wacc_raises
()
agent/tests/quantlib/valuation/test_dcf.py:451
Methodtest_terminal_order_cannot_be_canceled
(self)
agent/tests/test_paper_order_lifecycle.py:125
Methodtest_terminal_orders_are_left_alone
(self)
agent/tests/test_paper_order_lifecycle.py:157
Methodtest_terminal_outcomes_are_kept_queryable
(self, status)
agent/tests/test_background_store_recovery.py:66
Methodtest_terminal_path_attaches_without_the_warning
( self, fake_mod: _FakeMT5Module, tmp_path: Path, monkeypatch: pytest.MonkeyPatch, caplog: pytest.LogC
agent/tests/test_mt5_loader.py:252
Functiontest_terminal_policy_uses_only_the_latest_mark
()
agent/tests/quantlib/test_fundmath.py:245
Functiontest_terminal_refusals_fail_closed
(tap_env, monkeypatch, decision)
agent/tests/test_tap_forward.py:159
Functiontest_terminal_value_blows_up_near_wacc_without_silent_truncation
As g approaches WACC the terminal value must explode, not be capped or silently truncated to a merely-large number.
agent/tests/quantlib/valuation/test_dcf.py:470
Functiontest_terminal_value_method_switch_changes_enterprise_value
()
agent/tests/quantlib/valuation/test_dcf.py:200
Functiontest_terminal_value_refuses_non_finite_inputs
()
agent/tests/quantlib/valuation/test_dcf.py:1088
Methodtest_terminal_value_sheet_cells_match_result
(self, workbook_path)
agent/tests/quantlib/valuation/test_artifact.py:633
Functiontest_tesouro_and_number_helpers
()
agent/tests/test_br_data.py:184
Functiontest_test_connection_envelope_shape_is_documented_by_default
()
agent/tests/test_channel_authoring_contract.py:84
Functiontest_test_endpoint_returns_no_secret_and_save_then_requires_confirmation
( client: TestClient, custom_store: CustomProviderStore, monkeypatch: pytest.MonkeyPatch, )
agent/tests/test_custom_provider_routes.py:117
Functiontest_test_registrations_do_not_leak_between_tests
()
agent/tests/scheduling/test_registry.py:198
Methodtest_tests_dir_skipped
(self)
tools/test_ci_env_var_gate.py:168
Functiontest_text_delta_maps_to_single_message_chunk
()
agent/tests/test_openbb_bridge/test_event_mapper.py:12
Functiontest_text_field_offload_round_trips_selected_fields
Long prompt/answer fields can be offloaded without forcing all blobs open.
agent/tests/test_trace_writer.py:96
Functiontest_text_only_end_to_end_keeps_content_and_text
()
agent/tests/test_mcp_result_dedup.py:122
Functiontest_text_only_json_lookalike_without_data_is_not_touched
()
agent/tests/test_mcp_result_dedup.py:116
Functiontest_text_only_mcp_response_is_supported
(adapter)
agent/tests/test_parallel_search.py:270
Functiontest_text_only_payload_is_returned_unchanged
()
agent/tests/test_mcp_result_dedup.py:110
Functiontest_text_that_is_not_the_derived_join_is_kept
()
agent/tests/test_mcp_result_dedup.py:166
Methodtest_thai_characters
(self)
agent/tests/test_persistent_memory.py:103
Functiontest_the_active_run_named_by_a_stand_in_is_the_file_named_bare
``run_dir/artifacts/metrics.csv`` grounded nothing while ``metrics.csv`` grounded the same draft: the final value, trade count and trade PnL were
agent/tests/test_grounding_backtest_outputs.py:954
Functiontest_the_active_runs_copy_belongs_to_the_backtest_the_archive_names
The loop copies each finished backtest into the active run; a ref to that copy names the backtest it came from, and only once it has come. ``run_d
agent/tests/test_grounding_backtest_outputs.py:525
Functiontest_the_ambiguity_correction_names_the_refs_to_use
()
agent/tests/test_grounding_role_hardening.py:863
Methodtest_the_auto_route_asks_for_daily_bars_too
(self, monkeypatch, recording_loader)
agent/tests/test_weekly_monthly_bars.py:137
Methodtest_the_backoff_schedule_crosses_the_quota_window
(self)
agent/tests/test_tushare_loader.py:450
Functiontest_the_band_does_not_apply_to_hold_mode
Hold mode never resizes, so a band there would be a silent no-op.
