SIA is a Self Improving AI framework to autonomously improve the performance of any AI system (Model / Agent) on a benchmark task.
This report summarizes scanner evidence for hexo-ai/sia, a Python project with 2,160 GitHub stars. SynthScan examined 19,735 lines of code across 120 source files, recording 288 pattern matches distributed across 14 syntactic categories. The legacy adjusted pattern signal is 24.3 (Elevated pattern signal); it is not an authorship estimate.
Use the findings to ask whether maintainers can explain intent, verify dependencies, simplify unnecessary scaffolding, and focus review effort. Risk Payload v1 adds ownership, supply-chain, over-engineering, and review-effort vectors when available. These are prompts for human verification, not probabilities, security findings, or measured maintenance costs.
Use each vector to guide manual verification of the underlying findings.
Can maintainers explain intent, assumptions, and trade-offs? 177 finding(s).
Are imports declared and package/API names verified? 37 finding(s).
Are wrappers, guards, and repeated structures necessary? 194 finding(s).
Which changes need focused human review? 288 finding(s).
This chart tracks the legacy adjusted pattern score across scan runs. Changes can reflect source changes, scan coverage, or detector/scoring versions; they do not measure AI adoption or prove that a team refactored code. Compare runs only when their scoring versions are compatible. The dashed line tracks the raw pattern-hit count.
Groups findings by configured severity weight. Severity is not a confidence estimate, security rating, or proof that a package is missing. Verify imports against the project's manifests and inspect each finding in its local design context.
This horizontal bar chart shows where configured pattern scores accumulate across directories. Treat the bars as locations to inspect, not as evidence of authorship or code quality. Use the file-level findings to decide whether a directory needs a focused ownership, dependency, complexity, or review check.
The scanner identified 288 distinct pattern matches across 14 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a detector match to inspect in context. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.
Reading the findings table: The Severity column indicates the configured finding weight (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — legacy comment-context matches receive a ×1.5 score weight. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier under the legacy scoring model.
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | sia/context_manager.py | 183 | def _format_metrics_comparison(self, prev_metrics: dict[str, Any], current_metrics: dict[str, Any]) -> str: | STRING |
| LOW | sia/prompts.py | 39 | def _build_weights_meta_prompt( | CODE |
| LOW | sia/prompts.py | 714 | def build_target_client_setup(provider: Provider, task_model: str) -> str: | STRING |
| LOW | sia/prepare_mlebench_dataset.py | 112 | def get_similar_tasks_from_gemini(competition_id: str, task_description: str) -> str: | CODE |
| LOW | sia/prepare_mlebench_dataset.py | 175 | def create_sample_task_descriptions(competition_id: str, tasks_dir: Path, similar_tasks: str) -> bool: | CODE |
| LOW | sia/orchestrator.py | 312 | def _run_target_agent_sandboxed( | STRING |
| LOW | sia/profiles.py | 103 | def load_target_agent_profile(name_or_path: str) -> TargetAgentProfile: | CODE |
| LOW | sia/agent_impls/pydantic_ai.py | 30 | def _require_provider_api_key(provider): | CODE |
| LOW | sia/agent_impls/pydantic_ai.py | 56 | def _malformed_response_message(model_name, provider, exc: JSONDecodeError) -> str: | CODE |
| LOW | …asks/longcot-chess/reference/reference_target_agent.py | 32 | def load_chess_hard_questions(dataset_dir: Path) -> List[Dict[str, Any]]: | CODE |
| LOW | tests/test_orchestrator_helpers.py | 8 | def test_load_single_trajectory(tmp_path): | CODE |
| LOW | tests/test_orchestrator_helpers.py | 18 | def test_load_multi_trajectory(tmp_path): | CODE |
