{ "schema_version": 1, "builder_plan": { "scenario": "A support or data assistant repeats work across related customer requests.", "change": "Try the paper's memory change in a replay of past tasks, with separate memory per customer.", "baseline": "Your current memory pipeline plus a no-memory control using the same agent.", "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": "Task success, stale memories, retrieval misses, response latency, and all memory maintenance costs.", "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": "2610.11188", "title": "What to Admit and How to Present: Governing Persistent Memory in LLM Agents", "summary": "The method replaces verbatim injection (or relevance-only selection) of persistent memory entries with inference-time admission policies that adjudicate entries or typed sub-units against deterministic eligibility rules before they enter the working context. It costs additional adjudication calls and policy engineering per memory write/read, adding latency and implementation complexity, with no retraining required. It can fail by increasing personalization failures—both designs did—and PS missed its preregistered improvement and personalization-preservation criteria, so beneficial memory use may be lost while safety improves.", "rationale": "The safety gains are statistically supported and implementable without retraining on constrained infrastructure, but the measured increase in personalization failures and missed preregistered criteria require workload-specific validation before production adoption.", "technique": "Policy-gated admission of persistent agent memory", "family": "memoria_e_contexto", "recommendation": "testar", "published": "2026-10-08", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.11188", "independent_implementations": 1, "total_implementations": 2, "stars": 132, "citations": 0, "claims": [ { "claim_id": "claim-01", "statement": "The method replaces verbatim injection (or relevance-only selection) of persistent memory entries with inference-time admission policies that adjudicate entries or typed sub-units against deterministic eligibility rules before they enter the working context. It costs additional adjudication calls and policy engineering per memory write/read, adding latency and implementation complexity, with no retraining required. It can fail by increasing personalization failures—both designs did—and PS missed its preregistered improvement and personalization-preservation criteria, so beneficial memory use may be lost while safety improves.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11188", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "FC and PS admission reduce pooled judge-assessed failure rates by 6.7 and 8.8 percentage points versus verbatim memory injection (p = 2.7e-7 and 4.1e-12); development-set cross-domain leakage falls by up to 29.5 points.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.11188", "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": "" }