{
  "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": "2610.23284",
  "title": "Annotrace: A Workbench for Step-Level Annotation of Agent Traces",
  "summary": "Annotrace replaces the per-study practice of hand-building throwaway scripts to label agent failure traces, importing six trace formats (OpenTelemetry GenAI, LangSmith, Langfuse among them) into an ordered, nested step list with a user-defined codebook, AI-proposed starting labels, and inter-annotator agreement measurement between human and model annotators. It costs only local tooling plus frontier-model API calls for label proposal and pre-annotation, with no GPU, cluster, or training requirement, fitting the constrained infrastructure directly. It can fail if the model-proposed codebook biases the human codebook toward the model's own error categories, if AI pre-answers anchor human judgment and inflate measured agreement, and if the six supported formats do not cover the team's actual tracing stack.",
  "rationale": "Structured failure labelling is a prerequisite for debugging and evaluating production agents, and this replaces bespoke annotation scripts at negligible infrastructure cost, with the caveat that agreement metrics must be checked against anchoring bias.",
  "technique": "Step-level annotation workbench for agent traces",
  "family": "avaliacao_de_agente",
  "recommendation": "adotar",
  "published": "2026-10-16",
  "as_of": "2026-10-19",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2610.23284",
  "independent_implementations": 0,
  "total_implementations": 1,
  "stars": 10,
  "citations": null,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "Annotrace replaces the per-study practice of hand-building throwaway scripts to label agent failure traces, importing six trace formats (OpenTelemetry GenAI, LangSmith, Langfuse among them) into an ordered, nested step list with a user-defined codebook, AI-proposed starting labels, and inter-annotator agreement measurement between human and model annotators. It costs only local tooling plus frontier-model API calls for label proposal and pre-annotation, with no GPU, cluster, or training requirement, fitting the constrained infrastructure directly. It can fail if the model-proposed codebook biases the human codebook toward the model's own error categories, if AI pre-answers anchor human judgment and inflate measured agreement, and if the six supported formats do not cover the team's actual tracing stack.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.23284",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "claim_id": "claim-02",
      "statement": "Validation claim: it reproduces a prior manual effort's 235 annotation units exactly, and two models labelling the same traces agree 91% of the time at kappa 0.82; no speed, memory, cost, or quality improvement factor is reported.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.23284",
      "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": ""
}