| LOW | tests/test_orchestrator_helpers.py | 32 | def test_load_missing_execution(tmp_path): | CODE |
| LOW | tests/test_orchestrator_helpers.py | 45 | def test_load_empty_multi_trajectory_folder(tmp_path): | CODE |
| LOW⚡ | tests/test_agent_impls.py | 11 | def test_registry_lists_builtin_agent_impls(): | CODE |
| LOW⚡ | tests/test_agent_impls.py | 15 | def test_get_agent_impl_returns_callable(): | CODE |
| LOW⚡ | tests/test_agent_impls.py | 20 | def test_get_agent_impl_unknown_raises(): | CODE |
| LOW⚡ | tests/test_agent_impls.py | 25 | def test_util_reexports_registry_run_agent(): | CODE |
| LOW⚡ | tests/test_agent_impls.py | 32 | def test_pydantic_ai_impl_runs_with_test_model(tmp_path): | CODE |
| LOW | tests/test_agent_impls.py | 51 | def test_pydantic_ai_model_passthrough(): | CODE |
| LOW | tests/test_agent_impls.py | 61 | def test_pydantic_ai_openrouter_uses_native_provider(monkeypatch): | CODE |
| LOW | tests/test_agent_impls.py | 80 | def test_pydantic_ai_openai_compatible_provider_requires_configured_api_key(monkeypatch): | CODE |
| LOW | tests/test_agent_impls.py | 97 | def test_pydantic_ai_impl_wraps_malformed_provider_json(tmp_path, monkeypatch): | CODE |
| LOW | tests/test_agent_impls.py | 132 | def test_openhands_model_gets_openai_prefix_for_compatible_provider(): | CODE |
| LOW | tests/test_agent_impls.py | 143 | def test_openhands_model_passthrough_without_compatible_provider(): | CODE |
| LOW | tests/test_agent_impls.py | 153 | def test_run_agent_threads_provider_to_agent_impl(): | CODE |
| LOW | tests/test_agent_impls.py | 171 | def test_openhands_model_uses_provider_declared_litellm_prefix(): | CODE |
| LOW | tests/test_agent_impls.py | 193 | def test_run_agent_forwards_model_canonical_name_only_when_set(): | CODE |
| LOW | tests/test_agent_impls.py | 225 | def test_openrouter_prefix_preserves_cache_control_on_the_wire(monkeypatch): | CODE |
| LOW | tests/test_generation_loop.py | 66 | def test_run_target_agent_success(mock_popen_cls, tmp_path): | CODE |
| LOW | tests/test_generation_loop.py | 98 | def test_run_target_agent_failure(mock_popen_cls, tmp_path): | CODE |
| LOW | tests/test_generation_loop.py | 126 | def test_single_generation_creates_context(mock_run_ta, mock_run_fb, tmp_path): | CODE |
| LOW | tests/test_generation_loop.py | 165 | def test_run_generation_directory_structure(mock_run_ta, mock_run_fb, tmp_path): | CODE |
| LOW | tests/test_generation_loop.py | 194 | def test_two_generations_with_feedback(mock_run_ta, mock_run_fb, mock_llm, tmp_path): | CODE |
| LOW⚡ | tests/test_providers.py | 10 | def test_bundled_providers_present(): | CODE |
| LOW⚡ | tests/test_providers.py | 14 | def test_load_anthropic_provider(): | CODE |
| LOW⚡ | tests/test_providers.py | 23 | def test_load_nebius_provider(): | CODE |
| LOW⚡ | tests/test_providers.py | 30 | def test_unknown_provider_name_raises(): | CODE |
| LOW⚡ | tests/test_providers.py | 35 | def test_load_provider_from_path(tmp_path): | CODE |
| LOW | tests/test_providers.py | 58 | def test_invalid_client_kind_raises(tmp_path): | CODE |
| LOW | tests/test_providers.py | 65 | def test_user_dir_overrides_bundled(tmp_path, monkeypatch): | CODE |
| LOW | tests/test_providers.py | 83 | def test_litellm_prefix_is_optional_and_defaults_to_none(tmp_path): | CODE |
| LOW | tests/test_feedback_context_golden.py | 19 | def test_success_single_with_results(tmp_path): | CODE |
| LOW | tests/test_feedback_context_golden.py | 40 | def test_failure_single_no_results(tmp_path): | CODE |
| LOW | tests/test_feedback_context_golden.py | 60 | def test_success_multi_with_results(tmp_path): | CODE |
| LOW | tests/test_prompts_snapshot.py | 31 | def test_meta_prompt_anthropic_provider_is_byte_identical(): | CODE |
| LOW | tests/test_prompts_snapshot.py | 42 | def test_meta_prompt_openai_provider_golden(): | CODE |
| LOW | tests/test_prompts_snapshot.py | 53 | def test_feedback_prompt_golden(): | CODE |
| LOW | tests/test_web.py | 79 | def test_get_run_detail_and_domains(runs_root): | CODE |
| LOW⚡ | tests/test_web.py | 95 | def test_eval_details_and_artifacts(runs_root): | CODE |
