{
  "schema_version": 1,
  "builder_plan": {
    "scenario": "An assistant chooses the wrong API or sends arguments your product cannot use.",
    "change": "Apply the proposed tool-selection or argument-generation change in a sandbox.",
    "baseline": "Your current agent with the same tool schemas and task budget.",
    "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": "Correct completed actions, invalid calls, retries, latency, and cost per completed 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": "2610.18616",
  "title": "Knowledge Graphs as Files: An Agent Harness Where Tool Creation and Tool Use Co-Evolve",
  "summary": "It replaces predefined graph query interfaces and fixed retrieval tools with an agent that writes and revises Python code directly over complete node and edge files, interleaved with shell search. The cost is higher per-query LLM token consumption and multi-turn code execution latency, plus an execution sandbox and the requirement that the knowledge graph be available as flat files. It can fail when the model writes incorrect or inefficient code over large node sets, when graphs exceed practical file-inspection sizes, or when unconstrained code execution introduces security and reproducibility risks in production.",
  "rationale": "The approach is training-free and runs on API-accessible models, so it fits constrained infrastructure, but the quality gain and token costs must be validated against the reader's specific graph and query workload before production adoption.",
  "technique": "Agent-written code tools over raw KG files",
  "family": "uso_de_ferramenta",
  "recommendation": "testar",
  "published": "2026-10-13",
  "as_of": "2026-10-18",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2610.18616",
  "independent_implementations": 0,
  "total_implementations": 1,
  "stars": 10,
  "citations": null,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "It replaces predefined graph query interfaces and fixed retrieval tools with an agent that writes and revises Python code directly over complete node and edge files, interleaved with shell search. The cost is higher per-query LLM token consumption and multi-turn code execution latency, plus an execution sandbox and the requirement that the knowledge graph be available as flat files. It can fail when the model writes incorrect or inefficient code over large node sets, when graphs exceed practical file-inspection sizes, or when unconstrained code execution introduces security and reproducibility risks in production.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.18616",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "claim_id": "claim-02",
      "statement": "Training-free DKGI exceeds the strongest published baseline by at least 6.7 Hit@1 points on every benchmark graph across three domains.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.18616",
      "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": ""
}