{ "schema_version": 1, "builder_plan": { "scenario": "An agent looks good in demos, but you cannot tell which product tasks it fails.", "change": "Use the study's evaluation idea to inspect a small set of tasks with human-checked answers.", "baseline": "Your current evaluator and a manually reviewed reference set.", "steps": [ "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.", "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.", "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.", "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test." ], "measure": "Agreement with human review, missed failures, false alarms, and evaluation cost per task.", "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation." }, "builder_review": null, "arxiv_id": "2610.11482", "title": "Evaluating Local Language Model Agents for Reproducible Data Engineering: An Empirical Software Engineering Study of Mobility Workflows", "summary": "The paper replaces one-shot prompt-and-hope code generation for data-engineering tasks with a closed-loop agent setup plus deterministic checkers, and provides a 15-task benchmark for validating complete agent configurations before adoption. The cost is additional runtime and orchestration complexity: the agent must iterate over intermediate artifacts, each model/mode/task configuration needs repeated scored runs, and tasks must be instrumented with deterministic validators to expose errors. It can fail when tasks lack verifiable intermediate artifacts, when the local model is too small (models at or below 2B parameters show little benefit), or when quantized configurations degrade capability below the reliability threshold the workflow requires.", "rationale": "The closed-loop workspace pattern runs on a single consumer GPU and shows large, reproducible quality gains, but the reader should validate it against their own tasks because the gains depend heavily on task verifiability and model capability.", "technique": "Benchmark of local LLM agents with closed-loop workspace for data engineering", "family": "avaliacao_de_agente", "recommendation": "testar", "published": "2026-10-08", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.11482", "independent_implementations": 1, "total_implementations": 2, "stars": 14, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "The paper replaces one-shot prompt-and-hope code generation for data-engineering tasks with a closed-loop agent setup plus deterministic checkers, and provides a 15-task benchmark for validating complete agent configurations before adoption. The cost is additional runtime and orchestration complexity: the agent must iterate over intermediate artifacts, each model/mode/task configuration needs repeated scored runs, and tasks must be instrumented with deterministic validators to expose errors. It can fail when tasks lack verifiable intermediate artifacts, when the local model is too small (models at or below 2B parameters show little benefit), or when quantized configurations degrade capability below the reliability threshold the workflow requires.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11482", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "The closed-loop workspace condition raises artifact pass rates by 26.7-52.0 percentage points over one-shot generation for models above 2B parameters; the best configuration reaches 85.3% artifact-level success, and a quantized 9B model reaches 69.3% with about 6.5 GB of memory.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11482", "source_page": null, "source_excerpt": "" } ], "exposure_map": [ { "dimension": "quality", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compute", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "latency", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "operations", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "compatibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "security", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "data_and_training", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" }, { "dimension": "reproducibility", "basis": "not_evaluated", "finding": "", "interpretation": "", "limitation": "", "next_check": "", "source_url": "", "source_page": null, "source_excerpt": "" } ], "risks": [], "minimum_test": [], "open_questions": [], "independent_tests": [], "lab_reproductions": [], "report_available": false, "validation_tier": "unknown", "evidence_tier": "unknown", "infrastructure_basis": "unknown", "equations": [], "equations_status": "not_fetched", "core_kind": "", "equations_fetched_at": "", "equations_source_url": "", "equations_source_sha256": "", "exposures_reviewed_at": "", "exposures_source_sha256": "" }