| LOW⚡ | tests/test_web.py | 103 | def test_trajectory_normalization(runs_root): | CODE |
| LOW⚡ | tests/test_web.py | 112 | def test_missing_lookups_return_none(runs_root): | CODE |
| LOW⚡ | tests/test_web.py | 119 | def test_path_traversal_is_blocked(runs_root, evil): | CODE |
| LOW⚡ | tests/test_cli_interface.py | 11 | def test_top_level_help_lists_subcommands(): | CODE |
| LOW⚡ | tests/test_cli_interface.py | 19 | def test_run_help_exposes_orchestrator_flags(): | CODE |
| LOW⚡ | tests/test_cli_interface.py | 27 | def test_web_help_exposes_server_flags(): | CODE |
| LOW⚡ | tests/test_cli_interface.py | 35 | def test_no_args_exits_nonzero(): | CODE |
| LOW⚡ | tests/test_cli_interface.py | 40 | def test_default_subcommand_is_run(): | CODE |
| LOW⚡ | tests/test_cli_interface.py | 45 | def test_invalid_task_exits_nonzero(): | CODE |
| LOW | tests/test_context_manager.py | 37 | def test_initialize_creates_context_md(context_mgr, run_dir): | CODE |
| 60 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM⚡ | sia/orchestrator.py | 847 | # ======================== | COMMENT |
| MEDIUM⚡ | sia/orchestrator.py | 849 | # ======================== | COMMENT |
| MEDIUM⚡ | sia/orchestrator.py | 853 | # ======================== | COMMENT |
| MEDIUM⚡ | sia/orchestrator.py | 855 | # ======================== | COMMENT |
| MEDIUM | sia/orchestrator.py | 86 | # ======================== | COMMENT |
| MEDIUM | sia/orchestrator.py | 88 | # ======================== | COMMENT |
| MEDIUM | sia/orchestrator.py | 870 | # ======================== | COMMENT |
| MEDIUM | sia/orchestrator.py | 872 | # ======================== | COMMENT |
| MEDIUM | sia/orchestrator.py | 896 | # ======================== | COMMENT |
| MEDIUM | sia/orchestrator.py | 898 | # ======================== | COMMENT |
| MEDIUM | sia/orchestrator.py | 917 | # ======================== | COMMENT |
| MEDIUM | sia/orchestrator.py | 919 | # ======================== | COMMENT |
| MEDIUM | sia/tasks/_shared/reference_target_agent.py | 11 | # ── Tool definitions ────────────────────────────────────────────────────────── | COMMENT |
| MEDIUM⚡ | sia/tasks/_shared/reference_target_agent.py | 50 | # ── Tool implementations ────────────────────────────────────────────────────── | COMMENT |
| MEDIUM | sia/tasks/_shared/reference_target_agent.py | 101 | # ── Multi-Trajectory Logger ─────────────────────────────────────────────────── | COMMENT |
| MEDIUM | sia/tasks/_shared/reference_target_agent.py | 173 | # ── Agent loop ──────────────────────────────────────────────────────────────── | COMMENT |
| MEDIUM | sia/tasks/_shared/reference_target_agent.py | 225 | # ── Entry point ─────────────────────────────────────────────────────────────── | COMMENT |
| MEDIUM | sia/tasks/_shared/reference_target_agent.py | 234 | # ── Multi-Trajectory Usage Example ──────────────────────────────────────────── | COMMENT |
| MEDIUM⚡ | sia/tasks/gpqa/reference/reference_target_agent.py | 27 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | sia/tasks/gpqa/reference/reference_target_agent.py | 29 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | sia/tasks/gpqa/reference/reference_target_agent.py | 39 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | sia/tasks/gpqa/reference/reference_target_agent.py | 41 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | sia/tasks/gpqa/reference/reference_target_agent.py | 46 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | sia/tasks/gpqa/reference/reference_target_agent.py | 48 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | sia/tasks/gpqa/reference/reference_target_agent.py | 231 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | sia/tasks/gpqa/reference/reference_target_agent.py | 233 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM | sia/tasks/gpqa/reference/reference_target_agent.py | 11 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM | sia/tasks/gpqa/reference/reference_target_agent.py | 13 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM | sia/tasks/gpqa/reference/reference_target_agent.py | 60 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM | sia/tasks/gpqa/reference/reference_target_agent.py | 62 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM | sia/tasks/gpqa/reference/reference_target_agent.py | 111 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM | sia/tasks/gpqa/reference/reference_target_agent.py | 113 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM | sia/tasks/gpqa/reference/reference_target_agent.py | 184 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM | sia/tasks/gpqa/reference/reference_target_agent.py | 186 | # ----------------------------------------------------------------------------- | COMMENT |
| MEDIUM | …/spaceship-titanic/reference/reference_target_agent.py | 11 | # ── Tool definitions ────────────────────────────────────────────────────────── | COMMENT |
| MEDIUM⚡ | …/spaceship-titanic/reference/reference_target_agent.py | 50 | # ── Tool implementations ────────────────────────────────────────────────────── | COMMENT |
| MEDIUM | …/spaceship-titanic/reference/reference_target_agent.py | 100 | # ── Agent loop ──────────────────────────────────────────────────────────────── | COMMENT |
| MEDIUM | …/spaceship-titanic/reference/reference_target_agent.py | 152 | # ── Entry point ─────────────────────────────────────────────────────────────── | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | EVALUATION_GUIDE.md | 97 | print(f"Error: {submission} not found") | CODE |
| MEDIUM | sia/prepare_mlebench_dataset.py | 39 | print(f"Error running mlebench prepare: {e}", file=sys.stderr) | CODE |
| MEDIUM | sia/prepare_mlebench_dataset.py | 61 | print(f"Error: Prepared directory not found: {prepared_dir}", file=sys.stderr) | CODE |
| LOW | sia/prepare_mlebench_dataset.py | 170 | except Exception as e: | CODE |
| LOW | sia/orchestrator.py | 412 | except Exception as e: | STRING |
| LOW | sia/agent_impls/claude.py | 93 | except Exception as e: | CODE |
| LOW | sia/agent_impls/pydantic_ai.py | 185 | except Exception as e: | CODE |
| LOW | sia/agent_impls/openhands.py | 126 | except Exception as e: | CODE |
| LOW | …asks/longcot-chess/reference/reference_target_agent.py | 139 | except Exception as e: | CODE |
| MEDIUM⚡ | …asks/longcot-chess/reference/reference_target_agent.py | 283 | print("Error: GEMINI_API_KEY environment variable not set") | STRING |
| MEDIUM⚡ | …asks/longcot-chess/reference/reference_target_agent.py | 290 | print(f"Error: Dataset directory does not exist: {args.dataset_dir}") | STRING |
| LOW⚡ | sia/tasks/_shared/reference_target_agent.py | 57 | except Exception as e: | CODE |
| LOW⚡ | sia/tasks/_shared/reference_target_agent.py | 67 | except Exception as e: | CODE |
| LOW | sia/tasks/_shared/reference_target_agent.py | 86 | except Exception as e: | CODE |
| LOW | sia/tasks/_shared/reference_target_agent.py | 155 | except Exception as e: | CODE |
| MEDIUM⚡ | sia/tasks/gpqa/reference/reference_target_agent.py | 226 | print(f"Error on question {question_result['question_id']}: {question_result['error']}") | CODE |
| LOW | sia/tasks/gpqa/reference/reference_target_agent.py | 153 | except Exception as e: | CODE |
| LOW | sia/tasks/gpqa/reference/reference_target_agent.py | 168 | except Exception as exc: | CODE |
| LOW⚡ | …/spaceship-titanic/reference/reference_target_agent.py | 57 | except Exception as e: | CODE |
| LOW⚡ | …/spaceship-titanic/reference/reference_target_agent.py | 67 | except Exception as e: | CODE |
| LOW | …/spaceship-titanic/reference/reference_target_agent.py | 86 | except Exception as e: | CODE |
| MEDIUM | sia/tasks/lawbench/data/public/evaluate.py | 78 | print(f"Error: No submission.csv or predictions.csv found in {args.gen_dir}") | CODE |
| LOW | sia/tasks/lawbench/data/public/evaluate.py | 130 | except Exception as e: | CODE |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 202 | except Exception as e: | CODE |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 299 | except Exception as e: | CODE |
| LOW | sia/web/server.py | 128 | except Exception as exc: # pragma: no cover - background best-effort | CODE |