agent/tests/test_rebalance_tolerance.py:90
Methodtest_the_benchmark_is_built_the_same_way
(self, recording_loader)
agent/tests/test_weekly_monthly_bars.py:162
Functiontest_the_best_alpha_named_by_the_block_is_the_best_alpha_in_the_rows
(monkeypatch)
agent/tests/test_bench_multiple_testing.py:73
Methodtest_the_block_is_identical_for_different_user_messages
(self)
agent/tests/test_agent_output_discipline.py:128
Methodtest_the_block_is_static_template_text_with_no_substitution
No format field lands inside the block, so no input can reach it.
agent/tests/test_agent_output_discipline.py:121
Functiontest_the_block_is_stripped_from_the_released_text
The contract is between the model and the gate; the user never sees it.
agent/tests/test_grounding_figures_contract.py:220
Functiontest_the_block_parses_the_spec_example
Every role, every optional field, exactly as §2 writes it.
agent/tests/test_grounding_figures_contract.py:103
Methodtest_the_block_sits_before_any_per_session_content
It is in the cacheable prefix, ahead of the timestamp. Per-run workspace state no longer reaches the system prompt at all — it rides t
agent/tests/test_agent_output_discipline.py:139
Functiontest_the_briefing_is_the_reply_the_user_would_see
The last terminal reply, so the channel and the app never disagree.
agent/tests/test_scheduled_delivery_wiring.py:65
Functiontest_the_bundle_agrees_with_the_standalone_functions
(rows)
agent/tests/quantlib/test_fundmath.py:360
Functiontest_the_bundle_reports_the_same_identity
(rows)
agent/tests/quantlib/test_fundmath.py:353
Functiontest_the_calendar_is_snapshotted_once_and_passed_to_the_engine
(monkeypatch)
agent/tests/test_paper_trade_5m_scope.py:274
Methodtest_the_candidate_list_stays_bounded
(self)
agent/tests/test_institutional_holdings_tool.py:1601
Functiontest_the_chain_continues_across_a_rotation
(tmp_path)
agent/tests/test_governance.py:727
Functiontest_the_chinese_prompt_example_passes_the_gate_it_describes
The example sits on the same two bars this module probes with.
agent/tests/test_grounding_figures_contract.py:752
Functiontest_the_code_range_gives_the_currency_the_venue_quotes
(code: str, currency: str)
agent/tests/test_hk_counter_currency.py:30
Functiontest_the_composite_summary_reports_no_ic_std_either
composite_ic is a plain dict, so None survives instead of becoming NaN. ``insufficient="equal"`` gives every row a weight, so the composite is
agent/tests/quantlib/test_combination_single_ic_std.py:60
Methodtest_the_config_accepts_both
(self)
agent/tests/test_weekly_monthly_bars.py:168
Functiontest_the_connection_store_refuses_a_control_character_in_an_account
(tmp_path)
agent/tests/test_robinhood_portfolio_source.py:288
Functiontest_the_connection_store_refuses_an_account_on_a_profile_that_takes_none
(tmp_path)
agent/tests/test_robinhood_portfolio_source.py:280
Functiontest_the_correction_asks_for_the_answer_alone
A live second draft opened with "The rejection was because…" — in English, to a user who wrote Chinese — and that sentence was released.
agent/tests/test_grounding_backtest_outputs.py:836
Methodtest_the_correction_names_the_way_out
(self, tmp_path: Path)
agent/tests/test_agent_output_discipline.py:289
Functiontest_the_correction_prompt_names_each_figure_and_its_reason
§6: the feedback is per number, not a restatement of the policy. Three things per figure — what was written, what was declared, what the evid
agent/tests/test_grounding_figures_contract.py:615
Functiontest_the_correction_prompt_names_the_figures_to_keep
A mixed rejection must not talk the model into deleting clean figures. zeus229's production run on #1622: 65 issues over a draft whose evidence
agent/tests/test_grounding_release.py:2368
Functiontest_the_crypto_transactions_shape_from_the_report_is_read
The Bitpanda transactions export identifies crypto by bare symbol only. Ground truth from the anonymized samples: the same asset is ``BTC_to_EUR`
agent/tests/test_extraetf_reader.py:714
Functiontest_the_declared_quote_currency_is_the_one_an_answer_must_name
A venue can list one issuer in two currencies (GGAL.BA ARS, GGALD.BA USD, #1566).
agent/tests/test_agent_grounding.py:3594
Functiontest_the_default_band_is_zero_and_changes_nothing
Existing runs must not be rewritten by the introduction of the band.
agent/tests/test_rebalance_tolerance.py:56
← previousnext →25,101–25,200 of 26,887, ranked by callers