| MEDIUM | sia/web/server.py | 125 | def _run() -> None: | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | sia/config.py | 3 | CODE | |
| LOW | sia/results.py | 8 | CODE | |
| LOW | sia/agent_reference.py | 21 | CODE | |
| LOW | sia/agent_reference.py | 26 | CODE | |
| LOW | sia/util.py | 7 | CODE | |
| LOW | sia/util.py | 7 | CODE | |
| LOW | sia/util.py | 7 | CODE | |
| LOW | sia/layout.py | 9 | CODE | |
| LOW | sia/config_files.py | 10 | CODE | |
| LOW | sia/run_setup.py | 7 | CODE | |
| LOW | sia/prompts.py | 7 | CODE | |
| LOW | sia/providers.py | 12 | CODE | |
| LOW | sia/cli.py | 20 | CODE | |
| LOW | sia/orchestrator.py | 59 | CODE | |
| LOW | sia/io_utils.py | 9 | CODE | |
| LOW | sia/api_keys.py | 3 | CODE | |
| LOW | sia/profiles.py | 17 | CODE | |
| LOW | sia/logging_setup.py | 8 | CODE | |
| LOW | sia/agent_impls/claude.py | 3 | CODE | |
| LOW | sia/agent_impls/__init__.py | 8 | CODE | |
| LOW | sia/agent_impls/__init__.py | 8 | CODE | |
| LOW | sia/agent_impls/__init__.py | 8 | CODE | |
| LOW | sia/agent_impls/__init__.py | 9 | CODE | |
| LOW | sia/agent_impls/__init__.py | 9 | CODE | |
| LOW | sia/agent_impls/__init__.py | 9 | CODE | |
| LOW | sia/agent_impls/__init__.py | 9 | CODE | |
| LOW | sia/agent_impls/pydantic_ai.py | 12 | CODE | |
| LOW | sia/agent_impls/openhands.py | 3 | CODE | |
| LOW | sia/agent_impls/base.py | 11 | CODE | |
| LOW | sia/web/server.py | 8 | CODE | |
| LOW | sia/web/__init__.py | 9 | CODE | |
| LOW | sia/web/__init__.py | 9 | CODE | |
| LOW | sia/web/__init__.py | 9 | CODE | |
| LOW | sia/web/runs.py | 24 | CODE | |
| LOW | tests/golden_master.py | 16 | CODE | |
| LOW | tests/test_gpqa_evaluator.py | 3 | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | sia/prompts.py | 351 | ### Step 1: Verify Dataset Alignment | COMMENT |
| LOW⚡ | sia/prompts.py | 358 | ### Step 2: Check the Sandbox (if using code execution) | COMMENT |
| LOW | sia/prompts.py | 400 | ### Step 3: Isolate the Problem | COMMENT |
| LOW⚡ | sia/prepare_mlebench_dataset.py | 237 | # Step 1: Run mlebench prepare | COMMENT |
| LOW⚡ | sia/prepare_mlebench_dataset.py | 242 | # Step 2: Copy dataset files | COMMENT |
| LOW⚡ | sia/prepare_mlebench_dataset.py | 247 | # Step 3: Move and rename description | COMMENT |
| LOW | sia/prepare_mlebench_dataset.py | 265 | # Step 6: Copy reference agent | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 97 | # Step 1: Load test data | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 109 | # Step 2: Load valid classes | COMMENT |
| LOW⚡ | sia/tasks/lawbench/reference/reference_target_agent.py | 122 | # Step 3: Load training data for few-shot examples (optional) | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 135 | # Step 4: Generate predictions with multi-trajectory logging | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 239 | # Step 5: Save submission | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 262 | # Step 6: Save summary execution log | COMMENT |
| LOW | docs/walkthrough.md | 5 | ## Step 1: Set up the task directory | COMMENT |
| LOW | docs/walkthrough.md | 41 | ## Step 2: Run the orchestrator | COMMENT |
| LOW | docs/walkthrough.md | 68 | ## Step 3: Analyze results | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | sia/prompts.py | 351 | ### Step 1: Verify Dataset Alignment | COMMENT |
| LOW⚡ | sia/prompts.py | 358 | ### Step 2: Check the Sandbox (if using code execution) | COMMENT |
| LOW | sia/prompts.py | 400 | ### Step 3: Isolate the Problem | COMMENT |
| LOW⚡ | sia/prepare_mlebench_dataset.py | 237 | # Step 1: Run mlebench prepare | COMMENT |
| LOW⚡ | sia/prepare_mlebench_dataset.py | 242 | # Step 2: Copy dataset files | COMMENT |
| LOW⚡ | sia/prepare_mlebench_dataset.py | 247 | # Step 3: Move and rename description | COMMENT |
| LOW | sia/prepare_mlebench_dataset.py | 265 | # Step 6: Copy reference agent | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 97 | # Step 1: Load test data | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 109 | # Step 2: Load valid classes | COMMENT |
| LOW⚡ | sia/tasks/lawbench/reference/reference_target_agent.py | 122 | # Step 3: Load training data for few-shot examples (optional) | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 135 | # Step 4: Generate predictions with multi-trajectory logging | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 239 | # Step 5: Save submission | COMMENT |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 262 | # Step 6: Save summary execution log | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | sia/context_manager.py | 203 | except (ValueError, TypeError): | CODE |
| HIGH | sia/orchestrator.py | 246 | except (json.JSONDecodeError, OSError): | CODE |
| HIGH | sia/orchestrator.py | 420 | except OSError: | CODE |
| HIGH | sia/tasks/gpqa/reference/reference_target_agent.py | 91 | except json.JSONDecodeError: | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | sia/context_manager.py | 267 | CODE | |
| LOW | sia/context_manager.py | 338 | CODE | |
| LOW | sia/context_manager.py | 385 | CODE | |
| LOW | sia/context_manager.py | 454 | CODE | |
| LOW | sia/orchestrator.py | 91 | CODE | |
| LOW | sia/orchestrator.py | 185 | CODE | |
| LOW | sia/orchestrator.py | 425 | CODE | |
| LOW | sia/agent_impls/claude.py | 13 | CODE | |
| LOW | …asks/longcot-chess/reference/reference_target_agent.py | 93 | CODE | |
| LOW | sia/tasks/_shared/reference_target_agent.py | 175 | CODE | |
| LOW | sia/tasks/gpqa/data/public/evaluate.py | 111 | CODE | |
| LOW | sia/tasks/gpqa/reference/reference_target_agent.py | 114 | CODE | |
| LOW | …/spaceship-titanic/reference/reference_target_agent.py | 102 | CODE | |
| LOW | sia/tasks/lawbench/data/public/evaluate.py | 50 | CODE | |
| LOW | sia/tasks/lawbench/data/public/evaluate.py | 119 | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | sia/context_manager.py | 372 | # Priority 3: Parse stdout (check both harness and weights mode log files) | COMMENT |
| MEDIUM | sia/orchestrator.py | 885 | # Build meta prompt based on focus mode (weights=RL tuning, harness=code improvement) | COMMENT |
| MEDIUM | sia/orchestrator.py | 662 | # Use train.py for weights mode (RL tuning), target_agent.py for harness mode | STRING |
| MEDIUM | sia/web/runs.py | 299 | # For stdout_log, check both harness and weights mode variants | COMMENT |
| MEDIUM | sia/web/runs.py | 365 | # For stdout_log, check both harness and weights mode variants | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | sia/profiles.py | 41 | # Canonical model id used only for SDK capability lookups (prompt caching, context window, | COMMENT |
| MEDIUM | sia/agent_impls/openhands.py | 84 | # would drop it and leave capability lookups (prompt caching, context window, cost) | COMMENT |
| MEDIUM⚡ | sia/tasks/lawbench/reference/reference_target_agent.py | 122 | # Step 3: Load training data for few-shot examples (optional) | COMMENT |
| MEDIUM | docs/architecture.md | 33 | │ ├── orchestrator.py # Main orchestration logic | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | sia/util.py | 9 | __all__ = ["available_agent_impls", "get_agent_impl", "run_agent"] | CODE |
| LOW | sia/orchestrator.py | 68 | __all__ = [ | CODE |
| LOW | sia/agent_impls/__init__.py | 11 | __all__ = ["available_agent_impls", "get_agent_impl", "register", "run_agent"] | CODE |
| LOW | sia/tasks/lawbench/reference/reference_target_agent.py | 27 | logger = logging.getLogger(__name__) | CODE |
| LOW | sia/web/__init__.py | 11 | __all__ = ["create_app", "serve", "serve_in_background"] | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | sia/context_manager.py | 142 | # Create a temporary directory for LLM execution | STRING |
| MEDIUM | tests/test_context_manager.py | 17 | # Create a minimal target_agent.py | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH⚡ | …asks/longcot-chess/reference/reference_target_agent.py | 285 | print(" export GEMINI_API_KEY='your-api-key-here'") | STRING |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | sia/orchestrator.py | 235 | # Check if results.json was created | COMMENT |
| LOW | sia/tasks/longcot-chess/data/public/evaluate.py | 100 | # Check if any permutation matches (for unordered answers) | COMMENT |